fish/immortalwrt
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge Official Source

Browse Source 
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
This commit is contained in:
Tianling Shen 2023-05-23 12:19:10 +08:00
commit bdf6d3135c
435 changed files with 68849 additions and 156 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
include/kernel-6.1 Normal file
View File

@ -0,0 +1,2 @@
LINUX_VERSION-6.1 = .29
LINUX_KERNEL_HASH-6.1.29 = 1e736cc9bd6036379a1d915e518abd4c2c94ad0fd1ea0da961c3489308b8fcfb

2
include/kernel-defaults.mk
View File

@ -75,6 +75,7 @@ endif
rm -f $(LINUX_DIR)/.config.prev
mv $(LINUX_DIR)/.config $(LINUX_DIR)/.config.old
$(call Kernel/SetInitramfs/PreConfigure)
echo "# CONFIG_INITRAMFS_PRESERVE_MTIME is not set" >> $(LINUX_DIR)/.config
ifneq ($(CONFIG_TARGET_ROOTFS_INITRAMFS_SEPARATE),y)
echo 'CONFIG_INITRAMFS_ROOT_UID=$(shell id -u)' >> $(LINUX_DIR)/.config
echo 'CONFIG_INITRAMFS_ROOT_GID=$(shell id -g)' >> $(LINUX_DIR)/.config
@ -99,6 +100,7 @@ define Kernel/SetNoInitramfs
grep -v INITRAMFS $(LINUX_DIR)/.config.old > $(LINUX_DIR)/.config.set
echo 'CONFIG_INITRAMFS_SOURCE=""' >> $(LINUX_DIR)/.config.set
echo '# CONFIG_INITRAMFS_FORCE is not set' >> $(LINUX_DIR)/.config.set
echo "# CONFIG_INITRAMFS_PRESERVE_MTIME is not set" >> $(LINUX_DIR)/.config.set
endef
define Kernel/Configure/Default

12
include/prereq-build.mk
View File

@ -175,25 +175,21 @@ $(eval $(call SetupHostCommand,install,Please install GNU 'install', \
$(eval $(call SetupHostCommand,perl,Please install Perl 5.x, \
perl --version | grep "perl.*v5"))
$(eval $(call CleanupPython2))
$(eval $(call SetupHostCommand,python,Please install Python >= 3.6, \
$(eval $(call SetupHostCommand,python,Please install Python >= 3.7, \
python3.11 -V 2>&1 | grep 'Python 3', \
python3.10 -V 2>&1 | grep 'Python 3', \
python3.9 -V 2>&1 | grep 'Python 3', \
python3.8 -V 2>&1 | grep 'Python 3', \
python3.7 -V 2>&1 | grep 'Python 3', \
python3.6 -V 2>&1 | grep 'Python 3', \
python3 -V 2>&1 | grep -E 'Python 3\.([6-9]|[0-9][0-9])\.?'))
python3 -V 2>&1 | grep -E 'Python 3\.([7-9]|[0-9][0-9])\.?'))
$(eval $(call SetupHostCommand,python3,Please install Python >= 3.6, \
$(eval $(call SetupHostCommand,python3,Please install Python >= 3.7, \
python3.11 -V 2>&1 | grep 'Python 3', \
python3.10 -V 2>&1 | grep 'Python 3', \
python3.9 -V 2>&1 | grep 'Python 3', \
python3.8 -V 2>&1 | grep 'Python 3', \
python3.7 -V 2>&1 | grep 'Python 3', \
python3.6 -V 2>&1 | grep 'Python 3', \
python3 -V 2>&1 | grep -E 'Python 3\.([6-9]|[0-9][0-9])\.?'))
python3 -V 2>&1 | grep -E 'Python 3\.([7-9]|[0-9][0-9])\.?'))
$(eval $(call TestHostCommand,python3-distutils, \
Please install the Python3 distutils module, \

24
include/prereq.mk
View File

@ -30,6 +30,8 @@ define Require
printf "Checking '$(1)'... "
if $(NO_TRACE_MAKE) -f $(firstword $(MAKEFILE_LIST)) check-$(1) >/dev/null 2>/dev/null; then \
echo 'ok.'; \
elif $(NO_TRACE_MAKE) -f $(firstword $(MAKEFILE_LIST)) check-$(1) >/dev/null 2>/dev/null; then \
echo 'updated.'; \
else \
echo 'failed.'; \
echo "$(PKG_NAME): $(strip $(2))" >> $(TMP_DIR)/.prereq-error; \
@ -75,18 +77,6 @@ define RequireCHeader
$$(eval $$(call Require,$(1),$(2)))
endef
define CleanupPython2
define Require/python2-cleanup
if [ -f "$(STAGING_DIR_HOST)/bin/python" ] && \
$(STAGING_DIR_HOST)/bin/python -V 2>&1 | \
grep -q 'Python 2'; then \
rm $(STAGING_DIR_HOST)/bin/python; \
fi
endef
$$(eval $$(call Require,python2-cleanup))
endef
define QuoteHostCommand
'$(subst ','"'"',$(strip $(1)))'
endef
@ -107,7 +97,7 @@ endef
# 3+: candidates
define SetupHostCommand
define Require/$(1)
[ -f "$(STAGING_DIR_HOST)/bin/$(strip $(1))" ] && exit 0; \
mkdir -p "$(STAGING_DIR_HOST)/bin"; \
for cmd in $(call QuoteHostCommand,$(3)) $(call QuoteHostCommand,$(4)) \
$(call QuoteHostCommand,$(5)) $(call QuoteHostCommand,$(6)) \
$(call QuoteHostCommand,$(7)) $(call QuoteHostCommand,$(8)) \
@ -117,9 +107,13 @@ define SetupHostCommand
bin="$$$$$$$$(PATH="$(subst $(space),:,$(filter-out $(STAGING_DIR_HOST)/%,$(subst :,$(space),$(PATH))))" \
command -v "$$$$$$$${cmd%% *}")"; \
if [ -x "$$$$$$$$bin" ] && eval "$$$$$$$$cmd" >/dev/null 2>/dev/null; then \
mkdir -p "$(STAGING_DIR_HOST)/bin"; \
case "$$$$$$$$(ls -dl -- $(STAGING_DIR_HOST)/bin/$(strip $(1)))" in \
*" -> $$$$$$$$bin"*) \
[ -x "$(STAGING_DIR_HOST)/bin/$(strip $(1))" ] && exit 0 \
;; \
esac; \
ln -sf "$$$$$$$$bin" "$(STAGING_DIR_HOST)/bin/$(strip $(1))"; \
exit 0; \
exit 1; \
fi; \
fi; \
done; \

3
package/boot/uboot-envtools/files/ath79
View File

@ -27,6 +27,9 @@ araknis,an-700-ap-i-ac|\
arduino,yun|\
buffalo,bhr-4grv2|\
devolo,magic-2-wifi|\
dlink,dir-859-a1|\
dlink,dir-859-a3|\
dlink,dir-869-a1|\
engenius,eap1200h|\
engenius,eap1750h|\
engenius,eap300-v2|\

4
package/boot/uboot-envtools/files/mediatek_filogic
View File

@ -58,6 +58,10 @@ xiaomi,redmi-router-ax6000-ubootmod)
ubootenv_add_uci_config "$envdev" "0x0" "0x1f000" "0x20000" "1"
ubootenv_add_uci_config "$envdev2" "0x0" "0x1f000" "0x20000" "1"
;;
zyxel,ex5601-t0)
local envdev=/dev/mtd$(find_mtd_index "u-boot-env")
ubootenv_add_uci_config "$envdev" "0x0" "0x20000" "0x40000" "2"
;;
esac
config_load ubootenv

6
package/kernel/linux/modules/crypto.mk
View File

@ -897,10 +897,12 @@ define KernelPackage/crypto-sha1/mpc85xx
AUTOLOAD+=$(call AutoLoad,09,sha1-ppc-spe)
endef
ifndef CONFIG_TARGET_uml
define KernelPackage/crypto-sha1/x86_64
FILES+=$(LINUX_DIR)/arch/x86/crypto/sha1-ssse3.ko
AUTOLOAD+=$(call AutoLoad,09,sha1-ssse3)
endef
endif
ifdef KernelPackage/crypto-sha1/$(ARCH)
KernelPackage/crypto-sha1/$(CRYPTO_TARGET)=\
@ -935,10 +937,12 @@ define KernelPackage/crypto-sha256/mpc85xx
AUTOLOAD+=$(call AutoLoad,09,sha256-ppc-spe)
endef
ifndef CONFIG_TARGET_uml
define KernelPackage/crypto-sha256/x86_64
FILES+=$(LINUX_DIR)/arch/x86/crypto/sha256-ssse3.ko
AUTOLOAD+=$(call AutoLoad,09,sha256-ssse3)
endef
endif
ifdef KernelPackage/crypto-sha256/$(ARCH)
KernelPackage/crypto-sha256/$(CRYPTO_TARGET)=\
@ -977,10 +981,12 @@ endef
KernelPackage/crypto-sha512/tegra=$(KernelPackage/crypto-sha512/arm)
ifndef CONFIG_TARGET_uml
define KernelPackage/crypto-sha512/x86_64
FILES+=$(LINUX_DIR)/arch/x86/crypto/sha512-ssse3.ko
AUTOLOAD+=$(call AutoLoad,09,sha512-ssse3)
endef
endif
ifdef KernelPackage/crypto-sha512/$(ARCH)
KernelPackage/crypto-sha512/$(CRYPTO_TARGET)=\

6
package/kernel/linux/modules/fs.mk
View File

@ -10,7 +10,7 @@ FS_MENU:=Filesystems
define KernelPackage/fs-9p
SUBMENU:=$(FS_MENU)
TITLE:=Plan 9 Resource Sharing Support
DEPENDS:=+kmod-9pnet
DEPENDS:=+kmod-9pnet +LINUX_6_1:kmod-fs-netfs
KCONFIG:=\
CONFIG_9P_FS \
CONFIG_9P_FS_POSIX_ACL=n \
@ -269,7 +269,9 @@ define KernelPackage/fs-fscache
CONFIG_FSCACHE_OBJECT_LIST=n \
CONFIG_CACHEFILES \
CONFIG_CACHEFILES_DEBUG=n \
CONFIG_CACHEFILES_HISTOGRAM=n
CONFIG_CACHEFILES_HISTOGRAM=n \
CONFIG_CACHEFILES_ERROR_INJECTION=n@ge5.17 \
CONFIG_CACHEFILES_ONDEMAND=n@ge5.19
FILES:= \
$(LINUX_DIR)/fs/fscache/fscache.ko \
$(LINUX_DIR)/fs/cachefiles/cachefiles.ko

2
package/kernel/linux/modules/hwmon.mk
View File

@ -52,7 +52,7 @@ define KernelPackage/hwmon-adt7410
$(LINUX_DIR)/drivers/hwmon/adt7x10.ko \
$(LINUX_DIR)/drivers/hwmon/adt7410.ko
AUTOLOAD:=$(call AutoLoad,60,adt7x10 adt7410)
$(call AddDepends/hwmon,+kmod-i2c-core)
$(call AddDepends/hwmon,+kmod-i2c-core +LINUX_6_1:kmod-regmap-core)
endef
define KernelPackage/hwmon-adt7410/description

16
package/kernel/linux/modules/i2c.mk
View File

@ -84,6 +84,20 @@ endef
$(eval $(call KernelPackage,i2c-algo-pcf))
I2C_CCGS_UCSI_MODULES:= \
CONFIG_I2C_CCGX_UCSI:drivers/i2c/busses/i2c-ccgx-ucsi
define KernelPackage/i2c-ccgs-ucsi
$(call i2c_defaults,$(I2C_CCGS_UCSI_MODULES),58)
TITLE:=Cypress CCGx Type-C controller
DEPENDS:=+kmod-i2c-core +kmod-regmap-core
HIDDEN:=y
endef
$(eval $(call KernelPackage,i2c-ccgs-ucsi))
I2C_DWCORE_MODULES:= \
CONFIG_I2C_DESIGNWARE_CORE:drivers/i2c/busses/i2c-designware-core
@ -103,7 +117,7 @@ I2C_DWPCI_MODULES:= \
define KernelPackage/i2c-designware-pci
$(call i2c_defaults,$(I2C_DWPCI_MODULES),59)
TITLE:=Synopsys DesignWare PCI
DEPENDS:=@PCI_SUPPORT +kmod-i2c-designware-core
DEPENDS:=@PCI_SUPPORT +kmod-i2c-designware-core +kmod-i2c-ccgs-ucsi
endef
define KernelPackage/i2c-designware-pci/description

1
package/kernel/linux/modules/lib.mk
View File

@ -134,6 +134,7 @@ define KernelPackage/lib-zstd
FILES:= \
$(LINUX_DIR)/crypto/zstd.ko \
$(LINUX_DIR)/lib/xxhash.ko \
$(LINUX_DIR)/lib/zstd/zstd_common.ko@ge6.1 \
$(LINUX_DIR)/lib/zstd/zstd_compress.ko \
$(LINUX_DIR)/lib/zstd/zstd_decompress.ko
AUTOLOAD:=$(call AutoProbe,xxhash zstd zstd_compress zstd_decompress)

6
package/kernel/linux/modules/netdevices.mk
View File

@ -1468,13 +1468,15 @@ $(eval $(call KernelPackage,sfc-falcon))
define KernelPackage/wwan
SUBMENU:=$(NETWORK_DEVICES_MENU)
TITLE:=WWAN Driver Core
KCONFIG:=CONFIG_WWAN
KCONFIG:= \
CONFIG_WWAN \
CONFIG_WWAN_DEBUGFS=y@ge5.17
FILES:=$(LINUX_DIR)/drivers/net/wwan/wwan.ko
AUTOLOAD:=$(call AutoProbe,wwan)
endef
define KernelPackage/wwan/description
his driver provides a common framework for WWAN drivers.
This driver provides a common framework for WWAN drivers.
endef
$(eval $(call KernelPackage,wwan))

10
package/kernel/linux/modules/netfilter.mk
View File

@ -1164,15 +1164,15 @@ define KernelPackage/nft-offload
DEPENDS:=@IPV6 +kmod-nf-flow +kmod-nft-nat
KCONFIG:= \
CONFIG_NF_FLOW_TABLE_INET \
CONFIG_NF_FLOW_TABLE_IPV4 \
CONFIG_NF_FLOW_TABLE_IPV6 \
CONFIG_NF_FLOW_TABLE_IPV4@lt5.17 \
CONFIG_NF_FLOW_TABLE_IPV6@lt5.17 \
CONFIG_NFT_FLOW_OFFLOAD
FILES:= \
$(LINUX_DIR)/net/netfilter/nf_flow_table_inet.ko \
$(LINUX_DIR)/net/ipv4/netfilter/nf_flow_table_ipv4.ko \
$(LINUX_DIR)/net/ipv6/netfilter/nf_flow_table_ipv6.ko \
$(LINUX_DIR)/net/ipv4/netfilter/nf_flow_table_ipv4.ko@lt5.17 \
$(LINUX_DIR)/net/ipv6/netfilter/nf_flow_table_ipv6.ko@lt5.17 \
$(LINUX_DIR)/net/netfilter/nft_flow_offload.ko
AUTOLOAD:=$(call AutoProbe,nf_flow_table_inet nf_flow_table_ipv4 nf_flow_table_ipv6 nft_flow_offload)
AUTOLOAD:=$(call AutoProbe,nf_flow_table_inet nf_flow_table_ipv4@lt5.17 nf_flow_table_ipv6@lt5.17 nft_flow_offload)
endef
$(eval $(call KernelPackage,nft-offload))

7
package/kernel/linux/modules/netsupport.mk
View File

@ -92,7 +92,9 @@ define KernelPackage/vxlan
+kmod-udptunnel4 \
+IPV6:kmod-udptunnel6
KCONFIG:=CONFIG_VXLAN
FILES:=$(LINUX_DIR)/drivers/net/vxlan.ko
FILES:= \
$(LINUX_DIR)/drivers/net/vxlan.ko@lt5.18 \
$(LINUX_DIR)/drivers/net/vxlan/vxlan.ko@ge5.18
AUTOLOAD:=$(call AutoLoad,13,vxlan)
endef
@ -1314,7 +1316,8 @@ define KernelPackage/9pnet
CONFIG_NET_9P \
CONFIG_NET_9P_DEBUG=n \
CONFIG_NET_9P_XEN=n \
CONFIG_NET_9P_VIRTIO
CONFIG_NET_9P_VIRTIO \
CONFIG_NET_9P_FD=n@ge5.17
FILES:= \
$(LINUX_DIR)/net/9p/9pnet.ko \
$(LINUX_DIR)/net/9p/9pnet_virtio.ko

3
package/kernel/linux/modules/other.mk
View File

@ -57,7 +57,8 @@ define KernelPackage/bluetooth
$(LINUX_DIR)/drivers/bluetooth/hci_uart.ko \
$(LINUX_DIR)/drivers/bluetooth/btusb.ko \
$(LINUX_DIR)/drivers/bluetooth/btintel.ko \
$(LINUX_DIR)/drivers/bluetooth/btrtl.ko
$(LINUX_DIR)/drivers/bluetooth/btrtl.ko \
$(LINUX_DIR)/drivers/bluetooth/btmtk.ko@ge5.17
AUTOLOAD:=$(call AutoProbe,bluetooth rfcomm bnep hidp hci_uart btusb)
endef

4
package/kernel/linux/modules/usb.mk
View File

@ -1156,7 +1156,9 @@ $(eval $(call KernelPackage,usb-net-aqc111))
define KernelPackage/usb-net-asix
TITLE:=Kernel module for USB-to-Ethernet Asix convertors
DEPENDS:=+kmod-libphy +kmod-net-selftests +kmod-mdio-devres +kmod-phy-ax88796b
DEPENDS:= \
+kmod-libphy +kmod-net-selftests +kmod-mdio-devres +kmod-phy-ax88796b \
+LINUX_6_1:kmod-phylink
KCONFIG:=CONFIG_USB_NET_AX8817X
FILES:=$(LINUX_DIR)/drivers/$(USBNET_DIR)/asix.ko
AUTOLOAD:=$(call AutoProbe,asix)

2
package/kernel/mac80211/Makefile
View File

@ -11,7 +11,7 @@ include $(INCLUDE_DIR)/kernel.mk
PKG_NAME:=mac80211
PKG_VERSION:=6.1.24
PKG_RELEASE:=2
PKG_RELEASE:=3
# PKG_SOURCE_URL:=@KERNEL/linux/kernel/projects/backports/stable/v5.15.58/
PKG_SOURCE_URL:=http://mirror2.openwrt.org/sources/
PKG_HASH:=5d39aca7e34c33cb9b3e366117b2e86841b7bdd37933679d6b1e61be6b150648

64
package/kernel/mac80211/patches/brcm/860-brcmfmac-register-wiphy-s-during-module_init.patch
View File

@ -1,64 +0,0 @@
From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <zajec5@gmail.com>
Date: Mon, 8 Jun 2015 16:11:40 +0200
Subject: [PATCH] brcmfmac: register wiphy(s) during module_init
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
This is needed by OpenWrt which expects all PHYs to be created after
module loads successfully.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
---
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/firmware.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/firmware.c
@@ -459,6 +459,7 @@ struct brcmf_fw {
u32 curpos;
unsigned int board_index;
void (*done)(struct device *dev, int err, struct brcmf_fw_request *req);
+ struct completion *completion;
};
#ifdef CONFIG_EFI
@@ -686,6 +687,8 @@ static void brcmf_fw_request_done(const
fwctx->req = NULL;
}
fwctx->done(fwctx->dev, ret, fwctx->req);
+ if (fwctx->completion)
+ complete(fwctx->completion);
kfree(fwctx);
}
@@ -751,6 +754,8 @@ int brcmf_fw_get_firmwares(struct device
{
struct brcmf_fw_item *first = &req->items[0];
struct brcmf_fw *fwctx;
+ struct completion completion;
+ unsigned long time_left;
char *alt_path = NULL;
int ret;
@@ -768,6 +773,9 @@ int brcmf_fw_get_firmwares(struct device
fwctx->dev = dev;
fwctx->req = req;
fwctx->done = fw_cb;
+
+ init_completion(&completion);
+ fwctx->completion = &completion;
/* First try alternative board-specific path if any */
if (fwctx->req->board_types[0])
@@ -787,6 +795,12 @@ int brcmf_fw_get_firmwares(struct device
if (ret < 0)
brcmf_fw_request_done(NULL, fwctx);
+
+ time_left = wait_for_completion_timeout(&completion,
+ msecs_to_jiffies(5000));
+ if (!time_left && fwctx)
+ fwctx->completion = NULL;
+
return 0;
}

4
package/libs/libcap/Makefile
View File

@ -6,12 +6,12 @@
include $(TOPDIR)/rules.mk
PKG_NAME:=libcap
PKG_VERSION:=2.68
PKG_VERSION:=2.69
PKG_RELEASE:=1
PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.xz
PKG_SOURCE_URL:=@KERNEL/linux/libs/security/linux-privs/libcap2
PKG_HASH:=90be3b6d41be5f81ae4b03ec76012b0d27c829293684f6c05b65d5f9cce724b2
PKG_HASH:=f311f8f3dad84699d0566d1d6f7ec943a9298b28f714cae3c931dfd57492d7eb
PKG_MAINTAINER:=Paul Wassi <p.wassi@gmx.at>
PKG_LICENSE:=GPL-2.0-only

4
package/libs/libxml2/Makefile
View File

@ -1,12 +1,12 @@
include $(TOPDIR)/rules.mk
PKG_NAME:=libxml2
PKG_VERSION:=2.11.3
PKG_VERSION:=2.11.4
PKG_RELEASE:=1
PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.xz
PKG_SOURCE_URL:=@GNOME/libxml2/$(basename $(PKG_VERSION))
PKG_HASH:=f1acae1664bda006cd81bfc238238217043d586d06659d5c0e3d1bcebe040870
PKG_HASH:=737e1d7f8ab3f139729ca13a2494fd17bf30ddb4b7a427cf336252cab57f57f7
PKG_LICENSE:=MIT
PKG_LICENSE_FILES:=COPYING

4
package/network/services/lldpd/Makefile
View File

@ -8,12 +8,12 @@
include $(TOPDIR)/rules.mk
PKG_NAME:=lldpd
PKG_VERSION:=1.0.16
PKG_VERSION:=1.0.17
PKG_RELEASE:=1
PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.gz
PKG_SOURCE_URL:=https://github.com/lldpd/lldpd/releases/download/$(PKG_VERSION)/
PKG_HASH:=7753c6e31e938923185f4e10c4ab328929729e22ee4a9687d08881fb82c092ee
PKG_HASH:=89ae691a4917ac9e0ec3b8b2c1e634cc402d43b804f98850c73bd1c7df380882
PKG_MAINTAINER:=Stijn Tintel <stijn@linux-ipv6.be>
PKG_LICENSE:=ISC

4
package/utils/busybox/Makefile
View File

@ -5,14 +5,14 @@
include $(TOPDIR)/rules.mk
PKG_NAME:=busybox
PKG_VERSION:=1.36.0
PKG_VERSION:=1.36.1
PKG_RELEASE:=1
PKG_FLAGS:=essential
PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.bz2
PKG_SOURCE_URL:=https://www.busybox.net/downloads \
http://sources.buildroot.net
PKG_HASH:=542750c8af7cb2630e201780b4f99f3dcceeb06f505b479ec68241c1e6af61a5
PKG_HASH:=b8cc24c9574d809e7279c3be349795c5d5ceb6fdf19ca709f80cde50e47de314
PKG_BUILD_DEPENDS:=BUSYBOX_CONFIG_PAM:libpam
PKG_BUILD_PARALLEL:=1

8
target/linux/ath25/files/drivers/net/phy/mvswitch.c
View File

@ -312,7 +312,11 @@ mvswitch_config_init(struct phy_device *pdev)
priv->orig_features = dev->features;
#ifdef HEADER_MODE
#if LINUX_VERSION_CODE >= KERNEL_VERSION(6,1,0)
dev->priv_flags |= IFF_NO_IP_ALIGN;
#else
dev->extra_priv_flags |= IFF_NO_IP_ALIGN;
#endif
dev->features |= NETIF_F_HW_VLAN_CTAG_RX | NETIF_F_HW_VLAN_CTAG_TX;
#else
dev->features |= NETIF_F_HW_VLAN_CTAG_RX;
@ -372,7 +376,11 @@ mvswitch_detach(struct phy_device *pdev)
dev->eth_mangle_rx = NULL;
dev->eth_mangle_tx = NULL;
dev->features = priv->orig_features;
#if LINUX_VERSION_CODE >= KERNEL_VERSION(6,1,0)
dev->priv_flags &= ~IFF_NO_IP_ALIGN;
#else
dev->extra_priv_flags &= ~IFF_NO_IP_ALIGN;
#endif
}
static void

1
target/linux/ath79/dts/qca9563_dlink_dir-859-a1.dts
View File

@ -1,7 +1,6 @@
// SPDX-License-Identifier: GPL-2.0-or-later OR MIT
#include "qca9563_dlink_dir-8x9-a1.dtsi"
#include <dt-bindings/gpio/gpio.h>
/ {
model = "D-Link DIR-859 A1";

64
target/linux/ath79/dts/qca9563_dlink_dir-859-a3.dts Normal file
View File

@ -0,0 +1,64 @@
// SPDX-License-Identifier: GPL-2.0-or-later OR MIT
#include "qca9563_dlink_dir-8x9-a1.dtsi"
#include <dt-bindings/leds/common.h>
/ {
compatible = "dlink,dir-859-a3", "qca,qca9563";
model = "D-Link DIR-859 A3";
aliases {
label-mac-device = &wmac;
led-boot = &led_power;
led-failsafe = &led_power;
led-running = &led_power;
led-upgrade = &led_power;
};
leds {
compatible = "gpio-leds";
led_power: power {
label = "green:power";
color = <LED_COLOR_ID_GREEN>;
function = LED_FUNCTION_POWER;
gpios = <&gpio 15 GPIO_ACTIVE_LOW>;
};
wan {
color = <LED_COLOR_ID_GREEN>;
function = LED_FUNCTION_WAN;
gpios = <&gpio 16 GPIO_ACTIVE_LOW>;
};
wlan {
color = <LED_COLOR_ID_GREEN>;
function = LED_FUNCTION_WLAN;
gpios = <&gpio 19 GPIO_ACTIVE_LOW>;
linux,default-trigger = "phy0tpt";
};
wps {
color = <LED_COLOR_ID_GREEN>;
function = LED_FUNCTION_WPS;
gpios = <&gpio 8 GPIO_ACTIVE_LOW>;
};
};
};
&ath10k {
nvmem-cells = <&calibration_ath10k>, <&macaddr_bdcfg_ethaddr 2>;
nvmem-cell-names = "calibration", "mac-address";
};
&bdcfg {
macaddr_bdcfg_ethaddr: ethaddr {
#nvmem-cell-cells = <1>;
};
};
&wmac {
nvmem-cells = <&calibration_ath9k>, <&macaddr_bdcfg_ethaddr 0>;
nvmem-cell-names = "calibration", "mac-address";
};

2
target/linux/ath79/dts/qca9563_dlink_dir-869-a1.dts
View File

@ -1,7 +1,6 @@
// SPDX-License-Identifier: GPL-2.0-or-later OR MIT
#include "qca9563_dlink_dir-8x9-a1.dtsi"
#include <dt-bindings/gpio/gpio.h>
/ {
model = "D-Link DIR-869 A1";
@ -28,4 +27,3 @@
};
};
};

19
target/linux/ath79/dts/qca9563_dlink_dir-8x9-a1.dtsi
View File

@ -6,17 +6,18 @@
#include <dt-bindings/input/input.h>
/ {
keys {
compatible = "gpio-keys";
wps {
label = "wps";
linux,code = <KEY_WPS_BUTTON>;
gpios = <&gpio 1 GPIO_ACTIVE_LOW>;
debounce-interval = <60>;
};
reset {
label = "reset";
linux,code = <KEY_RESTART>;
gpios = <&gpio 2 GPIO_ACTIVE_LOW>;
debounce-interval = <60>;
@ -27,7 +28,7 @@
&pcie {
status = "okay";
wifi@0,0 {
ath10k: wifi@0,0 {
compatible = "qcom,ath10k";
reg = <0x0000 0 0 0 0>;
@ -42,7 +43,7 @@
flash@0 {
compatible = "jedec,spi-nor";
reg = <0>;
spi-max-frequency = <50000000>;
spi-max-frequency = <25000000>;
partitions {
compatible = "fixed-partitions";
@ -51,19 +52,19 @@
partition@0 {
label = "bootloader";
reg = <0x000000 0x40000>;
reg = <0x000000 0x040000>;
read-only;
};
partition@40000 {
bdcfg: partition@40000 {
compatible = "u-boot,env";
label = "bdcfg";
reg = <0x040000 0x10000>;
read-only;
reg = <0x040000 0x010000>;
};
partition@50000 {
label = "devdata";
reg = <0x050000 0x10000>;
reg = <0x050000 0x010000>;
read-only;
compatible = "nvmem-cells";
@ -81,7 +82,7 @@
partition@60000 {
label = "devconf";
reg = <0x060000 0x10000>;
reg = <0x060000 0x010000>;
read-only;
};

3
target/linux/ath79/generic/base-files/etc/board.d/01_leds
View File

@ -234,6 +234,9 @@ dlink,dap-1365-a1)
dlink,dir-859-a1)
ucidef_set_led_switch "internet" "WAN" "green:internet" "switch0" "0x20"
;;
dlink,dir-859-a3)
ucidef_set_led_switch "wan" "WAN" "green:wan" "switch0" "0x20"
;;
engenius,ens202ext-v1|\
engenius,enstationac-v1)
ucidef_set_rssimon "wlan0" "200000" "1"

5
target/linux/ath79/generic/base-files/etc/board.d/02_network
View File

@ -291,6 +291,7 @@ ath79_setup_interfaces()
dlink,dir-842-c2|\
dlink,dir-842-c3|\
dlink,dir-859-a1|\
dlink,dir-859-a3|\
dlink,dir-869-a1|\
engenius,epg5000|\
engenius,esr1200|\
@ -661,6 +662,10 @@ ath79_setup_macs()
lan_mac=$(mtd_get_mac_text "devdata" 0xc9)
wan_mac=$(mtd_get_mac_text "devdata" 0x79)
;;
dlink,dir-859-a3)
lan_mac=$(get_mac_label)
wan_mac=$(macaddr_add "$lan_mac" 3)
;;
qihoo,c301|\
wd,mynet-n600|\
wd,mynet-n750)

2
target/linux/ath79/generic/base-files/etc/uci-defaults/09_fix-checksum
View File

@ -22,6 +22,8 @@ dlink,dap-3662-a1)
fixwrgg
;;
dlink,dir-629-a1|\
dlink,dir-859-a1|\
dlink,dir-859-a3|\
dlink,dir-869-a1|\
qihoo,c301)
fix_seama_header

2
target/linux/ath79/generic/config-default
View File

@ -15,6 +15,8 @@ CONFIG_MARVELL_PHY=y
CONFIG_MICREL_PHY=y
CONFIG_MTD_REDBOOT_PARTS=y
CONFIG_MTD_SPLIT_EVA_FW=y
CONFIG_NVMEM_SYSFS=y
CONFIG_NVMEM_U_BOOT_ENV=y
CONFIG_PHY_AR7100_USB=y
CONFIG_PHY_AR7200_USB=y
CONFIG_REALTEK_PHY=y

14
target/linux/ath79/image/generic.mk
View File

@ -1144,18 +1144,28 @@ define Device/dlink_dir-842-c3
endef
TARGET_DEVICES += dlink_dir-842-c3
define Device/dlink_dir-859-a1
define Device/dlink_dir-859-ax
$(Device/seama)
SOC := qca9563
DEVICE_VENDOR := D-Link
DEVICE_MODEL := DIR-859
DEVICE_VARIANT := A1
IMAGE_SIZE := 15872k
DEVICE_PACKAGES := kmod-usb2 kmod-ath10k-ct-smallbuffers ath10k-firmware-qca988x-ct
SEAMA_SIGNATURE := wrgac37_dlink.2013gui_dir859
endef
define Device/dlink_dir-859-a1
$(Device/dlink_dir-859-ax)
DEVICE_VARIANT := A1
endef
TARGET_DEVICES += dlink_dir-859-a1
define Device/dlink_dir-859-a3
$(Device/dlink_dir-859-ax)
DEVICE_VARIANT := A3
endef
TARGET_DEVICES += dlink_dir-859-a3
define Device/dlink_dir-869-a1
$(Device/seama)
SOC := qca9563

187
target/linux/generic/backport-6.1/020-v6.3-19-mm-add-vma_has_recency.patch Normal file
View File

@ -0,0 +1,187 @@
From 6c7f552a48b49a8612786a28a2239fbc24fac289 Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Fri, 30 Dec 2022 14:52:51 -0700
Subject: [PATCH 19/29] mm: add vma_has_recency()
Add vma_has_recency() to indicate whether a VMA may exhibit temporal
locality that the LRU algorithm relies on.
This function returns false for VMAs marked by VM_SEQ_READ or
VM_RAND_READ. While the former flag indicates linear access, i.e., a
special case of spatial locality, both flags indicate a lack of temporal
locality, i.e., the reuse of an area within a relatively small duration.
"Recency" is chosen over "locality" to avoid confusion between temporal
and spatial localities.
Before this patch, the active/inactive LRU only ignored the accessed bit
from VMAs marked by VM_SEQ_READ. After this patch, the active/inactive
LRU and MGLRU share the same logic: they both ignore the accessed bit if
vma_has_recency() returns false.
For the active/inactive LRU, the following fio test showed a [6, 8]%
increase in IOPS when randomly accessing mapped files under memory
pressure.
kb=$(awk '/MemTotal/ { print $2 }' /proc/meminfo)
kb=$((kb - 8*1024*1024))
modprobe brd rd_nr=1 rd_size=$kb
dd if=/dev/zero of=/dev/ram0 bs=1M
mkfs.ext4 /dev/ram0
mount /dev/ram0 /mnt/
swapoff -a
fio --name=test --directory=/mnt/ --ioengine=mmap --numjobs=8 \
--size=8G --rw=randrw --time_based --runtime=10m \
--group_reporting
The discussion that led to this patch is here [1]. Additional test
results are available in that thread.
[1] https://lore.kernel.org/r/Y31s%2FK8T85jh05wH@google.com/
Link: https://lkml.kernel.org/r/20221230215252.2628425-1-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Andrea Righi <andrea.righi@canonical.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
include/linux/mm_inline.h | 9 +++++++++
mm/memory.c | 8 ++++----
mm/rmap.c | 42 +++++++++++++++++----------------------
mm/vmscan.c | 5 ++++-
4 files changed, 35 insertions(+), 29 deletions(-)
--- a/include/linux/mm_inline.h
+++ b/include/linux/mm_inline.h
@@ -578,4 +578,12 @@ pte_install_uffd_wp_if_needed(struct vm_
#endif
}
+static inline bool vma_has_recency(struct vm_area_struct *vma)
+{
+ if (vma->vm_flags & (VM_SEQ_READ | VM_RAND_READ))
+ return false;
+
+ return true;
+}
+
#endif
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -1435,8 +1435,7 @@ again:
force_flush = 1;
set_page_dirty(page);
}
- if (pte_young(ptent) &&
- likely(!(vma->vm_flags & VM_SEQ_READ)))
+ if (pte_young(ptent) && likely(vma_has_recency(vma)))
mark_page_accessed(page);
}
rss[mm_counter(page)]--;
@@ -5170,8 +5169,8 @@ static inline void mm_account_fault(stru
#ifdef CONFIG_LRU_GEN
static void lru_gen_enter_fault(struct vm_area_struct *vma)
{
- /* the LRU algorithm doesn't apply to sequential or random reads */
- current->in_lru_fault = !(vma->vm_flags & (VM_SEQ_READ | VM_RAND_READ));
+ /* the LRU algorithm only applies to accesses with recency */
+ current->in_lru_fault = vma_has_recency(vma);
}
static void lru_gen_exit_fault(void)
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -823,25 +823,14 @@ static bool folio_referenced_one(struct
}
if (pvmw.pte) {
- if (lru_gen_enabled() && pte_young(*pvmw.pte) &&
- !(vma->vm_flags & (VM_SEQ_READ | VM_RAND_READ))) {
+ if (lru_gen_enabled() && pte_young(*pvmw.pte)) {
lru_gen_look_around(&pvmw);
referenced++;
}
if (ptep_clear_flush_young_notify(vma, address,
- pvmw.pte)) {
- /*
- * Don't treat a reference through
- * a sequentially read mapping as such.
- * If the folio has been used in another mapping,
- * we will catch it; if this other mapping is
- * already gone, the unmap path will have set
- * the referenced flag or activated the folio.
- */
- if (likely(!(vma->vm_flags & VM_SEQ_READ)))
- referenced++;
- }
+ pvmw.pte))
+ referenced++;
} else if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE)) {
if (pmdp_clear_flush_young_notify(vma, address,
pvmw.pmd))
@@ -875,7 +864,20 @@ static bool invalid_folio_referenced_vma
struct folio_referenced_arg *pra = arg;
struct mem_cgroup *memcg = pra->memcg;
- if (!mm_match_cgroup(vma->vm_mm, memcg))
+ /*
+ * Ignore references from this mapping if it has no recency. If the
+ * page has been used in another mapping, we will catch it; if this
+ * other mapping is already gone, the unmap path will have set the
+ * referenced flag or activated the page in zap_pte_range().
+ */
+ if (!vma_has_recency(vma))
+ return true;
+
+ /*
+ * If we are reclaiming on behalf of a cgroup, skip counting on behalf
+ * of references from different cgroups.
+ */
+ if (memcg && !mm_match_cgroup(vma->vm_mm, memcg))
return true;
return false;
@@ -906,6 +908,7 @@ int folio_referenced(struct folio *folio
.arg = (void *)&pra,
.anon_lock = folio_lock_anon_vma_read,
.try_lock = true,
+ .invalid_vma = invalid_folio_referenced_vma,
};
*vm_flags = 0;
@@ -921,15 +924,6 @@ int folio_referenced(struct folio *folio
return 1;
}
- /*
- * If we are reclaiming on behalf of a cgroup, skip
- * counting on behalf of references from different
- * cgroups
- */
- if (memcg) {
- rwc.invalid_vma = invalid_folio_referenced_vma;
- }
-
rmap_walk(folio, &rwc);
*vm_flags = pra.vm_flags;
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3766,7 +3766,10 @@ static int should_skip_vma(unsigned long
if (is_vm_hugetlb_page(vma))
return true;
- if (vma->vm_flags & (VM_LOCKED | VM_SPECIAL | VM_SEQ_READ | VM_RAND_READ))
+ if (!vma_has_recency(vma))
+ return true;
+
+ if (vma->vm_flags & (VM_LOCKED | VM_SPECIAL))
return true;
if (vma == get_gate_vma(vma->vm_mm))

125
target/linux/generic/backport-6.1/020-v6.3-20-mm-support-POSIX_FADV_NOREUSE.patch Normal file
View File

@ -0,0 +1,125 @@
From 686c3d4f71de9e0e7a27f03a5617a712385f90cd Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Fri, 30 Dec 2022 14:52:52 -0700
Subject: [PATCH 20/29] mm: support POSIX_FADV_NOREUSE
This patch adds POSIX_FADV_NOREUSE to vma_has_recency() so that the LRU
algorithm can ignore access to mapped files marked by this flag.
The advantages of POSIX_FADV_NOREUSE are:
1. Unlike MADV_SEQUENTIAL and MADV_RANDOM, it does not alter the
default readahead behavior.
2. Unlike MADV_SEQUENTIAL and MADV_RANDOM, it does not split VMAs and
therefore does not take mmap_lock.
3. Unlike MADV_COLD, setting it has a negligible cost, regardless of
how many pages it affects.
Its limitations are:
1. Like POSIX_FADV_RANDOM and POSIX_FADV_SEQUENTIAL, it currently does
not support range. IOW, its scope is the entire file.
2. It currently does not ignore access through file descriptors.
Specifically, for the active/inactive LRU, given a file page shared
by two users and one of them having set POSIX_FADV_NOREUSE on the
file, this page will be activated upon the second user accessing
it. This corner case can be covered by checking POSIX_FADV_NOREUSE
before calling mark_page_accessed() on the read path. But it is
considered not worth the effort.
There have been a few attempts to support POSIX_FADV_NOREUSE, e.g., [1].
This time the goal is to fill a niche: a few desktop applications, e.g.,
large file transferring and video encoding/decoding, want fast file
streaming with mmap() rather than direct IO. Among those applications, an
SVT-AV1 regression was reported when running with MGLRU [2]. The
following test can reproduce that regression.
kb=$(awk '/MemTotal/ { print $2 }' /proc/meminfo)
kb=$((kb - 8*1024*1024))
modprobe brd rd_nr=1 rd_size=$kb
dd if=/dev/zero of=/dev/ram0 bs=1M
mkfs.ext4 /dev/ram0
mount /dev/ram0 /mnt/
swapoff -a
fallocate -l 8G /mnt/swapfile
mkswap /mnt/swapfile
swapon /mnt/swapfile
wget http://ultravideo.cs.tut.fi/video/Bosphorus_3840x2160_120fps_420_8bit_YUV_Y4M.7z
7z e -o/mnt/ Bosphorus_3840x2160_120fps_420_8bit_YUV_Y4M.7z
SvtAv1EncApp --preset 12 -w 3840 -h 2160 \
-i /mnt/Bosphorus_3840x2160.y4m
For MGLRU, the following change showed a [9-11]% increase in FPS,
which makes it on par with the active/inactive LRU.
patch Source/App/EncApp/EbAppMain.c <<EOF
31a32
> #include <fcntl.h>
35d35
< #include <fcntl.h> /* _O_BINARY */
117a118
> posix_fadvise(config->mmap.fd, 0, 0, POSIX_FADV_NOREUSE);
EOF
[1] https://lore.kernel.org/r/1308923350-7932-1-git-send-email-andrea@betterlinux.com/
[2] https://openbenchmarking.org/result/2209259-PTS-MGLRU8GB57
Link: https://lkml.kernel.org/r/20221230215252.2628425-2-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Andrea Righi <andrea.righi@canonical.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
include/linux/fs.h | 2 ++
include/linux/mm_inline.h | 3 +++
mm/fadvise.c | 5 ++++-
3 files changed, 9 insertions(+), 1 deletion(-)
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -166,6 +166,8 @@ typedef int (dio_iodone_t)(struct kiocb
/* File supports DIRECT IO */
#define FMODE_CAN_ODIRECT ((__force fmode_t)0x400000)
+#define FMODE_NOREUSE ((__force fmode_t)0x400000)
+
/* File was opened by fanotify and shouldn't generate fanotify events */
#define FMODE_NONOTIFY ((__force fmode_t)0x4000000)
--- a/include/linux/mm_inline.h
+++ b/include/linux/mm_inline.h
@@ -583,6 +583,9 @@ static inline bool vma_has_recency(struc
if (vma->vm_flags & (VM_SEQ_READ | VM_RAND_READ))
return false;
+ if (vma->vm_file && (vma->vm_file->f_mode & FMODE_NOREUSE))
+ return false;
+
return true;
}
--- a/mm/fadvise.c
+++ b/mm/fadvise.c
@@ -80,7 +80,7 @@ int generic_fadvise(struct file *file, l
case POSIX_FADV_NORMAL:
file->f_ra.ra_pages = bdi->ra_pages;
spin_lock(&file->f_lock);
- file->f_mode &= ~FMODE_RANDOM;
+ file->f_mode &= ~(FMODE_RANDOM | FMODE_NOREUSE);
spin_unlock(&file->f_lock);
break;
case POSIX_FADV_RANDOM:
@@ -107,6 +107,9 @@ int generic_fadvise(struct file *file, l
force_page_cache_readahead(mapping, file, start_index, nrpages);
break;
case POSIX_FADV_NOREUSE:
+ spin_lock(&file->f_lock);
+ file->f_mode |= FMODE_NOREUSE;
+ spin_unlock(&file->f_lock);
break;
case POSIX_FADV_DONTNEED:
__filemap_fdatawrite_range(mapping, offset, endbyte,

348
target/linux/generic/backport-6.1/020-v6.3-21-mm-multi-gen-LRU-rename-lru_gen_struct-to-lru_gen_pa.patch Normal file
View File

@ -0,0 +1,348 @@
From 348fdbada9fb3f0bf1a53651be46319105af187f Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:18:59 -0700
Subject: [PATCH 21/29] mm: multi-gen LRU: rename lru_gen_struct to
lru_gen_folio
Patch series "mm: multi-gen LRU: memcg LRU", v3.
Overview
========
An memcg LRU is a per-node LRU of memcgs. It is also an LRU of LRUs,
since each node and memcg combination has an LRU of folios (see
mem_cgroup_lruvec()).
Its goal is to improve the scalability of global reclaim, which is
critical to system-wide memory overcommit in data centers. Note that
memcg reclaim is currently out of scope.
Its memory bloat is a pointer to each lruvec and negligible to each
pglist_data. In terms of traversing memcgs during global reclaim, it
improves the best-case complexity from O(n) to O(1) and does not affect
the worst-case complexity O(n). Therefore, on average, it has a sublinear
complexity in contrast to the current linear complexity.
The basic structure of an memcg LRU can be understood by an analogy to
the active/inactive LRU (of folios):
1. It has the young and the old (generations), i.e., the counterparts
to the active and the inactive;
2. The increment of max_seq triggers promotion, i.e., the counterpart
to activation;
3. Other events trigger similar operations, e.g., offlining an memcg
triggers demotion, i.e., the counterpart to deactivation.
In terms of global reclaim, it has two distinct features:
1. Sharding, which allows each thread to start at a random memcg (in
the old generation) and improves parallelism;
2. Eventual fairness, which allows direct reclaim to bail out at will
and reduces latency without affecting fairness over some time.
The commit message in patch 6 details the workflow:
https://lore.kernel.org/r/20221222041905.2431096-7-yuzhao@google.com/
The following is a simple test to quickly verify its effectiveness.
Test design:
1. Create multiple memcgs.
2. Each memcg contains a job (fio).
3. All jobs access the same amount of memory randomly.
4. The system does not experience global memory pressure.
5. Periodically write to the root memory.reclaim.
Desired outcome:
1. All memcgs have similar pgsteal counts, i.e., stddev(pgsteal)
over mean(pgsteal) is close to 0%.
2. The total pgsteal is close to the total requested through
memory.reclaim, i.e., sum(pgsteal) over sum(requested) is close
to 100%.
Actual outcome [1]:
MGLRU off MGLRU on
stddev(pgsteal) / mean(pgsteal) 75% 20%
sum(pgsteal) / sum(requested) 425% 95%
####################################################################
MEMCGS=128
for ((memcg = 0; memcg < $MEMCGS; memcg++)); do
mkdir /sys/fs/cgroup/memcg$memcg
done
start() {
echo $BASHPID > /sys/fs/cgroup/memcg$memcg/cgroup.procs
fio -name=memcg$memcg --numjobs=1 --ioengine=mmap \
--filename=/dev/zero --size=1920M --rw=randrw \
--rate=64m,64m --random_distribution=random \
--fadvise_hint=0 --time_based --runtime=10h \
--group_reporting --minimal
}
for ((memcg = 0; memcg < $MEMCGS; memcg++)); do
start &
done
sleep 600
for ((i = 0; i < 600; i++)); do
echo 256m >/sys/fs/cgroup/memory.reclaim
sleep 6
done
for ((memcg = 0; memcg < $MEMCGS; memcg++)); do
grep "pgsteal " /sys/fs/cgroup/memcg$memcg/memory.stat
done
####################################################################
[1]: This was obtained from running the above script (touches less
than 256GB memory) on an EPYC 7B13 with 512GB DRAM for over an
hour.
This patch (of 8):
The new name lru_gen_folio will be more distinct from the coming
lru_gen_memcg.
Link: https://lkml.kernel.org/r/20221222041905.2431096-1-yuzhao@google.com
Link: https://lkml.kernel.org/r/20221222041905.2431096-2-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
include/linux/mm_inline.h | 4 ++--
include/linux/mmzone.h | 6 +++---
mm/vmscan.c | 34 +++++++++++++++++-----------------
mm/workingset.c | 4 ++--
4 files changed, 24 insertions(+), 24 deletions(-)
--- a/include/linux/mm_inline.h
+++ b/include/linux/mm_inline.h
@@ -178,7 +178,7 @@ static inline void lru_gen_update_size(s
int zone = folio_zonenum(folio);
int delta = folio_nr_pages(folio);
enum lru_list lru = type * LRU_INACTIVE_FILE;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
VM_WARN_ON_ONCE(old_gen != -1 && old_gen >= MAX_NR_GENS);
VM_WARN_ON_ONCE(new_gen != -1 && new_gen >= MAX_NR_GENS);
@@ -224,7 +224,7 @@ static inline bool lru_gen_add_folio(str
int gen = folio_lru_gen(folio);
int type = folio_is_file_lru(folio);
int zone = folio_zonenum(folio);
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
VM_WARN_ON_ONCE_FOLIO(gen != -1, folio);
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -404,7 +404,7 @@ enum {
* The number of pages in each generation is eventually consistent and therefore
* can be transiently negative when reset_batch_size() is pending.
*/
-struct lru_gen_struct {
+struct lru_gen_folio {
/* the aging increments the youngest generation number */
unsigned long max_seq;
/* the eviction increments the oldest generation numbers */
@@ -461,7 +461,7 @@ struct lru_gen_mm_state {
struct lru_gen_mm_walk {
/* the lruvec under reclaim */
struct lruvec *lruvec;
- /* unstable max_seq from lru_gen_struct */
+ /* unstable max_seq from lru_gen_folio */
unsigned long max_seq;
/* the next address within an mm to scan */
unsigned long next_addr;
@@ -524,7 +524,7 @@ struct lruvec {
unsigned long flags;
#ifdef CONFIG_LRU_GEN
/* evictable pages divided into generations */
- struct lru_gen_struct lrugen;
+ struct lru_gen_folio lrugen;
/* to concurrently iterate lru_gen_mm_list */
struct lru_gen_mm_state mm_state;
#endif
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3190,7 +3190,7 @@ static int get_nr_gens(struct lruvec *lr
static bool __maybe_unused seq_is_valid(struct lruvec *lruvec)
{
- /* see the comment on lru_gen_struct */
+ /* see the comment on lru_gen_folio */
return get_nr_gens(lruvec, LRU_GEN_FILE) >= MIN_NR_GENS &&
get_nr_gens(lruvec, LRU_GEN_FILE) <= get_nr_gens(lruvec, LRU_GEN_ANON) &&
get_nr_gens(lruvec, LRU_GEN_ANON) <= MAX_NR_GENS;
@@ -3596,7 +3596,7 @@ struct ctrl_pos {
static void read_ctrl_pos(struct lruvec *lruvec, int type, int tier, int gain,
struct ctrl_pos *pos)
{
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
int hist = lru_hist_from_seq(lrugen->min_seq[type]);
pos->refaulted = lrugen->avg_refaulted[type][tier] +
@@ -3611,7 +3611,7 @@ static void read_ctrl_pos(struct lruvec
static void reset_ctrl_pos(struct lruvec *lruvec, int type, bool carryover)
{
int hist, tier;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
bool clear = carryover ? NR_HIST_GENS == 1 : NR_HIST_GENS > 1;
unsigned long seq = carryover ? lrugen->min_seq[type] : lrugen->max_seq + 1;
@@ -3688,7 +3688,7 @@ static int folio_update_gen(struct folio
static int folio_inc_gen(struct lruvec *lruvec, struct folio *folio, bool reclaiming)
{
int type = folio_is_file_lru(folio);
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
int new_gen, old_gen = lru_gen_from_seq(lrugen->min_seq[type]);
unsigned long new_flags, old_flags = READ_ONCE(folio->flags);
@@ -3733,7 +3733,7 @@ static void update_batch_size(struct lru
static void reset_batch_size(struct lruvec *lruvec, struct lru_gen_mm_walk *walk)
{
int gen, type, zone;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
walk->batched = 0;
@@ -4253,7 +4253,7 @@ static bool inc_min_seq(struct lruvec *l
{
int zone;
int remaining = MAX_LRU_BATCH;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
int new_gen, old_gen = lru_gen_from_seq(lrugen->min_seq[type]);
if (type == LRU_GEN_ANON && !can_swap)
@@ -4289,7 +4289,7 @@ static bool try_to_inc_min_seq(struct lr
{
int gen, type, zone;
bool success = false;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
DEFINE_MIN_SEQ(lruvec);
VM_WARN_ON_ONCE(!seq_is_valid(lruvec));
@@ -4310,7 +4310,7 @@ next:
;
}
- /* see the comment on lru_gen_struct */
+ /* see the comment on lru_gen_folio */
if (can_swap) {
min_seq[LRU_GEN_ANON] = min(min_seq[LRU_GEN_ANON], min_seq[LRU_GEN_FILE]);
min_seq[LRU_GEN_FILE] = max(min_seq[LRU_GEN_ANON], lrugen->min_seq[LRU_GEN_FILE]);
@@ -4332,7 +4332,7 @@ static void inc_max_seq(struct lruvec *l
{
int prev, next;
int type, zone;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
spin_lock_irq(&lruvec->lru_lock);
@@ -4390,7 +4390,7 @@ static bool try_to_inc_max_seq(struct lr
bool success;
struct lru_gen_mm_walk *walk;
struct mm_struct *mm = NULL;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
VM_WARN_ON_ONCE(max_seq > READ_ONCE(lrugen->max_seq));
@@ -4455,7 +4455,7 @@ static bool should_run_aging(struct lruv
unsigned long old = 0;
unsigned long young = 0;
unsigned long total = 0;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
for (type = !can_swap; type < ANON_AND_FILE; type++) {
@@ -4740,7 +4740,7 @@ static bool sort_folio(struct lruvec *lr
int delta = folio_nr_pages(folio);
int refs = folio_lru_refs(folio);
int tier = lru_tier_from_refs(refs);
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
VM_WARN_ON_ONCE_FOLIO(gen >= MAX_NR_GENS, folio);
@@ -4840,7 +4840,7 @@ static int scan_folios(struct lruvec *lr
int scanned = 0;
int isolated = 0;
int remaining = MAX_LRU_BATCH;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
VM_WARN_ON_ONCE(!list_empty(list));
@@ -5240,7 +5240,7 @@ done:
static bool __maybe_unused state_is_valid(struct lruvec *lruvec)
{
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
if (lrugen->enabled) {
enum lru_list lru;
@@ -5522,7 +5522,7 @@ static void lru_gen_seq_show_full(struct
int i;
int type, tier;
int hist = lru_hist_from_seq(seq);
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
for (tier = 0; tier < MAX_NR_TIERS; tier++) {
seq_printf(m, " %10d", tier);
@@ -5572,7 +5572,7 @@ static int lru_gen_seq_show(struct seq_f
unsigned long seq;
bool full = !debugfs_real_fops(m->file)->write;
struct lruvec *lruvec = v;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
int nid = lruvec_pgdat(lruvec)->node_id;
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
DEFINE_MAX_SEQ(lruvec);
@@ -5826,7 +5826,7 @@ void lru_gen_init_lruvec(struct lruvec *
{
int i;
int gen, type, zone;
- struct lru_gen_struct *lrugen = &lruvec->lrugen;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
lrugen->max_seq = MIN_NR_GENS + 1;
lrugen->enabled = lru_gen_enabled();
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -223,7 +223,7 @@ static void *lru_gen_eviction(struct fol
unsigned long token;
unsigned long min_seq;
struct lruvec *lruvec;
- struct lru_gen_struct *lrugen;
+ struct lru_gen_folio *lrugen;
int type = folio_is_file_lru(folio);
int delta = folio_nr_pages(folio);
int refs = folio_lru_refs(folio);
@@ -252,7 +252,7 @@ static void lru_gen_refault(struct folio
unsigned long token;
unsigned long min_seq;
struct lruvec *lruvec;
- struct lru_gen_struct *lrugen;
+ struct lru_gen_folio *lrugen;
struct mem_cgroup *memcg;
struct pglist_data *pgdat;
int type = folio_is_file_lru(folio);

162
target/linux/generic/backport-6.1/020-v6.3-22-mm-multi-gen-LRU-rename-lrugen-lists-to-lrugen-pages.patch Normal file
View File

@ -0,0 +1,162 @@
From afd37e73db04c7e6b47411120ac5f6a7eca51fec Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:19:00 -0700
Subject: [PATCH 22/29] mm: multi-gen LRU: rename lrugen->lists[] to
lrugen->folios[]
lru_gen_folio will be chained into per-node lists by the coming
lrugen->list.
Link: https://lkml.kernel.org/r/20221222041905.2431096-3-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
include/linux/mm_inline.h | 4 ++--
include/linux/mmzone.h | 8 ++++----
mm/vmscan.c | 20 ++++++++++----------
3 files changed, 16 insertions(+), 16 deletions(-)
--- a/include/linux/mm_inline.h
+++ b/include/linux/mm_inline.h
@@ -256,9 +256,9 @@ static inline bool lru_gen_add_folio(str
lru_gen_update_size(lruvec, folio, -1, gen);
/* for folio_rotate_reclaimable() */
if (reclaiming)
- list_add_tail(&folio->lru, &lrugen->lists[gen][type][zone]);
+ list_add_tail(&folio->lru, &lrugen->folios[gen][type][zone]);
else
- list_add(&folio->lru, &lrugen->lists[gen][type][zone]);
+ list_add(&folio->lru, &lrugen->folios[gen][type][zone]);
return true;
}
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -312,7 +312,7 @@ enum lruvec_flags {
* They form a sliding window of a variable size [MIN_NR_GENS, MAX_NR_GENS]. An
* offset within MAX_NR_GENS, i.e., gen, indexes the LRU list of the
* corresponding generation. The gen counter in folio->flags stores gen+1 while
- * a page is on one of lrugen->lists[]. Otherwise it stores 0.
+ * a page is on one of lrugen->folios[]. Otherwise it stores 0.
*
* A page is added to the youngest generation on faulting. The aging needs to
* check the accessed bit at least twice before handing this page over to the
@@ -324,8 +324,8 @@ enum lruvec_flags {
* rest of generations, if they exist, are considered inactive. See
* lru_gen_is_active().
*
- * PG_active is always cleared while a page is on one of lrugen->lists[] so that
- * the aging needs not to worry about it. And it's set again when a page
+ * PG_active is always cleared while a page is on one of lrugen->folios[] so
+ * that the aging needs not to worry about it. And it's set again when a page
* considered active is isolated for non-reclaiming purposes, e.g., migration.
* See lru_gen_add_folio() and lru_gen_del_folio().
*
@@ -412,7 +412,7 @@ struct lru_gen_folio {
/* the birth time of each generation in jiffies */
unsigned long timestamps[MAX_NR_GENS];
/* the multi-gen LRU lists, lazily sorted on eviction */
- struct list_head lists[MAX_NR_GENS][ANON_AND_FILE][MAX_NR_ZONES];
+ struct list_head folios[MAX_NR_GENS][ANON_AND_FILE][MAX_NR_ZONES];
/* the multi-gen LRU sizes, eventually consistent */
long nr_pages[MAX_NR_GENS][ANON_AND_FILE][MAX_NR_ZONES];
/* the exponential moving average of refaulted */
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -4261,7 +4261,7 @@ static bool inc_min_seq(struct lruvec *l
/* prevent cold/hot inversion if force_scan is true */
for (zone = 0; zone < MAX_NR_ZONES; zone++) {
- struct list_head *head = &lrugen->lists[old_gen][type][zone];
+ struct list_head *head = &lrugen->folios[old_gen][type][zone];
while (!list_empty(head)) {
struct folio *folio = lru_to_folio(head);
@@ -4272,7 +4272,7 @@ static bool inc_min_seq(struct lruvec *l
VM_WARN_ON_ONCE_FOLIO(folio_zonenum(folio) != zone, folio);
new_gen = folio_inc_gen(lruvec, folio, false);
- list_move_tail(&folio->lru, &lrugen->lists[new_gen][type][zone]);
+ list_move_tail(&folio->lru, &lrugen->folios[new_gen][type][zone]);
if (!--remaining)
return false;
@@ -4300,7 +4300,7 @@ static bool try_to_inc_min_seq(struct lr
gen = lru_gen_from_seq(min_seq[type]);
for (zone = 0; zone < MAX_NR_ZONES; zone++) {
- if (!list_empty(&lrugen->lists[gen][type][zone]))
+ if (!list_empty(&lrugen->folios[gen][type][zone]))
goto next;
}
@@ -4765,7 +4765,7 @@ static bool sort_folio(struct lruvec *lr
/* promoted */
if (gen != lru_gen_from_seq(lrugen->min_seq[type])) {
- list_move(&folio->lru, &lrugen->lists[gen][type][zone]);
+ list_move(&folio->lru, &lrugen->folios[gen][type][zone]);
return true;
}
@@ -4774,7 +4774,7 @@ static bool sort_folio(struct lruvec *lr
int hist = lru_hist_from_seq(lrugen->min_seq[type]);
gen = folio_inc_gen(lruvec, folio, false);
- list_move_tail(&folio->lru, &lrugen->lists[gen][type][zone]);
+ list_move_tail(&folio->lru, &lrugen->folios[gen][type][zone]);
WRITE_ONCE(lrugen->protected[hist][type][tier - 1],
lrugen->protected[hist][type][tier - 1] + delta);
@@ -4786,7 +4786,7 @@ static bool sort_folio(struct lruvec *lr
if (folio_test_locked(folio) || folio_test_writeback(folio) ||
(type == LRU_GEN_FILE && folio_test_dirty(folio))) {
gen = folio_inc_gen(lruvec, folio, true);
- list_move(&folio->lru, &lrugen->lists[gen][type][zone]);
+ list_move(&folio->lru, &lrugen->folios[gen][type][zone]);
return true;
}
@@ -4853,7 +4853,7 @@ static int scan_folios(struct lruvec *lr
for (zone = sc->reclaim_idx; zone >= 0; zone--) {
LIST_HEAD(moved);
int skipped = 0;
- struct list_head *head = &lrugen->lists[gen][type][zone];
+ struct list_head *head = &lrugen->folios[gen][type][zone];
while (!list_empty(head)) {
struct folio *folio = lru_to_folio(head);
@@ -5253,7 +5253,7 @@ static bool __maybe_unused state_is_vali
int gen, type, zone;
for_each_gen_type_zone(gen, type, zone) {
- if (!list_empty(&lrugen->lists[gen][type][zone]))
+ if (!list_empty(&lrugen->folios[gen][type][zone]))
return false;
}
}
@@ -5298,7 +5298,7 @@ static bool drain_evictable(struct lruve
int remaining = MAX_LRU_BATCH;
for_each_gen_type_zone(gen, type, zone) {
- struct list_head *head = &lruvec->lrugen.lists[gen][type][zone];
+ struct list_head *head = &lruvec->lrugen.folios[gen][type][zone];
while (!list_empty(head)) {
bool success;
@@ -5835,7 +5835,7 @@ void lru_gen_init_lruvec(struct lruvec *
lrugen->timestamps[i] = jiffies;
for_each_gen_type_zone(gen, type, zone)
- INIT_LIST_HEAD(&lrugen->lists[gen][type][zone]);
+ INIT_LIST_HEAD(&lrugen->folios[gen][type][zone]);
lruvec->mm_state.seq = MIN_NR_GENS;
init_waitqueue_head(&lruvec->mm_state.wait);

187
target/linux/generic/backport-6.1/020-v6.3-23-mm-multi-gen-LRU-remove-eviction-fairness-safeguard.patch Normal file
View File

@ -0,0 +1,187 @@
From ce45f1c4b32cf69b166f56ef5bc6c761e06ed4e5 Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:19:01 -0700
Subject: [PATCH 23/29] mm: multi-gen LRU: remove eviction fairness safeguard
Recall that the eviction consumes the oldest generation: first it
bucket-sorts folios whose gen counters were updated by the aging and
reclaims the rest; then it increments lrugen->min_seq.
The current eviction fairness safeguard for global reclaim has a
dilemma: when there are multiple eligible memcgs, should it continue
or stop upon meeting the reclaim goal? If it continues, it overshoots
and increases direct reclaim latency; if it stops, it loses fairness
between memcgs it has taken memory away from and those it has yet to.
With memcg LRU, the eviction, while ensuring eventual fairness, will
stop upon meeting its goal. Therefore the current eviction fairness
safeguard for global reclaim will not be needed.
Note that memcg LRU only applies to global reclaim. For memcg reclaim,
the eviction will continue, even if it is overshooting. This becomes
unconditional due to code simplification.
Link: https://lkml.kernel.org/r/20221222041905.2431096-4-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/vmscan.c | 82 +++++++++++++++--------------------------------------
1 file changed, 23 insertions(+), 59 deletions(-)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -448,6 +448,11 @@ static bool cgroup_reclaim(struct scan_c
return sc->target_mem_cgroup;
}
+static bool global_reclaim(struct scan_control *sc)
+{
+ return !sc->target_mem_cgroup || mem_cgroup_is_root(sc->target_mem_cgroup);
+}
+
/**
* writeback_throttling_sane - is the usual dirty throttling mechanism available?
* @sc: scan_control in question
@@ -498,6 +503,11 @@ static bool cgroup_reclaim(struct scan_c
return false;
}
+static bool global_reclaim(struct scan_control *sc)
+{
+ return true;
+}
+
static bool writeback_throttling_sane(struct scan_control *sc)
{
return true;
@@ -4996,8 +5006,7 @@ static int isolate_folios(struct lruvec
return scanned;
}
-static int evict_folios(struct lruvec *lruvec, struct scan_control *sc, int swappiness,
- bool *need_swapping)
+static int evict_folios(struct lruvec *lruvec, struct scan_control *sc, int swappiness)
{
int type;
int scanned;
@@ -5086,9 +5095,6 @@ retry:
goto retry;
}
- if (need_swapping && type == LRU_GEN_ANON)
- *need_swapping = true;
-
return scanned;
}
@@ -5127,67 +5133,26 @@ done:
return min_seq[!can_swap] + MIN_NR_GENS <= max_seq ? nr_to_scan : 0;
}
-static bool should_abort_scan(struct lruvec *lruvec, unsigned long seq,
- struct scan_control *sc, bool need_swapping)
+static unsigned long get_nr_to_reclaim(struct scan_control *sc)
{
- int i;
- DEFINE_MAX_SEQ(lruvec);
-
- if (!current_is_kswapd()) {
- /* age each memcg at most once to ensure fairness */
- if (max_seq - seq > 1)
- return true;
-
- /* over-swapping can increase allocation latency */
- if (sc->nr_reclaimed >= sc->nr_to_reclaim && need_swapping)
- return true;
-
- /* give this thread a chance to exit and free its memory */
- if (fatal_signal_pending(current)) {
- sc->nr_reclaimed += MIN_LRU_BATCH;
- return true;
- }
-
- if (cgroup_reclaim(sc))
- return false;
- } else if (sc->nr_reclaimed - sc->last_reclaimed < sc->nr_to_reclaim)
- return false;
-
- /* keep scanning at low priorities to ensure fairness */
- if (sc->priority > DEF_PRIORITY - 2)
- return false;
-
- /*
- * A minimum amount of work was done under global memory pressure. For
- * kswapd, it may be overshooting. For direct reclaim, the allocation
- * may succeed if all suitable zones are somewhat safe. In either case,
- * it's better to stop now, and restart later if necessary.
- */
- for (i = 0; i <= sc->reclaim_idx; i++) {
- unsigned long wmark;
- struct zone *zone = lruvec_pgdat(lruvec)->node_zones + i;
-
- if (!managed_zone(zone))
- continue;
-
- wmark = current_is_kswapd() ? high_wmark_pages(zone) : low_wmark_pages(zone);
- if (wmark > zone_page_state(zone, NR_FREE_PAGES))
- return false;
- }
+ /* don't abort memcg reclaim to ensure fairness */
+ if (!global_reclaim(sc))
+ return -1;
- sc->nr_reclaimed += MIN_LRU_BATCH;
+ /* discount the previous progress for kswapd */
+ if (current_is_kswapd())
+ return sc->nr_to_reclaim + sc->last_reclaimed;
- return true;
+ return max(sc->nr_to_reclaim, compact_gap(sc->order));
}
static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
{
struct blk_plug plug;
bool need_aging = false;
- bool need_swapping = false;
unsigned long scanned = 0;
unsigned long reclaimed = sc->nr_reclaimed;
- DEFINE_MAX_SEQ(lruvec);
+ unsigned long nr_to_reclaim = get_nr_to_reclaim(sc);
lru_add_drain();
@@ -5211,7 +5176,7 @@ static void lru_gen_shrink_lruvec(struct
if (!nr_to_scan)
goto done;
- delta = evict_folios(lruvec, sc, swappiness, &need_swapping);
+ delta = evict_folios(lruvec, sc, swappiness);
if (!delta)
goto done;
@@ -5219,7 +5184,7 @@ static void lru_gen_shrink_lruvec(struct
if (scanned >= nr_to_scan)
break;
- if (should_abort_scan(lruvec, max_seq, sc, need_swapping))
+ if (sc->nr_reclaimed >= nr_to_reclaim)
break;
cond_resched();
@@ -5669,7 +5634,7 @@ static int run_eviction(struct lruvec *l
if (sc->nr_reclaimed >= nr_to_reclaim)
return 0;
- if (!evict_folios(lruvec, sc, swappiness, NULL))
+ if (!evict_folios(lruvec, sc, swappiness))
return 0;
cond_resched();

287
target/linux/generic/backport-6.1/020-v6.3-24-mm-multi-gen-LRU-remove-aging-fairness-safeguard.patch Normal file
View File

@ -0,0 +1,287 @@
From e20b7386fccc18c791796eb1dc1a91eee3ccf801 Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:19:02 -0700
Subject: [PATCH 24/29] mm: multi-gen LRU: remove aging fairness safeguard
Recall that the aging produces the youngest generation: first it scans
for accessed folios and updates their gen counters; then it increments
lrugen->max_seq.
The current aging fairness safeguard for kswapd uses two passes to
ensure the fairness to multiple eligible memcgs. On the first pass,
which is shared with the eviction, it checks whether all eligible
memcgs are low on cold folios. If so, it requires a second pass, on
which it ages all those memcgs at the same time.
With memcg LRU, the aging, while ensuring eventual fairness, will run
when necessary. Therefore the current aging fairness safeguard for
kswapd will not be needed.
Note that memcg LRU only applies to global reclaim. For memcg reclaim,
the aging can be unfair to different memcgs, i.e., their
lrugen->max_seq can be incremented at different paces.
Link: https://lkml.kernel.org/r/20221222041905.2431096-5-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/vmscan.c | 126 ++++++++++++++++++++++++----------------------------
1 file changed, 59 insertions(+), 67 deletions(-)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -136,7 +136,6 @@ struct scan_control {
#ifdef CONFIG_LRU_GEN
/* help kswapd make better choices among multiple memcgs */
- unsigned int memcgs_need_aging:1;
unsigned long last_reclaimed;
#endif
@@ -4458,7 +4457,7 @@ done:
return true;
}
-static bool should_run_aging(struct lruvec *lruvec, unsigned long max_seq, unsigned long *min_seq,
+static bool should_run_aging(struct lruvec *lruvec, unsigned long max_seq,
struct scan_control *sc, bool can_swap, unsigned long *nr_to_scan)
{
int gen, type, zone;
@@ -4467,6 +4466,13 @@ static bool should_run_aging(struct lruv
unsigned long total = 0;
struct lru_gen_folio *lrugen = &lruvec->lrugen;
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
+ DEFINE_MIN_SEQ(lruvec);
+
+ /* whether this lruvec is completely out of cold folios */
+ if (min_seq[!can_swap] + MIN_NR_GENS > max_seq) {
+ *nr_to_scan = 0;
+ return true;
+ }
for (type = !can_swap; type < ANON_AND_FILE; type++) {
unsigned long seq;
@@ -4495,8 +4501,6 @@ static bool should_run_aging(struct lruv
* stalls when the number of generations reaches MIN_NR_GENS. Hence, the
* ideal number of generations is MIN_NR_GENS+1.
*/
- if (min_seq[!can_swap] + MIN_NR_GENS > max_seq)
- return true;
if (min_seq[!can_swap] + MIN_NR_GENS < max_seq)
return false;
@@ -4515,40 +4519,54 @@ static bool should_run_aging(struct lruv
return false;
}
-static bool age_lruvec(struct lruvec *lruvec, struct scan_control *sc, unsigned long min_ttl)
+static bool lruvec_is_sizable(struct lruvec *lruvec, struct scan_control *sc)
{
- bool need_aging;
- unsigned long nr_to_scan;
- int swappiness = get_swappiness(lruvec, sc);
+ int gen, type, zone;
+ unsigned long total = 0;
+ bool can_swap = get_swappiness(lruvec, sc);
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
DEFINE_MAX_SEQ(lruvec);
DEFINE_MIN_SEQ(lruvec);
- VM_WARN_ON_ONCE(sc->memcg_low_reclaim);
+ for (type = !can_swap; type < ANON_AND_FILE; type++) {
+ unsigned long seq;
- mem_cgroup_calculate_protection(NULL, memcg);
+ for (seq = min_seq[type]; seq <= max_seq; seq++) {
+ gen = lru_gen_from_seq(seq);
- if (mem_cgroup_below_min(memcg))
- return false;
+ for (zone = 0; zone < MAX_NR_ZONES; zone++)
+ total += max(READ_ONCE(lrugen->nr_pages[gen][type][zone]), 0L);
+ }
+ }
- need_aging = should_run_aging(lruvec, max_seq, min_seq, sc, swappiness, &nr_to_scan);
+ /* whether the size is big enough to be helpful */
+ return mem_cgroup_online(memcg) ? (total >> sc->priority) : total;
+}
- if (min_ttl) {
- int gen = lru_gen_from_seq(min_seq[LRU_GEN_FILE]);
- unsigned long birth = READ_ONCE(lruvec->lrugen.timestamps[gen]);
+static bool lruvec_is_reclaimable(struct lruvec *lruvec, struct scan_control *sc,
+ unsigned long min_ttl)
+{
+ int gen;
+ unsigned long birth;
+ struct mem_cgroup *memcg = lruvec_memcg(lruvec);
+ DEFINE_MIN_SEQ(lruvec);
- if (time_is_after_jiffies(birth + min_ttl))
- return false;
+ VM_WARN_ON_ONCE(sc->memcg_low_reclaim);
- /* the size is likely too small to be helpful */
- if (!nr_to_scan && sc->priority != DEF_PRIORITY)
- return false;
- }
+ /* see the comment on lru_gen_folio */
+ gen = lru_gen_from_seq(min_seq[LRU_GEN_FILE]);
+ birth = READ_ONCE(lruvec->lrugen.timestamps[gen]);
- if (need_aging)
- try_to_inc_max_seq(lruvec, max_seq, sc, swappiness, false);
+ if (time_is_after_jiffies(birth + min_ttl))
+ return false;
- return true;
+ if (!lruvec_is_sizable(lruvec, sc))
+ return false;
+
+ mem_cgroup_calculate_protection(NULL, memcg);
+
+ return !mem_cgroup_below_min(memcg);
}
/* to protect the working set of the last N jiffies */
@@ -4557,46 +4575,32 @@ static unsigned long lru_gen_min_ttl __r
static void lru_gen_age_node(struct pglist_data *pgdat, struct scan_control *sc)
{
struct mem_cgroup *memcg;
- bool success = false;
unsigned long min_ttl = READ_ONCE(lru_gen_min_ttl);
VM_WARN_ON_ONCE(!current_is_kswapd());
sc->last_reclaimed = sc->nr_reclaimed;
- /*
- * To reduce the chance of going into the aging path, which can be
- * costly, optimistically skip it if the flag below was cleared in the
- * eviction path. This improves the overall performance when multiple
- * memcgs are available.
- */
- if (!sc->memcgs_need_aging) {
- sc->memcgs_need_aging = true;
+ /* check the order to exclude compaction-induced reclaim */
+ if (!min_ttl || sc->order || sc->priority == DEF_PRIORITY)
return;
- }
-
- set_mm_walk(pgdat);
memcg = mem_cgroup_iter(NULL, NULL, NULL);
do {
struct lruvec *lruvec = mem_cgroup_lruvec(memcg, pgdat);
- if (age_lruvec(lruvec, sc, min_ttl))
- success = true;
+ if (lruvec_is_reclaimable(lruvec, sc, min_ttl)) {
+ mem_cgroup_iter_break(NULL, memcg);
+ return;
+ }
cond_resched();
} while ((memcg = mem_cgroup_iter(NULL, memcg, NULL)));
- clear_mm_walk();
-
- /* check the order to exclude compaction-induced reclaim */
- if (success || !min_ttl || sc->order)
- return;
-
/*
* The main goal is to OOM kill if every generation from all memcgs is
* younger than min_ttl. However, another possibility is all memcgs are
- * either below min or empty.
+ * either too small or below min.
*/
if (mutex_trylock(&oom_lock)) {
struct oom_control oc = {
@@ -5104,33 +5108,27 @@ retry:
* reclaim.
*/
static unsigned long get_nr_to_scan(struct lruvec *lruvec, struct scan_control *sc,
- bool can_swap, bool *need_aging)
+ bool can_swap)
{
unsigned long nr_to_scan;
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
DEFINE_MAX_SEQ(lruvec);
- DEFINE_MIN_SEQ(lruvec);
if (mem_cgroup_below_min(memcg) ||
(mem_cgroup_below_low(memcg) && !sc->memcg_low_reclaim))
return 0;
- *need_aging = should_run_aging(lruvec, max_seq, min_seq, sc, can_swap, &nr_to_scan);
- if (!*need_aging)
+ if (!should_run_aging(lruvec, max_seq, sc, can_swap, &nr_to_scan))
return nr_to_scan;
/* skip the aging path at the default priority */
if (sc->priority == DEF_PRIORITY)
- goto done;
+ return nr_to_scan;
- /* leave the work to lru_gen_age_node() */
- if (current_is_kswapd())
- return 0;
+ try_to_inc_max_seq(lruvec, max_seq, sc, can_swap, false);
- if (try_to_inc_max_seq(lruvec, max_seq, sc, can_swap, false))
- return nr_to_scan;
-done:
- return min_seq[!can_swap] + MIN_NR_GENS <= max_seq ? nr_to_scan : 0;
+ /* skip this lruvec as it's low on cold folios */
+ return 0;
}
static unsigned long get_nr_to_reclaim(struct scan_control *sc)
@@ -5149,9 +5147,7 @@ static unsigned long get_nr_to_reclaim(s
static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
{
struct blk_plug plug;
- bool need_aging = false;
unsigned long scanned = 0;
- unsigned long reclaimed = sc->nr_reclaimed;
unsigned long nr_to_reclaim = get_nr_to_reclaim(sc);
lru_add_drain();
@@ -5172,13 +5168,13 @@ static void lru_gen_shrink_lruvec(struct
else
swappiness = 0;
- nr_to_scan = get_nr_to_scan(lruvec, sc, swappiness, &need_aging);
+ nr_to_scan = get_nr_to_scan(lruvec, sc, swappiness);
if (!nr_to_scan)
- goto done;
+ break;
delta = evict_folios(lruvec, sc, swappiness);
if (!delta)
- goto done;
+ break;
scanned += delta;
if (scanned >= nr_to_scan)
@@ -5190,10 +5186,6 @@ static void lru_gen_shrink_lruvec(struct
cond_resched();
}
- /* see the comment in lru_gen_age_node() */
- if (sc->nr_reclaimed - reclaimed >= MIN_LRU_BATCH && !need_aging)
- sc->memcgs_need_aging = false;
-done:
clear_mm_walk();
blk_finish_plug(&plug);

161
target/linux/generic/backport-6.1/020-v6.3-25-mm-multi-gen-LRU-shuffle-should_run_aging.patch Normal file
View File

@ -0,0 +1,161 @@
From 107d54931df3c28d81648122e219bf0034ef4e99 Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:19:03 -0700
Subject: [PATCH 25/29] mm: multi-gen LRU: shuffle should_run_aging()
Move should_run_aging() next to its only caller left.
Link: https://lkml.kernel.org/r/20221222041905.2431096-6-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/vmscan.c | 124 ++++++++++++++++++++++++++--------------------------
1 file changed, 62 insertions(+), 62 deletions(-)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -4457,68 +4457,6 @@ done:
return true;
}
-static bool should_run_aging(struct lruvec *lruvec, unsigned long max_seq,
- struct scan_control *sc, bool can_swap, unsigned long *nr_to_scan)
-{
- int gen, type, zone;
- unsigned long old = 0;
- unsigned long young = 0;
- unsigned long total = 0;
- struct lru_gen_folio *lrugen = &lruvec->lrugen;
- struct mem_cgroup *memcg = lruvec_memcg(lruvec);
- DEFINE_MIN_SEQ(lruvec);
-
- /* whether this lruvec is completely out of cold folios */
- if (min_seq[!can_swap] + MIN_NR_GENS > max_seq) {
- *nr_to_scan = 0;
- return true;
- }
-
- for (type = !can_swap; type < ANON_AND_FILE; type++) {
- unsigned long seq;
-
- for (seq = min_seq[type]; seq <= max_seq; seq++) {
- unsigned long size = 0;
-
- gen = lru_gen_from_seq(seq);
-
- for (zone = 0; zone < MAX_NR_ZONES; zone++)
- size += max(READ_ONCE(lrugen->nr_pages[gen][type][zone]), 0L);
-
- total += size;
- if (seq == max_seq)
- young += size;
- else if (seq + MIN_NR_GENS == max_seq)
- old += size;
- }
- }
-
- /* try to scrape all its memory if this memcg was deleted */
- *nr_to_scan = mem_cgroup_online(memcg) ? (total >> sc->priority) : total;
-
- /*
- * The aging tries to be lazy to reduce the overhead, while the eviction
- * stalls when the number of generations reaches MIN_NR_GENS. Hence, the
- * ideal number of generations is MIN_NR_GENS+1.
- */
- if (min_seq[!can_swap] + MIN_NR_GENS < max_seq)
- return false;
-
- /*
- * It's also ideal to spread pages out evenly, i.e., 1/(MIN_NR_GENS+1)
- * of the total number of pages for each generation. A reasonable range
- * for this average portion is [1/MIN_NR_GENS, 1/(MIN_NR_GENS+2)]. The
- * aging cares about the upper bound of hot pages, while the eviction
- * cares about the lower bound of cold pages.
- */
- if (young * MIN_NR_GENS > total)
- return true;
- if (old * (MIN_NR_GENS + 2) < total)
- return true;
-
- return false;
-}
-
static bool lruvec_is_sizable(struct lruvec *lruvec, struct scan_control *sc)
{
int gen, type, zone;
@@ -5102,6 +5040,68 @@ retry:
return scanned;
}
+static bool should_run_aging(struct lruvec *lruvec, unsigned long max_seq,
+ struct scan_control *sc, bool can_swap, unsigned long *nr_to_scan)
+{
+ int gen, type, zone;
+ unsigned long old = 0;
+ unsigned long young = 0;
+ unsigned long total = 0;
+ struct lru_gen_folio *lrugen = &lruvec->lrugen;
+ struct mem_cgroup *memcg = lruvec_memcg(lruvec);
+ DEFINE_MIN_SEQ(lruvec);
+
+ /* whether this lruvec is completely out of cold folios */
+ if (min_seq[!can_swap] + MIN_NR_GENS > max_seq) {
+ *nr_to_scan = 0;
+ return true;
+ }
+
+ for (type = !can_swap; type < ANON_AND_FILE; type++) {
+ unsigned long seq;
+
+ for (seq = min_seq[type]; seq <= max_seq; seq++) {
+ unsigned long size = 0;
+
+ gen = lru_gen_from_seq(seq);
+
+ for (zone = 0; zone < MAX_NR_ZONES; zone++)
+ size += max(READ_ONCE(lrugen->nr_pages[gen][type][zone]), 0L);
+
+ total += size;
+ if (seq == max_seq)
+ young += size;
+ else if (seq + MIN_NR_GENS == max_seq)
+ old += size;
+ }
+ }
+
+ /* try to scrape all its memory if this memcg was deleted */
+ *nr_to_scan = mem_cgroup_online(memcg) ? (total >> sc->priority) : total;
+
+ /*
+ * The aging tries to be lazy to reduce the overhead, while the eviction
+ * stalls when the number of generations reaches MIN_NR_GENS. Hence, the
+ * ideal number of generations is MIN_NR_GENS+1.
+ */
+ if (min_seq[!can_swap] + MIN_NR_GENS < max_seq)
+ return false;
+
+ /*
+ * It's also ideal to spread pages out evenly, i.e., 1/(MIN_NR_GENS+1)
+ * of the total number of pages for each generation. A reasonable range
+ * for this average portion is [1/MIN_NR_GENS, 1/(MIN_NR_GENS+2)]. The
+ * aging cares about the upper bound of hot pages, while the eviction
+ * cares about the lower bound of cold pages.
+ */
+ if (young * MIN_NR_GENS > total)
+ return true;
+ if (old * (MIN_NR_GENS + 2) < total)
+ return true;
+
+ return false;
+}
+
/*
* For future optimizations:
* 1. Defer try_to_inc_max_seq() to workqueues to reduce latency for memcg

868
target/linux/generic/backport-6.1/020-v6.3-26-mm-multi-gen-LRU-per-node-lru_gen_page-lists.patch Normal file
View File

@ -0,0 +1,868 @@
From fa6363828d314e837c5f79e97ea5e8c0d2f7f062 Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:19:04 -0700
Subject: [PATCH 26/29] mm: multi-gen LRU: per-node lru_gen_folio lists
For each node, memcgs are divided into two generations: the old and
the young. For each generation, memcgs are randomly sharded into
multiple bins to improve scalability. For each bin, an RCU hlist_nulls
is virtually divided into three segments: the head, the tail and the
default.
An onlining memcg is added to the tail of a random bin in the old
generation. The eviction starts at the head of a random bin in the old
generation. The per-node memcg generation counter, whose reminder (mod
2) indexes the old generation, is incremented when all its bins become
empty.
There are four operations:
1. MEMCG_LRU_HEAD, which moves an memcg to the head of a random bin in
its current generation (old or young) and updates its "seg" to
"head";
2. MEMCG_LRU_TAIL, which moves an memcg to the tail of a random bin in
its current generation (old or young) and updates its "seg" to
"tail";
3. MEMCG_LRU_OLD, which moves an memcg to the head of a random bin in
the old generation, updates its "gen" to "old" and resets its "seg"
to "default";
4. MEMCG_LRU_YOUNG, which moves an memcg to the tail of a random bin
in the young generation, updates its "gen" to "young" and resets
its "seg" to "default".
The events that trigger the above operations are:
1. Exceeding the soft limit, which triggers MEMCG_LRU_HEAD;
2. The first attempt to reclaim an memcg below low, which triggers
MEMCG_LRU_TAIL;
3. The first attempt to reclaim an memcg below reclaimable size
threshold, which triggers MEMCG_LRU_TAIL;
4. The second attempt to reclaim an memcg below reclaimable size
threshold, which triggers MEMCG_LRU_YOUNG;
5. Attempting to reclaim an memcg below min, which triggers
MEMCG_LRU_YOUNG;
6. Finishing the aging on the eviction path, which triggers
MEMCG_LRU_YOUNG;
7. Offlining an memcg, which triggers MEMCG_LRU_OLD.
Note that memcg LRU only applies to global reclaim, and the
round-robin incrementing of their max_seq counters ensures the
eventual fairness to all eligible memcgs. For memcg reclaim, it still
relies on mem_cgroup_iter().
Link: https://lkml.kernel.org/r/20221222041905.2431096-7-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
include/linux/memcontrol.h | 10 +
include/linux/mm_inline.h | 17 ++
include/linux/mmzone.h | 117 +++++++++++-
mm/memcontrol.c | 16 ++
mm/folio_alloc.c | 1 +
mm/vmscan.c | 373 +++++++++++++++++++++++++++++++++----
6 files changed, 499 insertions(+), 35 deletions(-)
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -790,6 +790,11 @@ static inline void obj_cgroup_put(struct
percpu_ref_put(&objcg->refcnt);
}
+static inline bool mem_cgroup_tryget(struct mem_cgroup *memcg)
+{
+ return !memcg || css_tryget(&memcg->css);
+}
+
static inline void mem_cgroup_put(struct mem_cgroup *memcg)
{
if (memcg)
@@ -1290,6 +1295,11 @@ static inline void obj_cgroup_put(struct
{
}
+static inline bool mem_cgroup_tryget(struct mem_cgroup *memcg)
+{
+ return true;
+}
+
static inline void mem_cgroup_put(struct mem_cgroup *memcg)
{
}
--- a/include/linux/mm_inline.h
+++ b/include/linux/mm_inline.h
@@ -122,6 +122,18 @@ static inline bool lru_gen_in_fault(void
return current->in_lru_fault;
}
+#ifdef CONFIG_MEMCG
+static inline int lru_gen_memcg_seg(struct lruvec *lruvec)
+{
+ return READ_ONCE(lruvec->lrugen.seg);
+}
+#else
+static inline int lru_gen_memcg_seg(struct lruvec *lruvec)
+{
+ return 0;
+}
+#endif
+
static inline int lru_gen_from_seq(unsigned long seq)
{
return seq % MAX_NR_GENS;
@@ -297,6 +309,11 @@ static inline bool lru_gen_in_fault(void
return false;
}
+static inline int lru_gen_memcg_seg(struct lruvec *lruvec)
+{
+ return 0;
+}
+
static inline bool lru_gen_add_folio(struct lruvec *lruvec, struct folio *folio, bool reclaiming)
{
return false;
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -7,6 +7,7 @@
#include <linux/spinlock.h>
#include <linux/list.h>
+#include <linux/list_nulls.h>
#include <linux/wait.h>
#include <linux/bitops.h>
#include <linux/cache.h>
@@ -367,6 +368,15 @@ struct page_vma_mapped_walk;
#define LRU_GEN_MASK ((BIT(LRU_GEN_WIDTH) - 1) << LRU_GEN_PGOFF)
#define LRU_REFS_MASK ((BIT(LRU_REFS_WIDTH) - 1) << LRU_REFS_PGOFF)
+/* see the comment on MEMCG_NR_GENS */
+enum {
+ MEMCG_LRU_NOP,
+ MEMCG_LRU_HEAD,
+ MEMCG_LRU_TAIL,
+ MEMCG_LRU_OLD,
+ MEMCG_LRU_YOUNG,
+};
+
#ifdef CONFIG_LRU_GEN
enum {
@@ -426,6 +436,14 @@ struct lru_gen_folio {
atomic_long_t refaulted[NR_HIST_GENS][ANON_AND_FILE][MAX_NR_TIERS];
/* whether the multi-gen LRU is enabled */
bool enabled;
+#ifdef CONFIG_MEMCG
+ /* the memcg generation this lru_gen_folio belongs to */
+ u8 gen;
+ /* the list segment this lru_gen_folio belongs to */
+ u8 seg;
+ /* per-node lru_gen_folio list for global reclaim */
+ struct hlist_nulls_node list;
+#endif
};
enum {
@@ -479,12 +497,87 @@ void lru_gen_init_lruvec(struct lruvec *
void lru_gen_look_around(struct page_vma_mapped_walk *pvmw);
#ifdef CONFIG_MEMCG
+
+/*
+ * For each node, memcgs are divided into two generations: the old and the
+ * young. For each generation, memcgs are randomly sharded into multiple bins
+ * to improve scalability. For each bin, the hlist_nulls is virtually divided
+ * into three segments: the head, the tail and the default.
+ *
+ * An onlining memcg is added to the tail of a random bin in the old generation.
+ * The eviction starts at the head of a random bin in the old generation. The
+ * per-node memcg generation counter, whose reminder (mod MEMCG_NR_GENS) indexes
+ * the old generation, is incremented when all its bins become empty.
+ *
+ * There are four operations:
+ * 1. MEMCG_LRU_HEAD, which moves an memcg to the head of a random bin in its
+ * current generation (old or young) and updates its "seg" to "head";
+ * 2. MEMCG_LRU_TAIL, which moves an memcg to the tail of a random bin in its
+ * current generation (old or young) and updates its "seg" to "tail";
+ * 3. MEMCG_LRU_OLD, which moves an memcg to the head of a random bin in the old
+ * generation, updates its "gen" to "old" and resets its "seg" to "default";
+ * 4. MEMCG_LRU_YOUNG, which moves an memcg to the tail of a random bin in the
+ * young generation, updates its "gen" to "young" and resets its "seg" to
+ * "default".
+ *
+ * The events that trigger the above operations are:
+ * 1. Exceeding the soft limit, which triggers MEMCG_LRU_HEAD;
+ * 2. The first attempt to reclaim an memcg below low, which triggers
+ * MEMCG_LRU_TAIL;
+ * 3. The first attempt to reclaim an memcg below reclaimable size threshold,
+ * which triggers MEMCG_LRU_TAIL;
+ * 4. The second attempt to reclaim an memcg below reclaimable size threshold,
+ * which triggers MEMCG_LRU_YOUNG;
+ * 5. Attempting to reclaim an memcg below min, which triggers MEMCG_LRU_YOUNG;
+ * 6. Finishing the aging on the eviction path, which triggers MEMCG_LRU_YOUNG;
+ * 7. Offlining an memcg, which triggers MEMCG_LRU_OLD.
+ *
+ * Note that memcg LRU only applies to global reclaim, and the round-robin
+ * incrementing of their max_seq counters ensures the eventual fairness to all
+ * eligible memcgs. For memcg reclaim, it still relies on mem_cgroup_iter().
+ */
+#define MEMCG_NR_GENS 2
+#define MEMCG_NR_BINS 8
+
+struct lru_gen_memcg {
+ /* the per-node memcg generation counter */
+ unsigned long seq;
+ /* each memcg has one lru_gen_folio per node */
+ unsigned long nr_memcgs[MEMCG_NR_GENS];
+ /* per-node lru_gen_folio list for global reclaim */
+ struct hlist_nulls_head fifo[MEMCG_NR_GENS][MEMCG_NR_BINS];
+ /* protects the above */
+ spinlock_t lock;
+};
+
+void lru_gen_init_pgdat(struct pglist_data *pgdat);
+
void lru_gen_init_memcg(struct mem_cgroup *memcg);
void lru_gen_exit_memcg(struct mem_cgroup *memcg);
-#endif
+void lru_gen_online_memcg(struct mem_cgroup *memcg);
+void lru_gen_offline_memcg(struct mem_cgroup *memcg);
+void lru_gen_release_memcg(struct mem_cgroup *memcg);
+void lru_gen_rotate_memcg(struct lruvec *lruvec, int op);
+
+#else /* !CONFIG_MEMCG */
+
+#define MEMCG_NR_GENS 1
+
+struct lru_gen_memcg {
+};
+
+static inline void lru_gen_init_pgdat(struct pglist_data *pgdat)
+{
+}
+
+#endif /* CONFIG_MEMCG */
#else /* !CONFIG_LRU_GEN */
+static inline void lru_gen_init_pgdat(struct pglist_data *pgdat)
+{
+}
+
static inline void lru_gen_init_lruvec(struct lruvec *lruvec)
{
}
@@ -494,6 +587,7 @@ static inline void lru_gen_look_around(s
}
#ifdef CONFIG_MEMCG
+
static inline void lru_gen_init_memcg(struct mem_cgroup *memcg)
{
}
@@ -501,7 +595,24 @@ static inline void lru_gen_init_memcg(st
static inline void lru_gen_exit_memcg(struct mem_cgroup *memcg)
{
}
-#endif
+
+static inline void lru_gen_online_memcg(struct mem_cgroup *memcg)
+{
+}
+
+static inline void lru_gen_offline_memcg(struct mem_cgroup *memcg)
+{
+}
+
+static inline void lru_gen_release_memcg(struct mem_cgroup *memcg)
+{
+}
+
+static inline void lru_gen_rotate_memcg(struct lruvec *lruvec, int op)
+{
+}
+
+#endif /* CONFIG_MEMCG */
#endif /* CONFIG_LRU_GEN */
@@ -1219,6 +1330,8 @@ typedef struct pglist_data {
#ifdef CONFIG_LRU_GEN
/* kswap mm walk data */
struct lru_gen_mm_walk mm_walk;
+ /* lru_gen_folio list */
+ struct lru_gen_memcg memcg_lru;
#endif
CACHELINE_PADDING(_pad2_);
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -477,6 +477,16 @@ static void mem_cgroup_update_tree(struc
struct mem_cgroup_per_node *mz;
struct mem_cgroup_tree_per_node *mctz;
+ if (lru_gen_enabled()) {
+ struct lruvec *lruvec = &memcg->nodeinfo[nid]->lruvec;
+
+ /* see the comment on MEMCG_NR_GENS */
+ if (soft_limit_excess(memcg) && lru_gen_memcg_seg(lruvec) != MEMCG_LRU_HEAD)
+ lru_gen_rotate_memcg(lruvec, MEMCG_LRU_HEAD);
+
+ return;
+ }
+
mctz = soft_limit_tree.rb_tree_per_node[nid];
if (!mctz)
return;
@@ -3522,6 +3532,9 @@ unsigned long mem_cgroup_soft_limit_recl
struct mem_cgroup_tree_per_node *mctz;
unsigned long excess;
+ if (lru_gen_enabled())
+ return 0;
+
if (order > 0)
return 0;
@@ -5382,6 +5395,7 @@ static int mem_cgroup_css_online(struct
if (unlikely(mem_cgroup_is_root(memcg)))
queue_delayed_work(system_unbound_wq, &stats_flush_dwork,
2UL*HZ);
+ lru_gen_online_memcg(memcg);
return 0;
offline_kmem:
memcg_offline_kmem(memcg);
@@ -5413,6 +5427,7 @@ static void mem_cgroup_css_offline(struc
memcg_offline_kmem(memcg);
reparent_shrinker_deferred(memcg);
wb_memcg_offline(memcg);
+ lru_gen_offline_memcg(memcg);
drain_all_stock(memcg);
@@ -5424,6 +5439,7 @@ static void mem_cgroup_css_released(stru
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
invalidate_reclaim_iterators(memcg);
+ lru_gen_release_memcg(memcg);
}
static void mem_cgroup_css_free(struct cgroup_subsys_state *css)
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -7957,6 +7957,7 @@ static void __init free_area_init_node(i
pgdat_set_deferred_range(pgdat);
free_area_init_core(pgdat);
+ lru_gen_init_pgdat(pgdat);
}
static void __init free_area_init_memoryless_node(int nid)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -54,6 +54,8 @@
#include <linux/shmem_fs.h>
#include <linux/ctype.h>
#include <linux/debugfs.h>
+#include <linux/rculist_nulls.h>
+#include <linux/random.h>
#include <asm/tlbflush.h>
#include <asm/div64.h>
@@ -134,11 +136,6 @@ struct scan_control {
/* Always discard instead of demoting to lower tier memory */
unsigned int no_demotion:1;
-#ifdef CONFIG_LRU_GEN
- /* help kswapd make better choices among multiple memcgs */
- unsigned long last_reclaimed;
-#endif
-
/* Allocation order */
s8 order;
@@ -3160,6 +3157,9 @@ DEFINE_STATIC_KEY_ARRAY_FALSE(lru_gen_ca
for ((type) = 0; (type) < ANON_AND_FILE; (type)++) \
for ((zone) = 0; (zone) < MAX_NR_ZONES; (zone)++)
+#define get_memcg_gen(seq) ((seq) % MEMCG_NR_GENS)
+#define get_memcg_bin(bin) ((bin) % MEMCG_NR_BINS)
+
static struct lruvec *get_lruvec(struct mem_cgroup *memcg, int nid)
{
struct pglist_data *pgdat = NODE_DATA(nid);
@@ -4443,8 +4443,7 @@ done:
if (sc->priority <= DEF_PRIORITY - 2)
wait_event_killable(lruvec->mm_state.wait,
max_seq < READ_ONCE(lrugen->max_seq));
-
- return max_seq < READ_ONCE(lrugen->max_seq);
+ return false;
}
VM_WARN_ON_ONCE(max_seq != READ_ONCE(lrugen->max_seq));
@@ -4517,8 +4516,6 @@ static void lru_gen_age_node(struct pgli
VM_WARN_ON_ONCE(!current_is_kswapd());
- sc->last_reclaimed = sc->nr_reclaimed;
-
/* check the order to exclude compaction-induced reclaim */
if (!min_ttl || sc->order || sc->priority == DEF_PRIORITY)
return;
@@ -5107,8 +5104,7 @@ static bool should_run_aging(struct lruv
* 1. Defer try_to_inc_max_seq() to workqueues to reduce latency for memcg
* reclaim.
*/
-static unsigned long get_nr_to_scan(struct lruvec *lruvec, struct scan_control *sc,
- bool can_swap)
+static long get_nr_to_scan(struct lruvec *lruvec, struct scan_control *sc, bool can_swap)
{
unsigned long nr_to_scan;
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
@@ -5125,10 +5121,8 @@ static unsigned long get_nr_to_scan(stru
if (sc->priority == DEF_PRIORITY)
return nr_to_scan;
- try_to_inc_max_seq(lruvec, max_seq, sc, can_swap, false);
-
/* skip this lruvec as it's low on cold folios */
- return 0;
+ return try_to_inc_max_seq(lruvec, max_seq, sc, can_swap, false) ? -1 : 0;
}
static unsigned long get_nr_to_reclaim(struct scan_control *sc)
@@ -5137,29 +5131,18 @@ static unsigned long get_nr_to_reclaim(s
if (!global_reclaim(sc))
return -1;
- /* discount the previous progress for kswapd */
- if (current_is_kswapd())
- return sc->nr_to_reclaim + sc->last_reclaimed;
-
return max(sc->nr_to_reclaim, compact_gap(sc->order));
}
-static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
+static bool try_to_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
{
- struct blk_plug plug;
+ long nr_to_scan;
unsigned long scanned = 0;
unsigned long nr_to_reclaim = get_nr_to_reclaim(sc);
- lru_add_drain();
-
- blk_start_plug(&plug);
-
- set_mm_walk(lruvec_pgdat(lruvec));
-
while (true) {
int delta;
int swappiness;
- unsigned long nr_to_scan;
if (sc->may_swap)
swappiness = get_swappiness(lruvec, sc);
@@ -5169,7 +5152,7 @@ static void lru_gen_shrink_lruvec(struct
swappiness = 0;
nr_to_scan = get_nr_to_scan(lruvec, sc, swappiness);
- if (!nr_to_scan)
+ if (nr_to_scan <= 0)
break;
delta = evict_folios(lruvec, sc, swappiness);
@@ -5186,10 +5169,250 @@ static void lru_gen_shrink_lruvec(struct
cond_resched();
}
+ /* whether try_to_inc_max_seq() was successful */
+ return nr_to_scan < 0;
+}
+
+static int shrink_one(struct lruvec *lruvec, struct scan_control *sc)
+{
+ bool success;
+ unsigned long scanned = sc->nr_scanned;
+ unsigned long reclaimed = sc->nr_reclaimed;
+ int seg = lru_gen_memcg_seg(lruvec);
+ struct mem_cgroup *memcg = lruvec_memcg(lruvec);
+ struct pglist_data *pgdat = lruvec_pgdat(lruvec);
+
+ /* see the comment on MEMCG_NR_GENS */
+ if (!lruvec_is_sizable(lruvec, sc))
+ return seg != MEMCG_LRU_TAIL ? MEMCG_LRU_TAIL : MEMCG_LRU_YOUNG;
+
+ mem_cgroup_calculate_protection(NULL, memcg);
+
+ if (mem_cgroup_below_min(memcg))
+ return MEMCG_LRU_YOUNG;
+
+ if (mem_cgroup_below_low(memcg)) {
+ /* see the comment on MEMCG_NR_GENS */
+ if (seg != MEMCG_LRU_TAIL)
+ return MEMCG_LRU_TAIL;
+
+ memcg_memory_event(memcg, MEMCG_LOW);
+ }
+
+ success = try_to_shrink_lruvec(lruvec, sc);
+
+ shrink_slab(sc->gfp_mask, pgdat->node_id, memcg, sc->priority);
+
+ vmpressure(sc->gfp_mask, memcg, false, sc->nr_scanned - scanned,
+ sc->nr_reclaimed - reclaimed);
+
+ sc->nr_reclaimed += current->reclaim_state->reclaimed_slab;
+ current->reclaim_state->reclaimed_slab = 0;
+
+ return success ? MEMCG_LRU_YOUNG : 0;
+}
+
+#ifdef CONFIG_MEMCG
+
+static void shrink_many(struct pglist_data *pgdat, struct scan_control *sc)
+{
+ int gen;
+ int bin;
+ int first_bin;
+ struct lruvec *lruvec;
+ struct lru_gen_folio *lrugen;
+ const struct hlist_nulls_node *pos;
+ int op = 0;
+ struct mem_cgroup *memcg = NULL;
+ unsigned long nr_to_reclaim = get_nr_to_reclaim(sc);
+
+ bin = first_bin = prandom_u32_max(MEMCG_NR_BINS);
+restart:
+ gen = get_memcg_gen(READ_ONCE(pgdat->memcg_lru.seq));
+
+ rcu_read_lock();
+
+ hlist_nulls_for_each_entry_rcu(lrugen, pos, &pgdat->memcg_lru.fifo[gen][bin], list) {
+ if (op)
+ lru_gen_rotate_memcg(lruvec, op);
+
+ mem_cgroup_put(memcg);
+
+ lruvec = container_of(lrugen, struct lruvec, lrugen);
+ memcg = lruvec_memcg(lruvec);
+
+ if (!mem_cgroup_tryget(memcg)) {
+ op = 0;
+ memcg = NULL;
+ continue;
+ }
+
+ rcu_read_unlock();
+
+ op = shrink_one(lruvec, sc);
+
+ if (sc->nr_reclaimed >= nr_to_reclaim)
+ goto success;
+
+ rcu_read_lock();
+ }
+
+ rcu_read_unlock();
+
+ /* restart if raced with lru_gen_rotate_memcg() */
+ if (gen != get_nulls_value(pos))
+ goto restart;
+
+ /* try the rest of the bins of the current generation */
+ bin = get_memcg_bin(bin + 1);
+ if (bin != first_bin)
+ goto restart;
+success:
+ if (op)
+ lru_gen_rotate_memcg(lruvec, op);
+
+ mem_cgroup_put(memcg);
+}
+
+static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
+{
+ struct blk_plug plug;
+
+ VM_WARN_ON_ONCE(global_reclaim(sc));
+
+ lru_add_drain();
+
+ blk_start_plug(&plug);
+
+ set_mm_walk(lruvec_pgdat(lruvec));
+
+ if (try_to_shrink_lruvec(lruvec, sc))
+ lru_gen_rotate_memcg(lruvec, MEMCG_LRU_YOUNG);
+
+ clear_mm_walk();
+
+ blk_finish_plug(&plug);
+}
+
+#else /* !CONFIG_MEMCG */
+
+static void shrink_many(struct pglist_data *pgdat, struct scan_control *sc)
+{
+ BUILD_BUG();
+}
+
+static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
+{
+ BUILD_BUG();
+}
+
+#endif
+
+static void set_initial_priority(struct pglist_data *pgdat, struct scan_control *sc)
+{
+ int priority;
+ unsigned long reclaimable;
+ struct lruvec *lruvec = mem_cgroup_lruvec(NULL, pgdat);
+
+ if (sc->priority != DEF_PRIORITY || sc->nr_to_reclaim < MIN_LRU_BATCH)
+ return;
+ /*
+ * Determine the initial priority based on ((total / MEMCG_NR_GENS) >>
+ * priority) * reclaimed_to_scanned_ratio = nr_to_reclaim, where the
+ * estimated reclaimed_to_scanned_ratio = inactive / total.
+ */
+ reclaimable = node_page_state(pgdat, NR_INACTIVE_FILE);
+ if (get_swappiness(lruvec, sc))
+ reclaimable += node_page_state(pgdat, NR_INACTIVE_ANON);
+
+ reclaimable /= MEMCG_NR_GENS;
+
+ /* round down reclaimable and round up sc->nr_to_reclaim */
+ priority = fls_long(reclaimable) - 1 - fls_long(sc->nr_to_reclaim - 1);
+
+ sc->priority = clamp(priority, 0, DEF_PRIORITY);
+}
+
+static void lru_gen_shrink_node(struct pglist_data *pgdat, struct scan_control *sc)
+{
+ struct blk_plug plug;
+ unsigned long reclaimed = sc->nr_reclaimed;
+
+ VM_WARN_ON_ONCE(!global_reclaim(sc));
+
+ lru_add_drain();
+
+ blk_start_plug(&plug);
+
+ set_mm_walk(pgdat);
+
+ set_initial_priority(pgdat, sc);
+
+ if (current_is_kswapd())
+ sc->nr_reclaimed = 0;
+
+ if (mem_cgroup_disabled())
+ shrink_one(&pgdat->__lruvec, sc);
+ else
+ shrink_many(pgdat, sc);
+
+ if (current_is_kswapd())
+ sc->nr_reclaimed += reclaimed;
+
clear_mm_walk();
blk_finish_plug(&plug);
+
+ /* kswapd should never fail */
+ pgdat->kswapd_failures = 0;
+}
+
+#ifdef CONFIG_MEMCG
+void lru_gen_rotate_memcg(struct lruvec *lruvec, int op)
+{
+ int seg;
+ int old, new;
+ int bin = prandom_u32_max(MEMCG_NR_BINS);
+ struct pglist_data *pgdat = lruvec_pgdat(lruvec);
+
+ spin_lock(&pgdat->memcg_lru.lock);
+
+ VM_WARN_ON_ONCE(hlist_nulls_unhashed(&lruvec->lrugen.list));
+
+ seg = 0;
+ new = old = lruvec->lrugen.gen;
+
+ /* see the comment on MEMCG_NR_GENS */
+ if (op == MEMCG_LRU_HEAD)
+ seg = MEMCG_LRU_HEAD;
+ else if (op == MEMCG_LRU_TAIL)
+ seg = MEMCG_LRU_TAIL;
+ else if (op == MEMCG_LRU_OLD)
+ new = get_memcg_gen(pgdat->memcg_lru.seq);
+ else if (op == MEMCG_LRU_YOUNG)
+ new = get_memcg_gen(pgdat->memcg_lru.seq + 1);
+ else
+ VM_WARN_ON_ONCE(true);
+
+ hlist_nulls_del_rcu(&lruvec->lrugen.list);
+
+ if (op == MEMCG_LRU_HEAD || op == MEMCG_LRU_OLD)
+ hlist_nulls_add_head_rcu(&lruvec->lrugen.list, &pgdat->memcg_lru.fifo[new][bin]);
+ else
+ hlist_nulls_add_tail_rcu(&lruvec->lrugen.list, &pgdat->memcg_lru.fifo[new][bin]);
+
+ pgdat->memcg_lru.nr_memcgs[old]--;
+ pgdat->memcg_lru.nr_memcgs[new]++;
+
+ lruvec->lrugen.gen = new;
+ WRITE_ONCE(lruvec->lrugen.seg, seg);
+
+ if (!pgdat->memcg_lru.nr_memcgs[old] && old == get_memcg_gen(pgdat->memcg_lru.seq))
+ WRITE_ONCE(pgdat->memcg_lru.seq, pgdat->memcg_lru.seq + 1);
+
+ spin_unlock(&pgdat->memcg_lru.lock);
}
+#endif
/******************************************************************************
* state change
@@ -5647,11 +5870,11 @@ static int run_cmd(char cmd, int memcg_i
if (!mem_cgroup_disabled()) {
rcu_read_lock();
+
memcg = mem_cgroup_from_id(memcg_id);
-#ifdef CONFIG_MEMCG
- if (memcg && !css_tryget(&memcg->css))
+ if (!mem_cgroup_tryget(memcg))
memcg = NULL;
-#endif
+
rcu_read_unlock();
if (!memcg)
@@ -5799,6 +6022,19 @@ void lru_gen_init_lruvec(struct lruvec *
}
#ifdef CONFIG_MEMCG
+
+void lru_gen_init_pgdat(struct pglist_data *pgdat)
+{
+ int i, j;
+
+ spin_lock_init(&pgdat->memcg_lru.lock);
+
+ for (i = 0; i < MEMCG_NR_GENS; i++) {
+ for (j = 0; j < MEMCG_NR_BINS; j++)
+ INIT_HLIST_NULLS_HEAD(&pgdat->memcg_lru.fifo[i][j], i);
+ }
+}
+
void lru_gen_init_memcg(struct mem_cgroup *memcg)
{
INIT_LIST_HEAD(&memcg->mm_list.fifo);
@@ -5822,7 +6058,69 @@ void lru_gen_exit_memcg(struct mem_cgrou
}
}
}
-#endif
+
+void lru_gen_online_memcg(struct mem_cgroup *memcg)
+{
+ int gen;
+ int nid;
+ int bin = prandom_u32_max(MEMCG_NR_BINS);
+
+ for_each_node(nid) {
+ struct pglist_data *pgdat = NODE_DATA(nid);
+ struct lruvec *lruvec = get_lruvec(memcg, nid);
+
+ spin_lock(&pgdat->memcg_lru.lock);
+
+ VM_WARN_ON_ONCE(!hlist_nulls_unhashed(&lruvec->lrugen.list));
+
+ gen = get_memcg_gen(pgdat->memcg_lru.seq);
+
+ hlist_nulls_add_tail_rcu(&lruvec->lrugen.list, &pgdat->memcg_lru.fifo[gen][bin]);
+ pgdat->memcg_lru.nr_memcgs[gen]++;
+
+ lruvec->lrugen.gen = gen;
+
+ spin_unlock(&pgdat->memcg_lru.lock);
+ }
+}
+
+void lru_gen_offline_memcg(struct mem_cgroup *memcg)
+{
+ int nid;
+
+ for_each_node(nid) {
+ struct lruvec *lruvec = get_lruvec(memcg, nid);
+
+ lru_gen_rotate_memcg(lruvec, MEMCG_LRU_OLD);
+ }
+}
+
+void lru_gen_release_memcg(struct mem_cgroup *memcg)
+{
+ int gen;
+ int nid;
+
+ for_each_node(nid) {
+ struct pglist_data *pgdat = NODE_DATA(nid);
+ struct lruvec *lruvec = get_lruvec(memcg, nid);
+
+ spin_lock(&pgdat->memcg_lru.lock);
+
+ VM_WARN_ON_ONCE(hlist_nulls_unhashed(&lruvec->lrugen.list));
+
+ gen = lruvec->lrugen.gen;
+
+ hlist_nulls_del_rcu(&lruvec->lrugen.list);
+ pgdat->memcg_lru.nr_memcgs[gen]--;
+
+ if (!pgdat->memcg_lru.nr_memcgs[gen] && gen == get_memcg_gen(pgdat->memcg_lru.seq))
+ WRITE_ONCE(pgdat->memcg_lru.seq, pgdat->memcg_lru.seq + 1);
+
+ spin_unlock(&pgdat->memcg_lru.lock);
+ }
+}
+
+#endif /* CONFIG_MEMCG */
static int __init init_lru_gen(void)
{
@@ -5849,6 +6147,10 @@ static void lru_gen_shrink_lruvec(struct
{
}
+static void lru_gen_shrink_node(struct pglist_data *pgdat, struct scan_control *sc)
+{
+}
+
#endif /* CONFIG_LRU_GEN */
static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)
@@ -5862,7 +6164,7 @@ static void shrink_lruvec(struct lruvec
bool proportional_reclaim;
struct blk_plug plug;
- if (lru_gen_enabled()) {
+ if (lru_gen_enabled() && !global_reclaim(sc)) {
lru_gen_shrink_lruvec(lruvec, sc);
return;
}
@@ -6105,6 +6407,11 @@ static void shrink_node(pg_data_t *pgdat
struct lruvec *target_lruvec;
bool reclaimable = false;
+ if (lru_gen_enabled() && global_reclaim(sc)) {
+ lru_gen_shrink_node(pgdat, sc);
+ return;
+ }
+
target_lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup, pgdat);
again:

196
target/linux/generic/backport-6.1/020-v6.3-27-mm-multi-gen-LRU-clarify-scan_control-flags.patch Normal file
View File

@ -0,0 +1,196 @@
From 93147736b5b3a21bea24313bfc7a696829932009 Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:19:05 -0700
Subject: [PATCH 27/29] mm: multi-gen LRU: clarify scan_control flags
Among the flags in scan_control:
1. sc->may_swap, which indicates swap constraint due to memsw.max, is
supported as usual.
2. sc->proactive, which indicates reclaim by memory.reclaim, may not
opportunistically skip the aging path, since it is considered less
latency sensitive.
3. !(sc->gfp_mask & __GFP_IO), which indicates IO constraint, lowers
swappiness to prioritize file LRU, since clean file folios are more
likely to exist.
4. sc->may_writefolio and sc->may_unmap, which indicates opportunistic
reclaim, are rejected, since unmapped clean folios are already
prioritized. Scanning for more of them is likely futile and can
cause high reclaim latency when there is a large number of memcgs.
The rest are handled by the existing code.
Link: https://lkml.kernel.org/r/20221222041905.2431096-8-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/vmscan.c | 55 +++++++++++++++++++++++++++--------------------------
1 file changed, 28 insertions(+), 27 deletions(-)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3185,6 +3185,9 @@ static int get_swappiness(struct lruvec
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
struct pglist_data *pgdat = lruvec_pgdat(lruvec);
+ if (!sc->may_swap)
+ return 0;
+
if (!can_demote(pgdat->node_id, sc) &&
mem_cgroup_get_nr_swap_pages(memcg) < MIN_LRU_BATCH)
return 0;
@@ -4226,7 +4229,7 @@ static void walk_mm(struct lruvec *lruve
} while (err == -EAGAIN);
}
-static struct lru_gen_mm_walk *set_mm_walk(struct pglist_data *pgdat)
+static struct lru_gen_mm_walk *set_mm_walk(struct pglist_data *pgdat, bool force_alloc)
{
struct lru_gen_mm_walk *walk = current->reclaim_state->mm_walk;
@@ -4234,7 +4237,7 @@ static struct lru_gen_mm_walk *set_mm_wa
VM_WARN_ON_ONCE(walk);
walk = &pgdat->mm_walk;
- } else if (!pgdat && !walk) {
+ } else if (!walk && force_alloc) {
VM_WARN_ON_ONCE(current_is_kswapd());
walk = kzalloc(sizeof(*walk), __GFP_HIGH | __GFP_NOMEMALLOC | __GFP_NOWARN);
@@ -4420,7 +4423,7 @@ static bool try_to_inc_max_seq(struct lr
goto done;
}
- walk = set_mm_walk(NULL);
+ walk = set_mm_walk(NULL, true);
if (!walk) {
success = iterate_mm_list_nowalk(lruvec, max_seq);
goto done;
@@ -4489,8 +4492,6 @@ static bool lruvec_is_reclaimable(struct
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
DEFINE_MIN_SEQ(lruvec);
- VM_WARN_ON_ONCE(sc->memcg_low_reclaim);
-
/* see the comment on lru_gen_folio */
gen = lru_gen_from_seq(min_seq[LRU_GEN_FILE]);
birth = READ_ONCE(lruvec->lrugen.timestamps[gen]);
@@ -4746,12 +4747,8 @@ static bool isolate_folio(struct lruvec
{
bool success;
- /* unmapping inhibited */
- if (!sc->may_unmap && folio_mapped(folio))
- return false;
-
/* swapping inhibited */
- if (!(sc->may_writepage && (sc->gfp_mask & __GFP_IO)) &&
+ if (!(sc->gfp_mask & __GFP_IO) &&
(folio_test_dirty(folio) ||
(folio_test_anon(folio) && !folio_test_swapcache(folio))))
return false;
@@ -4848,9 +4845,8 @@ static int scan_folios(struct lruvec *lr
__count_vm_events(PGSCAN_ANON + type, isolated);
/*
- * There might not be eligible pages due to reclaim_idx, may_unmap and
- * may_writepage. Check the remaining to prevent livelock if it's not
- * making progress.
+ * There might not be eligible pages due to reclaim_idx. Check the
+ * remaining to prevent livelock if it's not making progress.
*/
return isolated || !remaining ? scanned : 0;
}
@@ -5110,8 +5106,7 @@ static long get_nr_to_scan(struct lruvec
struct mem_cgroup *memcg = lruvec_memcg(lruvec);
DEFINE_MAX_SEQ(lruvec);
- if (mem_cgroup_below_min(memcg) ||
- (mem_cgroup_below_low(memcg) && !sc->memcg_low_reclaim))
+ if (mem_cgroup_below_min(memcg))
return 0;
if (!should_run_aging(lruvec, max_seq, sc, can_swap, &nr_to_scan))
@@ -5139,17 +5134,14 @@ static bool try_to_shrink_lruvec(struct
long nr_to_scan;
unsigned long scanned = 0;
unsigned long nr_to_reclaim = get_nr_to_reclaim(sc);
+ int swappiness = get_swappiness(lruvec, sc);
+
+ /* clean file folios are more likely to exist */
+ if (swappiness && !(sc->gfp_mask & __GFP_IO))
+ swappiness = 1;
while (true) {
int delta;
- int swappiness;
-
- if (sc->may_swap)
- swappiness = get_swappiness(lruvec, sc);
- else if (!cgroup_reclaim(sc) && get_swappiness(lruvec, sc))
- swappiness = 1;
- else
- swappiness = 0;
nr_to_scan = get_nr_to_scan(lruvec, sc, swappiness);
if (nr_to_scan <= 0)
@@ -5279,12 +5271,13 @@ static void lru_gen_shrink_lruvec(struct
struct blk_plug plug;
VM_WARN_ON_ONCE(global_reclaim(sc));
+ VM_WARN_ON_ONCE(!sc->may_writepage || !sc->may_unmap);
lru_add_drain();
blk_start_plug(&plug);
- set_mm_walk(lruvec_pgdat(lruvec));
+ set_mm_walk(NULL, sc->proactive);
if (try_to_shrink_lruvec(lruvec, sc))
lru_gen_rotate_memcg(lruvec, MEMCG_LRU_YOUNG);
@@ -5340,11 +5333,19 @@ static void lru_gen_shrink_node(struct p
VM_WARN_ON_ONCE(!global_reclaim(sc));
+ /*
+ * Unmapped clean folios are already prioritized. Scanning for more of
+ * them is likely futile and can cause high reclaim latency when there
+ * is a large number of memcgs.
+ */
+ if (!sc->may_writepage || !sc->may_unmap)
+ goto done;
+
lru_add_drain();
blk_start_plug(&plug);
- set_mm_walk(pgdat);
+ set_mm_walk(NULL, sc->proactive);
set_initial_priority(pgdat, sc);
@@ -5362,7 +5363,7 @@ static void lru_gen_shrink_node(struct p
clear_mm_walk();
blk_finish_plug(&plug);
-
+done:
/* kswapd should never fail */
pgdat->kswapd_failures = 0;
}
@@ -5934,7 +5935,7 @@ static ssize_t lru_gen_seq_write(struct
set_task_reclaim_state(current, &sc.reclaim_state);
flags = memalloc_noreclaim_save();
blk_start_plug(&plug);
- if (!set_mm_walk(NULL)) {
+ if (!set_mm_walk(NULL, true)) {
err = -ENOMEM;
goto done;
}

34
target/linux/generic/backport-6.1/020-v6.3-28-mm-multi-gen-LRU-simplify-arch_has_hw_pte_young-chec.patch Normal file
View File

@ -0,0 +1,34 @@
From cf3297e4c7a928da8b2b2f0baff2f9c69ea57952 Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Wed, 21 Dec 2022 21:19:06 -0700
Subject: [PATCH 28/29] mm: multi-gen LRU: simplify arch_has_hw_pte_young()
check
Scanning page tables when hardware does not set the accessed bit has
no real use cases.
Link: https://lkml.kernel.org/r/20221222041905.2431096-9-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michael Larabel <Michael@MichaelLarabel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/vmscan.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -4418,7 +4418,7 @@ static bool try_to_inc_max_seq(struct lr
* handful of PTEs. Spreading the work out over a period of time usually
* is less efficient, but it avoids bursty page faults.
*/
- if (!force_scan && !(arch_has_hw_pte_young() && get_cap(LRU_GEN_MM_WALK))) {
+ if (!arch_has_hw_pte_young() || !get_cap(LRU_GEN_MM_WALK)) {
success = iterate_mm_list_nowalk(lruvec, max_seq);
goto done;
}

88
target/linux/generic/backport-6.1/020-v6.3-29-mm-multi-gen-LRU-avoid-futile-retries.patch Normal file
View File

@ -0,0 +1,88 @@
From cc67f962cc53f6e1dfa92eb85b7b26fe83a3c66f Mon Sep 17 00:00:00 2001
From: Yu Zhao <yuzhao@google.com>
Date: Mon, 13 Feb 2023 00:53:22 -0700
Subject: [PATCH 29/29] mm: multi-gen LRU: avoid futile retries
Recall that the per-node memcg LRU has two generations and they alternate
when the last memcg (of a given node) is moved from one to the other.
Each generation is also sharded into multiple bins to improve scalability.
A reclaimer starts with a random bin (in the old generation) and, if it
fails, it will retry, i.e., to try the rest of the bins.
If a reclaimer fails with the last memcg, it should move this memcg to the
young generation first, which causes the generations to alternate, and
then retry. Otherwise, the retries will be futile because all other bins
are empty.
Link: https://lkml.kernel.org/r/20230213075322.1416966-1-yuzhao@google.com
Fixes: e4dde56cd208 ("mm: multi-gen LRU: per-node lru_gen_folio lists")
Signed-off-by: Yu Zhao <yuzhao@google.com>
Reported-by: T.J. Mercier <tjmercier@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/vmscan.c | 25 +++++++++++++++----------
1 file changed, 15 insertions(+), 10 deletions(-)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -5208,18 +5208,20 @@ static int shrink_one(struct lruvec *lru
static void shrink_many(struct pglist_data *pgdat, struct scan_control *sc)
{
+ int op;
int gen;
int bin;
int first_bin;
struct lruvec *lruvec;
struct lru_gen_folio *lrugen;
+ struct mem_cgroup *memcg;
const struct hlist_nulls_node *pos;
- int op = 0;
- struct mem_cgroup *memcg = NULL;
unsigned long nr_to_reclaim = get_nr_to_reclaim(sc);
bin = first_bin = prandom_u32_max(MEMCG_NR_BINS);
restart:
+ op = 0;
+ memcg = NULL;
gen = get_memcg_gen(READ_ONCE(pgdat->memcg_lru.seq));
rcu_read_lock();
@@ -5243,14 +5245,22 @@ restart:
op = shrink_one(lruvec, sc);
- if (sc->nr_reclaimed >= nr_to_reclaim)
- goto success;
-
rcu_read_lock();
+
+ if (sc->nr_reclaimed >= nr_to_reclaim)
+ break;
}
rcu_read_unlock();
+ if (op)
+ lru_gen_rotate_memcg(lruvec, op);
+
+ mem_cgroup_put(memcg);
+
+ if (sc->nr_reclaimed >= nr_to_reclaim)
+ return;
+
/* restart if raced with lru_gen_rotate_memcg() */
if (gen != get_nulls_value(pos))
goto restart;
@@ -5259,11 +5269,6 @@ restart:
bin = get_memcg_bin(bin + 1);
if (bin != first_bin)
goto restart;
-success:
- if (op)
- lru_gen_rotate_memcg(lruvec, op);
-
- mem_cgroup_put(memcg);
}
static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc)

65
target/linux/generic/backport-6.1/406-v6.2-0001-mtd-core-simplify-a-bit-code-find-partition-matching.patch Normal file
View File

@ -0,0 +1,65 @@
From 63db0cb35e1cb3b3c134906d1062f65513fdda2d Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl>
Date: Tue, 4 Oct 2022 10:37:09 +0200
Subject: [PATCH] mtd: core: simplify (a bit) code find partition-matching
dynamic OF node
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
1. Don't hardcode "partition-" string twice
2. Use simpler logic & use ->name to avoid of_property_read_string()
3. Use mtd_get_of_node() helper
Cc: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20221004083710.27704-1-zajec5@gmail.com
---
drivers/mtd/mtdcore.c | 16 +++++++---------
1 file changed, 7 insertions(+), 9 deletions(-)
--- a/drivers/mtd/mtdcore.c
+++ b/drivers/mtd/mtdcore.c
@@ -551,18 +551,16 @@ static void mtd_check_of_node(struct mtd
struct device_node *partitions, *parent_dn, *mtd_dn = NULL;
const char *pname, *prefix = "partition-";
int plen, mtd_name_len, offset, prefix_len;
- struct mtd_info *parent;
bool found = false;
/* Check if MTD already has a device node */
- if (dev_of_node(&mtd->dev))
+ if (mtd_get_of_node(mtd))
return;
/* Check if a partitions node exist */
if (!mtd_is_partition(mtd))
return;
- parent = mtd->parent;
- parent_dn = of_node_get(dev_of_node(&parent->dev));
+ parent_dn = of_node_get(mtd_get_of_node(mtd->parent));
if (!parent_dn)
return;
@@ -575,15 +573,15 @@ static void mtd_check_of_node(struct mtd
/* Search if a partition is defined with the same name */
for_each_child_of_node(partitions, mtd_dn) {
- offset = 0;
-
/* Skip partition with no/wrong prefix */
- if (!of_node_name_prefix(mtd_dn, "partition-"))
+ if (!of_node_name_prefix(mtd_dn, prefix))
continue;
/* Label have priority. Check that first */
- if (of_property_read_string(mtd_dn, "label", &pname)) {
- of_property_read_string(mtd_dn, "name", &pname);
+ if (!of_property_read_string(mtd_dn, "label", &pname)) {
+ offset = 0;
+ } else {
+ pname = mtd_dn->name;
offset = prefix_len;
}

84
target/linux/generic/backport-6.1/406-v6.2-0002-mtd-core-try-to-find-OF-node-for-every-MTD-partition.patch Normal file
View File

@ -0,0 +1,84 @@
From ddb8cefb7af288950447ca6eeeafb09977dab56f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl>
Date: Tue, 4 Oct 2022 10:37:10 +0200
Subject: [PATCH] mtd: core: try to find OF node for every MTD partition
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
So far this feature was limited to the top-level "nvmem-cells" node.
There are multiple parsers creating partitions and subpartitions
dynamically. Extend that code to handle them too.
This allows finding partition-* node for every MTD (sub)partition.
Random example:
partitions {
compatible = "brcm,bcm947xx-cfe-partitions";
partition-firmware {
compatible = "brcm,trx";
partition-loader {
};
};
};
Cc: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20221004083710.27704-2-zajec5@gmail.com
---
drivers/mtd/mtdcore.c | 18 ++++++------------
1 file changed, 6 insertions(+), 12 deletions(-)
--- a/drivers/mtd/mtdcore.c
+++ b/drivers/mtd/mtdcore.c
@@ -551,20 +551,22 @@ static void mtd_check_of_node(struct mtd
struct device_node *partitions, *parent_dn, *mtd_dn = NULL;
const char *pname, *prefix = "partition-";
int plen, mtd_name_len, offset, prefix_len;
- bool found = false;
/* Check if MTD already has a device node */
if (mtd_get_of_node(mtd))
return;
- /* Check if a partitions node exist */
if (!mtd_is_partition(mtd))
return;
+
parent_dn = of_node_get(mtd_get_of_node(mtd->parent));
if (!parent_dn)
return;
- partitions = of_get_child_by_name(parent_dn, "partitions");
+ if (mtd_is_partition(mtd->parent))
+ partitions = of_node_get(parent_dn);
+ else
+ partitions = of_get_child_by_name(parent_dn, "partitions");
if (!partitions)
goto exit_parent;
@@ -588,19 +590,11 @@ static void mtd_check_of_node(struct mtd
plen = strlen(pname) - offset;
if (plen == mtd_name_len &&
!strncmp(mtd->name, pname + offset, plen)) {
- found = true;
+ mtd_set_of_node(mtd, mtd_dn);
break;
}
}
- if (!found)
- goto exit_partitions;
-
- /* Set of_node only for nvmem */
- if (of_device_is_compatible(mtd_dn, "nvmem-cells"))
- mtd_set_of_node(mtd, mtd_dn);
-
-exit_partitions:
of_node_put(partitions);
exit_parent:
of_node_put(parent_dn);

47
target/linux/generic/backport-6.1/408-v6.2-mtd-core-set-ROOT_DEV-for-partitions-marked-as-rootf.patch Normal file
View File

@ -0,0 +1,47 @@
From 26422ac78e9d8767bd4aabfbae616b15edbf6a1b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl>
Date: Sat, 22 Oct 2022 23:13:18 +0200
Subject: [PATCH] mtd: core: set ROOT_DEV for partitions marked as rootfs in DT
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
This adds support for "linux,rootfs" binding that is used to mark flash
partition containing rootfs. It's useful for devices using device tree
that don't have bootloader passing root info in cmdline.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20221022211318.32009-2-zajec5@gmail.com
---
drivers/mtd/mtdcore.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
--- a/drivers/mtd/mtdcore.c
+++ b/drivers/mtd/mtdcore.c
@@ -28,6 +28,7 @@
#include <linux/leds.h>
#include <linux/debugfs.h>
#include <linux/nvmem-provider.h>
+#include <linux/root_dev.h>
#include <linux/mtd/mtd.h>
#include <linux/mtd/partitions.h>
@@ -737,6 +738,17 @@ int add_mtd_device(struct mtd_info *mtd)
not->add(mtd);
mutex_unlock(&mtd_table_mutex);
+
+ if (of_find_property(mtd_get_of_node(mtd), "linux,rootfs", NULL)) {
+ if (IS_BUILTIN(CONFIG_MTD)) {
+ pr_info("mtd: setting mtd%d (%s) as root device\n", mtd->index, mtd->name);
+ ROOT_DEV = MKDEV(MTD_BLOCK_MAJOR, mtd->index);
+ } else {
+ pr_warn("mtd: can't set mtd%d (%s) as root device - mtd must be builtin\n",
+ mtd->index, mtd->name);
+ }
+ }
+
/* We _know_ we aren't being removed, because
our caller is still holding us here. So none
of this try_ nonsense, and no bitching about it

229
target/linux/generic/backport-6.1/421-v6.2-mtd-parsers-add-TP-Link-SafeLoader-partitions-table-.patch Normal file
View File

@ -0,0 +1,229 @@
From aec4d5f5ffd0f0092bd9dc21ea90e0bc237d4b74 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl>
Date: Sat, 15 Oct 2022 11:29:50 +0200
Subject: [PATCH] mtd: parsers: add TP-Link SafeLoader partitions table parser
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
This parser deals with most TP-Link home routers. It reads info about
partitions and registers them in the MTD subsystem.
Example from TP-Link Archer C5 V2:
spi-nor spi0.0: s25fl128s1 (16384 Kbytes)
15 tplink-safeloader partitions found on MTD device spi0.0
Creating 15 MTD partitions on "spi0.0":
0x000000000000-0x000000040000 : "fs-uboot"
0x000000040000-0x000000440000 : "os-image"
0x000000440000-0x000000e40000 : "rootfs"
0x000000e40000-0x000000e40200 : "default-mac"
0x000000e40200-0x000000e40400 : "pin"
0x000000e40400-0x000000e40600 : "product-info"
0x000000e50000-0x000000e60000 : "partition-table"
0x000000e60000-0x000000e60200 : "soft-version"
0x000000e61000-0x000000e70000 : "support-list"
0x000000e70000-0x000000e80000 : "profile"
0x000000e80000-0x000000e90000 : "default-config"
0x000000e90000-0x000000ee0000 : "user-config"
0x000000ee0000-0x000000fe0000 : "log"
0x000000fe0000-0x000000ff0000 : "radio_bk"
0x000000ff0000-0x000001000000 : "radio"
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20221015092950.27467-2-zajec5@gmail.com
---
drivers/mtd/parsers/Kconfig | 15 +++
drivers/mtd/parsers/Makefile | 1 +
drivers/mtd/parsers/tplink_safeloader.c | 150 ++++++++++++++++++++++++
3 files changed, 166 insertions(+)
create mode 100644 drivers/mtd/parsers/tplink_safeloader.c
--- a/drivers/mtd/parsers/Kconfig
+++ b/drivers/mtd/parsers/Kconfig
@@ -123,6 +123,21 @@ config MTD_AFS_PARTS
for your particular device. It won't happen automatically. The
'physmap' map driver (CONFIG_MTD_PHYSMAP) does this, for example.
+config MTD_PARSER_TPLINK_SAFELOADER
+ tristate "TP-Link Safeloader partitions parser"
+ depends on MTD && (ARCH_BCM_5301X || ATH79 || SOC_MT7620 || SOC_MT7621 || COMPILE_TEST)
+ help
+ TP-Link home routers use flash partitions to store various data. Info
+ about flash space layout is stored in a partitions table using a
+ custom ASCII-based format.
+
+ That format was first found in devices with SafeLoader bootloader and
+ was named after it. Later it was adapted to CFE and U-Boot
+ bootloaders.
+
+ This driver reads partitions table, parses it and creates MTD
+ partitions.
+
config MTD_PARSER_TRX
tristate "Parser for TRX format partitions"
depends on MTD && (BCM47XX || ARCH_BCM_5301X || ARCH_MEDIATEK || RALINK || COMPILE_TEST)
--- a/drivers/mtd/parsers/Makefile
+++ b/drivers/mtd/parsers/Makefile
@@ -10,6 +10,7 @@ ofpart-$(CONFIG_MTD_OF_PARTS_BCM4908) +=
ofpart-$(CONFIG_MTD_OF_PARTS_LINKSYS_NS)+= ofpart_linksys_ns.o
obj-$(CONFIG_MTD_PARSER_IMAGETAG) += parser_imagetag.o
obj-$(CONFIG_MTD_AFS_PARTS) += afs.o
+obj-$(CONFIG_MTD_PARSER_TPLINK_SAFELOADER) += tplink_safeloader.o
obj-$(CONFIG_MTD_PARSER_TRX) += parser_trx.o
obj-$(CONFIG_MTD_SERCOMM_PARTS) += scpart.o
obj-$(CONFIG_MTD_SHARPSL_PARTS) += sharpslpart.o
--- /dev/null
+++ b/drivers/mtd/parsers/tplink_safeloader.c
@@ -0,0 +1,150 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright © 2022 Rafał Miłecki <rafal@milecki.pl>
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/mtd/mtd.h>
+#include <linux/mtd/partitions.h>
+#include <linux/of.h>
+#include <linux/slab.h>
+
+#define TPLINK_SAFELOADER_DATA_OFFSET 4
+#define TPLINK_SAFELOADER_MAX_PARTS 32
+
+struct safeloader_cmn_header {
+ __be32 size;
+ uint32_t unused;
+} __packed;
+
+static void *mtd_parser_tplink_safeloader_read_table(struct mtd_info *mtd)
+{
+ struct safeloader_cmn_header hdr;
+ struct device_node *np;
+ size_t bytes_read;
+ size_t offset;
+ size_t size;
+ char *buf;
+ int err;
+
+ np = mtd_get_of_node(mtd);
+ if (mtd_is_partition(mtd))
+ of_node_get(np);
+ else
+ np = of_get_child_by_name(np, "partitions");
+
+ if (of_property_read_u32(np, "partitions-table-offset", (u32 *)&offset)) {
+ pr_err("Failed to get partitions table offset\n");
+ goto err_put;
+ }
+
+ err = mtd_read(mtd, offset, sizeof(hdr), &bytes_read, (uint8_t *)&hdr);
+ if (err && !mtd_is_bitflip(err)) {
+ pr_err("Failed to read from %s at 0x%zx\n", mtd->name, offset);
+ goto err_put;
+ }
+
+ size = be32_to_cpu(hdr.size);
+
+ buf = kmalloc(size + 1, GFP_KERNEL);
+ if (!buf)
+ goto err_put;
+
+ err = mtd_read(mtd, offset + sizeof(hdr), size, &bytes_read, buf);
+ if (err && !mtd_is_bitflip(err)) {
+ pr_err("Failed to read from %s at 0x%zx\n", mtd->name, offset + sizeof(hdr));
+ goto err_kfree;
+ }
+
+ buf[size] = '\0';
+
+ of_node_put(np);
+
+ return buf;
+
+err_kfree:
+ kfree(buf);
+err_put:
+ of_node_put(np);
+ return NULL;
+}
+
+static int mtd_parser_tplink_safeloader_parse(struct mtd_info *mtd,
+ const struct mtd_partition **pparts,
+ struct mtd_part_parser_data *data)
+{
+ struct mtd_partition *parts;
+ char name[65];
+ size_t offset;
+ size_t bytes;
+ char *buf;
+ int idx;
+ int err;
+
+ parts = kcalloc(TPLINK_SAFELOADER_MAX_PARTS, sizeof(*parts), GFP_KERNEL);
+ if (!parts) {
+ err = -ENOMEM;
+ goto err_out;
+ }
+
+ buf = mtd_parser_tplink_safeloader_read_table(mtd);
+ if (!buf) {
+ err = -ENOENT;
+ goto err_out;
+ }
+
+ for (idx = 0, offset = TPLINK_SAFELOADER_DATA_OFFSET;
+ idx < TPLINK_SAFELOADER_MAX_PARTS &&
+ sscanf(buf + offset, "partition %64s base 0x%llx size 0x%llx%zn\n",
+ name, &parts[idx].offset, &parts[idx].size, &bytes) == 3;
+ idx++, offset += bytes + 1) {
+ parts[idx].name = kstrdup(name, GFP_KERNEL);
+ if (!parts[idx].name) {
+ err = -ENOMEM;
+ goto err_free;
+ }
+ }
+
+ if (idx == TPLINK_SAFELOADER_MAX_PARTS)
+ pr_warn("Reached maximum number of partitions!\n");
+
+ kfree(buf);
+
+ *pparts = parts;
+
+ return idx;
+
+err_free:
+ for (idx -= 1; idx >= 0; idx--)
+ kfree(parts[idx].name);
+err_out:
+ return err;
+};
+
+static void mtd_parser_tplink_safeloader_cleanup(const struct mtd_partition *pparts,
+ int nr_parts)
+{
+ int i;
+
+ for (i = 0; i < nr_parts; i++)
+ kfree(pparts[i].name);
+
+ kfree(pparts);
+}
+
+static const struct of_device_id mtd_parser_tplink_safeloader_of_match_table[] = {
+ { .compatible = "tplink,safeloader-partitions" },
+ {},
+};
+MODULE_DEVICE_TABLE(of, mtd_parser_tplink_safeloader_of_match_table);
+
+static struct mtd_part_parser mtd_parser_tplink_safeloader = {
+ .parse_fn = mtd_parser_tplink_safeloader_parse,
+ .cleanup = mtd_parser_tplink_safeloader_cleanup,
+ .name = "tplink-safeloader",
+ .of_match_table = mtd_parser_tplink_safeloader_of_match_table,
+};
+module_mtd_part_parser(mtd_parser_tplink_safeloader);
+
+MODULE_LICENSE("GPL");

35
target/linux/generic/backport-6.1/423-v6.3-mtd-spinand-macronix-use-scratch-buffer-for-DMA-oper.patch Normal file
View File

@ -0,0 +1,35 @@
From ebed787a0becb9354f0a23620a5130cccd6c730c Mon Sep 17 00:00:00 2001
From: Daniel Golle <daniel@makrotopia.org>
Date: Thu, 19 Jan 2023 03:45:43 +0000
Subject: [PATCH] mtd: spinand: macronix: use scratch buffer for DMA operation
The mx35lf1ge4ab_get_eccsr() function uses an SPI DMA operation to
read the eccsr, hence the buffer should not be on stack. Since commit
380583227c0c7f ("spi: spi-mem: Add extra sanity checks on the op param")
the kernel emmits a warning and blocks such operations.
Use the scratch buffer to get eccsr instead of trying to directly read
into a stack-allocated variable.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/Y8i85zM0u4XdM46z@makrotopia.org
---
drivers/mtd/nand/spi/macronix.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
--- a/drivers/mtd/nand/spi/macronix.c
+++ b/drivers/mtd/nand/spi/macronix.c
@@ -83,9 +83,10 @@ static int mx35lf1ge4ab_ecc_get_status(s
* in order to avoid forcing the wear-leveling layer to move
* data around if it's not necessary.
*/
- if (mx35lf1ge4ab_get_eccsr(spinand, &eccsr))
+ if (mx35lf1ge4ab_get_eccsr(spinand, spinand->scratchbuf))
return nanddev_get_ecc_conf(nand)->strength;
+ eccsr = *spinand->scratchbuf;
if (WARN_ON(eccsr > nanddev_get_ecc_conf(nand)->strength ||
!eccsr))
return nanddev_get_ecc_conf(nand)->strength;

47
target/linux/generic/backport-6.1/424-v6.4-0004-mtd-core-prepare-mtd_otp_nvmem_add-to-handle-EPROBE_.patch Normal file
View File

@ -0,0 +1,47 @@
From 281f7a6c1a33fffcde32001bacbb4f672140fbf9 Mon Sep 17 00:00:00 2001
From: Michael Walle <michael@walle.cc>
Date: Wed, 8 Mar 2023 09:20:21 +0100
Subject: [PATCH] mtd: core: prepare mtd_otp_nvmem_add() to handle
-EPROBE_DEFER
NVMEM soon will get the ability for nvmem layouts and these might
not be ready when nvmem_register() is called and thus it might
return -EPROBE_DEFER. Don't print the error message in this case.
Signed-off-by: Michael Walle <michael@walle.cc>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20230308082021.870459-4-michael@walle.cc
---
drivers/mtd/mtdcore.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)
--- a/drivers/mtd/mtdcore.c
+++ b/drivers/mtd/mtdcore.c
@@ -953,8 +953,8 @@ static int mtd_otp_nvmem_add(struct mtd_
nvmem = mtd_otp_nvmem_register(mtd, "user-otp", size,
mtd_nvmem_user_otp_reg_read);
if (IS_ERR(nvmem)) {
- dev_err(dev, "Failed to register OTP NVMEM device\n");
- return PTR_ERR(nvmem);
+ err = PTR_ERR(nvmem);
+ goto err;
}
mtd->otp_user_nvmem = nvmem;
}
@@ -971,7 +971,6 @@ static int mtd_otp_nvmem_add(struct mtd_
nvmem = mtd_otp_nvmem_register(mtd, "factory-otp", size,
mtd_nvmem_fact_otp_reg_read);
if (IS_ERR(nvmem)) {
- dev_err(dev, "Failed to register OTP NVMEM device\n");
err = PTR_ERR(nvmem);
goto err;
}
@@ -983,7 +982,7 @@ static int mtd_otp_nvmem_add(struct mtd_
err:
nvmem_unregister(mtd->otp_user_nvmem);
- return err;
+ return dev_err_probe(dev, err, "Failed to register OTP NVMEM device\n");
}
/**

56
target/linux/generic/backport-6.1/610-v6.3-net-page_pool-use-in_softirq-instead.patch Normal file
View File

@ -0,0 +1,56 @@
From: Qingfang DENG <qingfang.deng@siflower.com.cn>
Date: Fri, 3 Feb 2023 09:16:11 +0800
Subject: [PATCH] net: page_pool: use in_softirq() instead
We use BH context only for synchronization, so we don't care if it's
actually serving softirq or not.
As a side node, in case of threaded NAPI, in_serving_softirq() will
return false because it's in process context with BH off, making
page_pool_recycle_in_cache() unreachable.
Signed-off-by: Qingfang DENG <qingfang.deng@siflower.com.cn>
---
--- a/include/net/page_pool.h
+++ b/include/net/page_pool.h
@@ -386,7 +386,7 @@ static inline void page_pool_nid_changed
static inline void page_pool_ring_lock(struct page_pool *pool)
__acquires(&pool->ring.producer_lock)
{
- if (in_serving_softirq())
+ if (in_softirq())
spin_lock(&pool->ring.producer_lock);
else
spin_lock_bh(&pool->ring.producer_lock);
@@ -395,7 +395,7 @@ static inline void page_pool_ring_lock(s
static inline void page_pool_ring_unlock(struct page_pool *pool)
__releases(&pool->ring.producer_lock)
{
- if (in_serving_softirq())
+ if (in_softirq())
spin_unlock(&pool->ring.producer_lock);
else
spin_unlock_bh(&pool->ring.producer_lock);
--- a/net/core/page_pool.c
+++ b/net/core/page_pool.c
@@ -511,8 +511,8 @@ static void page_pool_return_page(struct
static bool page_pool_recycle_in_ring(struct page_pool *pool, struct page *page)
{
int ret;
- /* BH protection not needed if current is serving softirq */
- if (in_serving_softirq())
+ /* BH protection not needed if current is softirq */
+ if (in_softirq())
ret = ptr_ring_produce(&pool->ring, page);
else
ret = ptr_ring_produce_bh(&pool->ring, page);
@@ -570,7 +570,7 @@ __page_pool_put_page(struct page_pool *p
page_pool_dma_sync_for_device(pool, page,
dma_sync_size);
- if (allow_direct && in_serving_softirq() &&
+ if (allow_direct && in_softirq() &&
page_pool_recycle_in_cache(page, pool))
return NULL;

41
target/linux/generic/backport-6.1/611-v6.3-net-add-helper-eth_addr_add.patch Normal file
View File

@ -0,0 +1,41 @@
From 7390609b0121a1b982c5ecdfcd72dc328e5784ee Mon Sep 17 00:00:00 2001
From: Michael Walle <michael@walle.cc>
Date: Mon, 6 Feb 2023 13:43:42 +0000
Subject: [PATCH] net: add helper eth_addr_add()
Add a helper to add an offset to a ethernet address. This comes in handy
if you have a base ethernet address for multiple interfaces.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20230206134356.839737-9-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
include/linux/etherdevice.h | 14 ++++++++++++++
1 file changed, 14 insertions(+)
--- a/include/linux/etherdevice.h
+++ b/include/linux/etherdevice.h
@@ -508,6 +508,20 @@ static inline void eth_addr_inc(u8 *addr
}
/**
+ * eth_addr_add() - Add (or subtract) an offset to/from the given MAC address.
+ *
+ * @offset: Offset to add.
+ * @addr: Pointer to a six-byte array containing Ethernet address to increment.
+ */
+static inline void eth_addr_add(u8 *addr, long offset)
+{
+ u64 u = ether_addr_to_u64(addr);
+
+ u += offset;
+ u64_to_ether_addr(u, addr);
+}
+
+/**
* is_etherdev_addr - Tell if given Ethernet address belongs to the device.
* @dev: Pointer to a device structure
* @addr: Pointer to a six-byte array containing the Ethernet address

394
target/linux/generic/backport-6.1/707-v6.3-net-pcs-add-driver-for-MediaTek-SGMII-PCS.patch Normal file
View File

@ -0,0 +1,394 @@
From 4765a9722e09765866e131ec31f7b9cf4c1f4854 Mon Sep 17 00:00:00 2001
From: Daniel Golle <daniel@makrotopia.org>
Date: Sun, 19 Mar 2023 12:57:50 +0000
Subject: [PATCH] net: pcs: add driver for MediaTek SGMII PCS
The SGMII core found in several MediaTek SoCs is identical to what can
also be found in MediaTek's MT7531 Ethernet switch IC.
As this has not always been clear, both drivers developed different
implementations to deal with the PCS.
Recently Alexander Couzens pointed out this fact which lead to the
development of this shared driver.
Add a dedicated driver, mostly by copying the code now found in the
Ethernet driver. The now redundant code will be removed by a follow-up
commit.
Suggested-by: Alexander Couzens <lynxis@fe80.eu>
Suggested-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
MAINTAINERS | 8 +
drivers/net/pcs/Kconfig | 7 +
drivers/net/pcs/Makefile | 1 +
drivers/net/pcs/pcs-mtk-lynxi.c | 305 ++++++++++++++++++++++++++++++
include/linux/pcs/pcs-mtk-lynxi.h | 13 ++
5 files changed, 334 insertions(+)
create mode 100644 drivers/net/pcs/pcs-mtk-lynxi.c
create mode 100644 include/linux/pcs/pcs-mtk-lynxi.h
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -12926,6 +12926,14 @@ L: netdev@vger.kernel.org
S: Maintained
F: drivers/net/ethernet/mediatek/
+MEDIATEK ETHERNET PCS DRIVER
+M: Alexander Couzens <lynxis@fe80.eu>
+M: Daniel Golle <daniel@makrotopia.org>
+L: netdev@vger.kernel.org
+S: Maintained
+F: drivers/net/pcs/pcs-mtk-lynxi.c
+F: include/linux/pcs/pcs-mtk-lynxi.h
+
MEDIATEK I2C CONTROLLER DRIVER
M: Qii Wang <qii.wang@mediatek.com>
L: linux-i2c@vger.kernel.org
--- a/drivers/net/pcs/Kconfig
+++ b/drivers/net/pcs/Kconfig
@@ -32,4 +32,11 @@ config PCS_ALTERA_TSE
This module provides helper functions for the Altera Triple Speed
Ethernet SGMII PCS, that can be found on the Intel Socfpga family.
+config PCS_MTK_LYNXI
+ tristate
+ select REGMAP
+ help
+ This module provides helpers to phylink for managing the LynxI PCS
+ which is part of MediaTek's SoC and Ethernet switch ICs.
+
endmenu
--- a/drivers/net/pcs/Makefile
+++ b/drivers/net/pcs/Makefile
@@ -7,3 +7,4 @@ obj-$(CONFIG_PCS_XPCS) += pcs_xpcs.o
obj-$(CONFIG_PCS_LYNX) += pcs-lynx.o
obj-$(CONFIG_PCS_RZN1_MIIC) += pcs-rzn1-miic.o
obj-$(CONFIG_PCS_ALTERA_TSE) += pcs-altera-tse.o
+obj-$(CONFIG_PCS_MTK_LYNXI) += pcs-mtk-lynxi.o
--- /dev/null
+++ b/drivers/net/pcs/pcs-mtk-lynxi.c
@@ -0,0 +1,305 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (c) 2018-2019 MediaTek Inc.
+/* A library for MediaTek SGMII circuit
+ *
+ * Author: Sean Wang <sean.wang@mediatek.com>
+ * Author: Alexander Couzens <lynxis@fe80.eu>
+ * Author: Daniel Golle <daniel@makrotopia.org>
+ *
+ */
+
+#include <linux/mdio.h>
+#include <linux/of.h>
+#include <linux/pcs/pcs-mtk-lynxi.h>
+#include <linux/phylink.h>
+#include <linux/regmap.h>
+
+/* SGMII subsystem config registers */
+/* BMCR (low 16) BMSR (high 16) */
+#define SGMSYS_PCS_CONTROL_1 0x0
+#define SGMII_BMCR GENMASK(15, 0)
+#define SGMII_BMSR GENMASK(31, 16)
+
+#define SGMSYS_PCS_DEVICE_ID 0x4
+#define SGMII_LYNXI_DEV_ID 0x4d544950
+
+#define SGMSYS_PCS_ADVERTISE 0x8
+#define SGMII_ADVERTISE GENMASK(15, 0)
+#define SGMII_LPA GENMASK(31, 16)
+
+#define SGMSYS_PCS_SCRATCH 0x14
+#define SGMII_DEV_VERSION GENMASK(31, 16)
+
+/* Register to programmable link timer, the unit in 2 * 8ns */
+#define SGMSYS_PCS_LINK_TIMER 0x18
+#define SGMII_LINK_TIMER_MASK GENMASK(19, 0)
+#define SGMII_LINK_TIMER_VAL(ns) FIELD_PREP(SGMII_LINK_TIMER_MASK, \
+ ((ns) / 2 / 8))
+
+/* Register to control remote fault */
+#define SGMSYS_SGMII_MODE 0x20
+#define SGMII_IF_MODE_SGMII BIT(0)
+#define SGMII_SPEED_DUPLEX_AN BIT(1)
+#define SGMII_SPEED_MASK GENMASK(3, 2)
+#define SGMII_SPEED_10 FIELD_PREP(SGMII_SPEED_MASK, 0)
+#define SGMII_SPEED_100 FIELD_PREP(SGMII_SPEED_MASK, 1)
+#define SGMII_SPEED_1000 FIELD_PREP(SGMII_SPEED_MASK, 2)
+#define SGMII_DUPLEX_HALF BIT(4)
+#define SGMII_REMOTE_FAULT_DIS BIT(8)
+
+/* Register to reset SGMII design */
+#define SGMSYS_RESERVED_0 0x34
+#define SGMII_SW_RESET BIT(0)
+
+/* Register to set SGMII speed, ANA RG_ Control Signals III */
+#define SGMII_PHY_SPEED_MASK GENMASK(3, 2)
+#define SGMII_PHY_SPEED_1_25G FIELD_PREP(SGMII_PHY_SPEED_MASK, 0)
+#define SGMII_PHY_SPEED_3_125G FIELD_PREP(SGMII_PHY_SPEED_MASK, 1)
+
+/* Register to power up QPHY */
+#define SGMSYS_QPHY_PWR_STATE_CTRL 0xe8
+#define SGMII_PHYA_PWD BIT(4)
+
+/* Register to QPHY wrapper control */
+#define SGMSYS_QPHY_WRAP_CTRL 0xec
+#define SGMII_PN_SWAP_MASK GENMASK(1, 0)
+#define SGMII_PN_SWAP_TX_RX (BIT(0) | BIT(1))
+
+/* struct mtk_pcs_lynxi - This structure holds each sgmii regmap andassociated
+ * data
+ * @regmap: The register map pointing at the range used to setup
+ * SGMII modes
+ * @dev: Pointer to device owning the PCS
+ * @ana_rgc3: The offset of register ANA_RGC3 relative to regmap
+ * @interface: Currently configured interface mode
+ * @pcs: Phylink PCS structure
+ * @flags: Flags indicating hardware properties
+ */
+struct mtk_pcs_lynxi {
+ struct regmap *regmap;
+ u32 ana_rgc3;
+ phy_interface_t interface;
+ struct phylink_pcs pcs;
+ u32 flags;
+};
+
+static struct mtk_pcs_lynxi *pcs_to_mtk_pcs_lynxi(struct phylink_pcs *pcs)
+{
+ return container_of(pcs, struct mtk_pcs_lynxi, pcs);
+}
+
+static void mtk_pcs_lynxi_get_state(struct phylink_pcs *pcs,
+ struct phylink_link_state *state)
+{
+ struct mtk_pcs_lynxi *mpcs = pcs_to_mtk_pcs_lynxi(pcs);
+ unsigned int bm, adv;
+
+ /* Read the BMSR and LPA */
+ regmap_read(mpcs->regmap, SGMSYS_PCS_CONTROL_1, &bm);
+ regmap_read(mpcs->regmap, SGMSYS_PCS_ADVERTISE, &adv);
+
+ phylink_mii_c22_pcs_decode_state(state, FIELD_GET(SGMII_BMSR, bm),
+ FIELD_GET(SGMII_LPA, adv));
+}
+
+static int mtk_pcs_lynxi_config(struct phylink_pcs *pcs, unsigned int mode,
+ phy_interface_t interface,
+ const unsigned long *advertising,
+ bool permit_pause_to_mac)
+{
+ struct mtk_pcs_lynxi *mpcs = pcs_to_mtk_pcs_lynxi(pcs);
+ bool mode_changed = false, changed, use_an;
+ unsigned int rgc3, sgm_mode, bmcr;
+ int advertise, link_timer;
+
+ advertise = phylink_mii_c22_pcs_encode_advertisement(interface,
+ advertising);
+ if (advertise < 0)
+ return advertise;
+
+ /* Clearing IF_MODE_BIT0 switches the PCS to BASE-X mode, and
+ * we assume that fixes it's speed at bitrate = line rate (in
+ * other words, 1000Mbps or 2500Mbps).
+ */
+ if (interface == PHY_INTERFACE_MODE_SGMII) {
+ sgm_mode = SGMII_IF_MODE_SGMII;
+ if (phylink_autoneg_inband(mode)) {
+ sgm_mode |= SGMII_REMOTE_FAULT_DIS |
+ SGMII_SPEED_DUPLEX_AN;
+ use_an = true;
+ } else {
+ use_an = false;
+ }
+ } else if (phylink_autoneg_inband(mode)) {
+ /* 1000base-X or 2500base-X autoneg */
+ sgm_mode = SGMII_REMOTE_FAULT_DIS;
+ use_an = linkmode_test_bit(ETHTOOL_LINK_MODE_Autoneg_BIT,
+ advertising);
+ } else {
+ /* 1000base-X or 2500base-X without autoneg */
+ sgm_mode = 0;
+ use_an = false;
+ }
+
+ if (use_an)
+ bmcr = BMCR_ANENABLE;
+ else
+ bmcr = 0;
+
+ if (mpcs->interface != interface) {
+ link_timer = phylink_get_link_timer_ns(interface);
+ if (link_timer < 0)
+ return link_timer;
+
+ /* PHYA power down */
+ regmap_set_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
+ SGMII_PHYA_PWD);
+
+ /* Reset SGMII PCS state */
+ regmap_set_bits(mpcs->regmap, SGMSYS_RESERVED_0,
+ SGMII_SW_RESET);
+
+ if (mpcs->flags & MTK_SGMII_FLAG_PN_SWAP)
+ regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_WRAP_CTRL,
+ SGMII_PN_SWAP_MASK,
+ SGMII_PN_SWAP_TX_RX);
+
+ if (interface == PHY_INTERFACE_MODE_2500BASEX)
+ rgc3 = SGMII_PHY_SPEED_3_125G;
+ else
+ rgc3 = SGMII_PHY_SPEED_1_25G;
+
+ /* Configure the underlying interface speed */
+ regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
+ SGMII_PHY_SPEED_MASK, rgc3);
+
+ /* Setup the link timer */
+ regmap_write(mpcs->regmap, SGMSYS_PCS_LINK_TIMER,
+ SGMII_LINK_TIMER_VAL(link_timer));
+
+ mpcs->interface = interface;
+ mode_changed = true;
+ }
+
+ /* Update the advertisement, noting whether it has changed */
+ regmap_update_bits_check(mpcs->regmap, SGMSYS_PCS_ADVERTISE,
+ SGMII_ADVERTISE, advertise, &changed);
+
+ /* Update the sgmsys mode register */
+ regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
+ SGMII_REMOTE_FAULT_DIS | SGMII_SPEED_DUPLEX_AN |
+ SGMII_IF_MODE_SGMII, sgm_mode);
+
+ /* Update the BMCR */
+ regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
+ BMCR_ANENABLE, bmcr);
+
+ /* Release PHYA power down state
+ * Only removing bit SGMII_PHYA_PWD isn't enough.
+ * There are cases when the SGMII_PHYA_PWD register contains 0x9 which
+ * prevents SGMII from working. The SGMII still shows link but no traffic
+ * can flow. Writing 0x0 to the PHYA_PWD register fix the issue. 0x0 was
+ * taken from a good working state of the SGMII interface.
+ * Unknown how much the QPHY needs but it is racy without a sleep.
+ * Tested on mt7622 & mt7986.
+ */
+ usleep_range(50, 100);
+ regmap_write(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, 0);
+
+ return changed || mode_changed;
+}
+
+static void mtk_pcs_lynxi_restart_an(struct phylink_pcs *pcs)
+{
+ struct mtk_pcs_lynxi *mpcs = pcs_to_mtk_pcs_lynxi(pcs);
+
+ regmap_set_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1, BMCR_ANRESTART);
+}
+
+static void mtk_pcs_lynxi_link_up(struct phylink_pcs *pcs, unsigned int mode,
+ phy_interface_t interface, int speed,
+ int duplex)
+{
+ struct mtk_pcs_lynxi *mpcs = pcs_to_mtk_pcs_lynxi(pcs);
+ unsigned int sgm_mode;
+
+ if (!phylink_autoneg_inband(mode)) {
+ /* Force the speed and duplex setting */
+ if (speed == SPEED_10)
+ sgm_mode = SGMII_SPEED_10;
+ else if (speed == SPEED_100)
+ sgm_mode = SGMII_SPEED_100;
+ else
+ sgm_mode = SGMII_SPEED_1000;
+
+ if (duplex != DUPLEX_FULL)
+ sgm_mode |= SGMII_DUPLEX_HALF;
+
+ regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
+ SGMII_DUPLEX_HALF | SGMII_SPEED_MASK,
+ sgm_mode);
+ }
+}
+
+static const struct phylink_pcs_ops mtk_pcs_lynxi_ops = {
+ .pcs_get_state = mtk_pcs_lynxi_get_state,
+ .pcs_config = mtk_pcs_lynxi_config,
+ .pcs_an_restart = mtk_pcs_lynxi_restart_an,
+ .pcs_link_up = mtk_pcs_lynxi_link_up,
+};
+
+struct phylink_pcs *mtk_pcs_lynxi_create(struct device *dev,
+ struct regmap *regmap, u32 ana_rgc3,
+ u32 flags)
+{
+ struct mtk_pcs_lynxi *mpcs;
+ u32 id, ver;
+ int ret;
+
+ ret = regmap_read(regmap, SGMSYS_PCS_DEVICE_ID, &id);
+ if (ret < 0)
+ return NULL;
+
+ if (id != SGMII_LYNXI_DEV_ID) {
+ dev_err(dev, "unknown PCS device id %08x\n", id);
+ return NULL;
+ }
+
+ ret = regmap_read(regmap, SGMSYS_PCS_SCRATCH, &ver);
+ if (ret < 0)
+ return NULL;
+
+ ver = FIELD_GET(SGMII_DEV_VERSION, ver);
+ if (ver != 0x1) {
+ dev_err(dev, "unknown PCS device version %04x\n", ver);
+ return NULL;
+ }
+
+ dev_dbg(dev, "MediaTek LynxI SGMII PCS (id 0x%08x, ver 0x%04x)\n", id,
+ ver);
+
+ mpcs = kzalloc(sizeof(*mpcs), GFP_KERNEL);
+ if (!mpcs)
+ return NULL;
+
+ mpcs->ana_rgc3 = ana_rgc3;
+ mpcs->regmap = regmap;
+ mpcs->flags = flags;
+ mpcs->pcs.ops = &mtk_pcs_lynxi_ops;
+ mpcs->pcs.poll = true;
+ mpcs->interface = PHY_INTERFACE_MODE_NA;
+
+ return &mpcs->pcs;
+}
+EXPORT_SYMBOL(mtk_pcs_lynxi_create);
+
+void mtk_pcs_lynxi_destroy(struct phylink_pcs *pcs)
+{
+ if (!pcs)
+ return;
+
+ kfree(pcs_to_mtk_pcs_lynxi(pcs));
+}
+EXPORT_SYMBOL(mtk_pcs_lynxi_destroy);
+
+MODULE_LICENSE("GPL");
--- /dev/null
+++ b/include/linux/pcs/pcs-mtk-lynxi.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LINUX_PCS_MTK_LYNXI_H
+#define __LINUX_PCS_MTK_LYNXI_H
+
+#include <linux/phylink.h>
+#include <linux/regmap.h>
+
+#define MTK_SGMII_FLAG_PN_SWAP BIT(0)
+struct phylink_pcs *mtk_pcs_lynxi_create(struct device *dev,
+ struct regmap *regmap,
+ u32 ana_rgc3, u32 flags);
+void mtk_pcs_lynxi_destroy(struct phylink_pcs *pcs);
+#endif

591
target/linux/generic/backport-6.1/729-01-v6.1-net-ethernet-mtk_wed-introduce-wed-mcu-support.patch Normal file
View File

@ -0,0 +1,591 @@
From: Sujuan Chen <sujuan.chen@mediatek.com>
Date: Sat, 5 Nov 2022 23:36:18 +0100
Subject: [PATCH] net: ethernet: mtk_wed: introduce wed mcu support
Introduce WED mcu support used to configure WED WO chip.
This is a preliminary patch in order to add RX Wireless
Ethernet Dispatch available on MT7986 SoC.
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
create mode 100644 drivers/net/ethernet/mediatek/mtk_wed_mcu.c
create mode 100644 drivers/net/ethernet/mediatek/mtk_wed_wo.h
--- a/drivers/net/ethernet/mediatek/Makefile
+++ b/drivers/net/ethernet/mediatek/Makefile
@@ -5,7 +5,7 @@
obj-$(CONFIG_NET_MEDIATEK_SOC) += mtk_eth.o
mtk_eth-y := mtk_eth_soc.o mtk_sgmii.o mtk_eth_path.o mtk_ppe.o mtk_ppe_debugfs.o mtk_ppe_offload.o
-mtk_eth-$(CONFIG_NET_MEDIATEK_SOC_WED) += mtk_wed.o
+mtk_eth-$(CONFIG_NET_MEDIATEK_SOC_WED) += mtk_wed.o mtk_wed_mcu.o
ifdef CONFIG_DEBUG_FS
mtk_eth-$(CONFIG_NET_MEDIATEK_SOC_WED) += mtk_wed_debugfs.o
endif
--- /dev/null
+++ b/drivers/net/ethernet/mediatek/mtk_wed_mcu.c
@@ -0,0 +1,359 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2022 MediaTek Inc.
+ *
+ * Author: Lorenzo Bianconi <lorenzo@kernel.org>
+ * Sujuan Chen <sujuan.chen@mediatek.com>
+ */
+
+#include <linux/firmware.h>
+#include <linux/of_address.h>
+#include <linux/of_reserved_mem.h>
+#include <linux/mfd/syscon.h>
+#include <linux/soc/mediatek/mtk_wed.h>
+
+#include "mtk_wed_regs.h"
+#include "mtk_wed_wo.h"
+#include "mtk_wed.h"
+
+static u32 wo_r32(struct mtk_wed_wo *wo, u32 reg)
+{
+ return readl(wo->boot.addr + reg);
+}
+
+static void wo_w32(struct mtk_wed_wo *wo, u32 reg, u32 val)
+{
+ writel(val, wo->boot.addr + reg);
+}
+
+static struct sk_buff *
+mtk_wed_mcu_msg_alloc(const void *data, int data_len)
+{
+ int length = sizeof(struct mtk_wed_mcu_hdr) + data_len;
+ struct sk_buff *skb;
+
+ skb = alloc_skb(length, GFP_KERNEL);
+ if (!skb)
+ return NULL;
+
+ memset(skb->head, 0, length);
+ skb_reserve(skb, sizeof(struct mtk_wed_mcu_hdr));
+ if (data && data_len)
+ skb_put_data(skb, data, data_len);
+
+ return skb;
+}
+
+static struct sk_buff *
+mtk_wed_mcu_get_response(struct mtk_wed_wo *wo, unsigned long expires)
+{
+ if (!time_is_after_jiffies(expires))
+ return NULL;
+
+ wait_event_timeout(wo->mcu.wait, !skb_queue_empty(&wo->mcu.res_q),
+ expires - jiffies);
+ return skb_dequeue(&wo->mcu.res_q);
+}
+
+void mtk_wed_mcu_rx_event(struct mtk_wed_wo *wo, struct sk_buff *skb)
+{
+ skb_queue_tail(&wo->mcu.res_q, skb);
+ wake_up(&wo->mcu.wait);
+}
+
+void mtk_wed_mcu_rx_unsolicited_event(struct mtk_wed_wo *wo,
+ struct sk_buff *skb)
+{
+ struct mtk_wed_mcu_hdr *hdr = (struct mtk_wed_mcu_hdr *)skb->data;
+
+ switch (hdr->cmd) {
+ case MTK_WED_WO_EVT_LOG_DUMP: {
+ const char *msg = (const char *)(skb->data + sizeof(*hdr));
+
+ dev_notice(wo->hw->dev, "%s\n", msg);
+ break;
+ }
+ case MTK_WED_WO_EVT_PROFILING: {
+ struct mtk_wed_wo_log_info *info;
+ u32 count = (skb->len - sizeof(*hdr)) / sizeof(*info);
+ int i;
+
+ info = (struct mtk_wed_wo_log_info *)(skb->data + sizeof(*hdr));
+ for (i = 0 ; i < count ; i++)
+ dev_notice(wo->hw->dev,
+ "SN:%u latency: total=%u, rro:%u, mod:%u\n",
+ le32_to_cpu(info[i].sn),
+ le32_to_cpu(info[i].total),
+ le32_to_cpu(info[i].rro),
+ le32_to_cpu(info[i].mod));
+ break;
+ }
+ case MTK_WED_WO_EVT_RXCNT_INFO:
+ break;
+ default:
+ break;
+ }
+
+ dev_kfree_skb(skb);
+}
+
+static int
+mtk_wed_mcu_skb_send_msg(struct mtk_wed_wo *wo, struct sk_buff *skb,
+ int id, int cmd, u16 *wait_seq, bool wait_resp)
+{
+ struct mtk_wed_mcu_hdr *hdr;
+
+ /* TODO: make it dynamic based on cmd */
+ wo->mcu.timeout = 20 * HZ;
+
+ hdr = (struct mtk_wed_mcu_hdr *)skb_push(skb, sizeof(*hdr));
+ hdr->cmd = cmd;
+ hdr->length = cpu_to_le16(skb->len);
+
+ if (wait_resp && wait_seq) {
+ u16 seq = ++wo->mcu.seq;
+
+ if (!seq)
+ seq = ++wo->mcu.seq;
+ *wait_seq = seq;
+
+ hdr->flag |= cpu_to_le16(MTK_WED_WARP_CMD_FLAG_NEED_RSP);
+ hdr->seq = cpu_to_le16(seq);
+ }
+ if (id == MTK_WED_MODULE_ID_WO)
+ hdr->flag |= cpu_to_le16(MTK_WED_WARP_CMD_FLAG_FROM_TO_WO);
+
+ dev_kfree_skb(skb);
+ return 0;
+}
+
+static int
+mtk_wed_mcu_parse_response(struct mtk_wed_wo *wo, struct sk_buff *skb,
+ int cmd, int seq)
+{
+ struct mtk_wed_mcu_hdr *hdr;
+
+ if (!skb) {
+ dev_err(wo->hw->dev, "Message %08x (seq %d) timeout\n",
+ cmd, seq);
+ return -ETIMEDOUT;
+ }
+
+ hdr = (struct mtk_wed_mcu_hdr *)skb->data;
+ if (le16_to_cpu(hdr->seq) != seq)
+ return -EAGAIN;
+
+ skb_pull(skb, sizeof(*hdr));
+ switch (cmd) {
+ case MTK_WED_WO_CMD_RXCNT_INFO:
+ default:
+ break;
+ }
+
+ return 0;
+}
+
+int mtk_wed_mcu_send_msg(struct mtk_wed_wo *wo, int id, int cmd,
+ const void *data, int len, bool wait_resp)
+{
+ unsigned long expires;
+ struct sk_buff *skb;
+ u16 seq;
+ int ret;
+
+ skb = mtk_wed_mcu_msg_alloc(data, len);
+ if (!skb)
+ return -ENOMEM;
+
+ mutex_lock(&wo->mcu.mutex);
+
+ ret = mtk_wed_mcu_skb_send_msg(wo, skb, id, cmd, &seq, wait_resp);
+ if (ret || !wait_resp)
+ goto unlock;
+
+ expires = jiffies + wo->mcu.timeout;
+ do {
+ skb = mtk_wed_mcu_get_response(wo, expires);
+ ret = mtk_wed_mcu_parse_response(wo, skb, cmd, seq);
+ dev_kfree_skb(skb);
+ } while (ret == -EAGAIN);
+
+unlock:
+ mutex_unlock(&wo->mcu.mutex);
+
+ return ret;
+}
+
+static int
+mtk_wed_get_memory_region(struct mtk_wed_wo *wo,
+ struct mtk_wed_wo_memory_region *region)
+{
+ struct reserved_mem *rmem;
+ struct device_node *np;
+ int index;
+
+ index = of_property_match_string(wo->hw->node, "memory-region-names",
+ region->name);
+ if (index < 0)
+ return index;
+
+ np = of_parse_phandle(wo->hw->node, "memory-region", index);
+ if (!np)
+ return -ENODEV;
+
+ rmem = of_reserved_mem_lookup(np);
+ of_node_put(np);
+
+ if (!rmem)
+ return -ENODEV;
+
+ region->phy_addr = rmem->base;
+ region->size = rmem->size;
+ region->addr = devm_ioremap(wo->hw->dev, region->phy_addr, region->size);
+
+ return !region->addr ? -EINVAL : 0;
+}
+
+static int
+mtk_wed_mcu_run_firmware(struct mtk_wed_wo *wo, const struct firmware *fw,
+ struct mtk_wed_wo_memory_region *region)
+{
+ const u8 *first_region_ptr, *region_ptr, *trailer_ptr, *ptr = fw->data;
+ const struct mtk_wed_fw_trailer *trailer;
+ const struct mtk_wed_fw_region *fw_region;
+
+ trailer_ptr = fw->data + fw->size - sizeof(*trailer);
+ trailer = (const struct mtk_wed_fw_trailer *)trailer_ptr;
+ region_ptr = trailer_ptr - trailer->num_region * sizeof(*fw_region);
+ first_region_ptr = region_ptr;
+
+ while (region_ptr < trailer_ptr) {
+ u32 length;
+
+ fw_region = (const struct mtk_wed_fw_region *)region_ptr;
+ length = le32_to_cpu(fw_region->len);
+
+ if (region->phy_addr != le32_to_cpu(fw_region->addr))
+ goto next;
+
+ if (region->size < length)
+ goto next;
+
+ if (first_region_ptr < ptr + length)
+ goto next;
+
+ if (region->shared && region->consumed)
+ return 0;
+
+ if (!region->shared || !region->consumed) {
+ memcpy_toio(region->addr, ptr, length);
+ region->consumed = true;
+ return 0;
+ }
+next:
+ region_ptr += sizeof(*fw_region);
+ ptr += length;
+ }
+
+ return -EINVAL;
+}
+
+static int
+mtk_wed_mcu_load_firmware(struct mtk_wed_wo *wo)
+{
+ static struct mtk_wed_wo_memory_region mem_region[] = {
+ [MTK_WED_WO_REGION_EMI] = {
+ .name = "wo-emi",
+ },
+ [MTK_WED_WO_REGION_ILM] = {
+ .name = "wo-ilm",
+ },
+ [MTK_WED_WO_REGION_DATA] = {
+ .name = "wo-data",
+ .shared = true,
+ },
+ };
+ const struct mtk_wed_fw_trailer *trailer;
+ const struct firmware *fw;
+ const char *fw_name;
+ u32 val, boot_cr;
+ int ret, i;
+
+ /* load firmware region metadata */
+ for (i = 0; i < ARRAY_SIZE(mem_region); i++) {
+ ret = mtk_wed_get_memory_region(wo, &mem_region[i]);
+ if (ret)
+ return ret;
+ }
+
+ wo->boot.name = "wo-boot";
+ ret = mtk_wed_get_memory_region(wo, &wo->boot);
+ if (ret)
+ return ret;
+
+ /* set dummy cr */
+ wed_w32(wo->hw->wed_dev, MTK_WED_SCR0 + 4 * MTK_WED_DUMMY_CR_FWDL,
+ wo->hw->index + 1);
+
+ /* load firmware */
+ fw_name = wo->hw->index ? MT7986_FIRMWARE_WO1 : MT7986_FIRMWARE_WO0;
+ ret = request_firmware(&fw, fw_name, wo->hw->dev);
+ if (ret)
+ return ret;
+
+ trailer = (void *)(fw->data + fw->size -
+ sizeof(struct mtk_wed_fw_trailer));
+ dev_info(wo->hw->dev,
+ "MTK WED WO Firmware Version: %.10s, Build Time: %.15s\n",
+ trailer->fw_ver, trailer->build_date);
+ dev_info(wo->hw->dev, "MTK WED WO Chip ID %02x Region %d\n",
+ trailer->chip_id, trailer->num_region);
+
+ for (i = 0; i < ARRAY_SIZE(mem_region); i++) {
+ ret = mtk_wed_mcu_run_firmware(wo, fw, &mem_region[i]);
+ if (ret)
+ goto out;
+ }
+
+ /* set the start address */
+ boot_cr = wo->hw->index ? MTK_WO_MCU_CFG_LS_WA_BOOT_ADDR_ADDR
+ : MTK_WO_MCU_CFG_LS_WM_BOOT_ADDR_ADDR;
+ wo_w32(wo, boot_cr, mem_region[MTK_WED_WO_REGION_EMI].phy_addr >> 16);
+ /* wo firmware reset */
+ wo_w32(wo, MTK_WO_MCU_CFG_LS_WF_MCCR_CLR_ADDR, 0xc00);
+
+ val = wo_r32(wo, MTK_WO_MCU_CFG_LS_WF_MCU_CFG_WM_WA_ADDR);
+ val |= wo->hw->index ? MTK_WO_MCU_CFG_LS_WF_WM_WA_WA_CPU_RSTB_MASK
+ : MTK_WO_MCU_CFG_LS_WF_WM_WA_WM_CPU_RSTB_MASK;
+ wo_w32(wo, MTK_WO_MCU_CFG_LS_WF_MCU_CFG_WM_WA_ADDR, val);
+out:
+ release_firmware(fw);
+
+ return ret;
+}
+
+static u32
+mtk_wed_mcu_read_fw_dl(struct mtk_wed_wo *wo)
+{
+ return wed_r32(wo->hw->wed_dev,
+ MTK_WED_SCR0 + 4 * MTK_WED_DUMMY_CR_FWDL);
+}
+
+int mtk_wed_mcu_init(struct mtk_wed_wo *wo)
+{
+ u32 val;
+ int ret;
+
+ skb_queue_head_init(&wo->mcu.res_q);
+ init_waitqueue_head(&wo->mcu.wait);
+ mutex_init(&wo->mcu.mutex);
+
+ ret = mtk_wed_mcu_load_firmware(wo);
+ if (ret)
+ return ret;
+
+ return readx_poll_timeout(mtk_wed_mcu_read_fw_dl, wo, val, !val,
+ 100, MTK_FW_DL_TIMEOUT);
+}
+
+MODULE_FIRMWARE(MT7986_FIRMWARE_WO0);
+MODULE_FIRMWARE(MT7986_FIRMWARE_WO1);
--- a/drivers/net/ethernet/mediatek/mtk_wed_regs.h
+++ b/drivers/net/ethernet/mediatek/mtk_wed_regs.h
@@ -152,6 +152,7 @@ struct mtk_wdma_desc {
#define MTK_WED_RING_RX(_n) (0x400 + (_n) * 0x10)
+#define MTK_WED_SCR0 0x3c0
#define MTK_WED_WPDMA_INT_TRIGGER 0x504
#define MTK_WED_WPDMA_INT_TRIGGER_RX_DONE BIT(1)
#define MTK_WED_WPDMA_INT_TRIGGER_TX_DONE GENMASK(5, 4)
--- /dev/null
+++ b/drivers/net/ethernet/mediatek/mtk_wed_wo.h
@@ -0,0 +1,150 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/* Copyright (C) 2022 Lorenzo Bianconi <lorenzo@kernel.org> */
+
+#ifndef __MTK_WED_WO_H
+#define __MTK_WED_WO_H
+
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+
+struct mtk_wed_hw;
+
+struct mtk_wed_mcu_hdr {
+ /* DW0 */
+ u8 version;
+ u8 cmd;
+ __le16 length;
+
+ /* DW1 */
+ __le16 seq;
+ __le16 flag;
+
+ /* DW2 */
+ __le32 status;
+
+ /* DW3 */
+ u8 rsv[20];
+};
+
+struct mtk_wed_wo_log_info {
+ __le32 sn;
+ __le32 total;
+ __le32 rro;
+ __le32 mod;
+};
+
+enum mtk_wed_wo_event {
+ MTK_WED_WO_EVT_LOG_DUMP = 0x1,
+ MTK_WED_WO_EVT_PROFILING = 0x2,
+ MTK_WED_WO_EVT_RXCNT_INFO = 0x3,
+};
+
+#define MTK_WED_MODULE_ID_WO 1
+#define MTK_FW_DL_TIMEOUT 4000000 /* us */
+#define MTK_WOCPU_TIMEOUT 2000000 /* us */
+
+enum {
+ MTK_WED_WARP_CMD_FLAG_RSP = BIT(0),
+ MTK_WED_WARP_CMD_FLAG_NEED_RSP = BIT(1),
+ MTK_WED_WARP_CMD_FLAG_FROM_TO_WO = BIT(2),
+};
+
+enum {
+ MTK_WED_WO_REGION_EMI,
+ MTK_WED_WO_REGION_ILM,
+ MTK_WED_WO_REGION_DATA,
+ MTK_WED_WO_REGION_BOOT,
+ __MTK_WED_WO_REGION_MAX,
+};
+
+enum mtk_wed_dummy_cr_idx {
+ MTK_WED_DUMMY_CR_FWDL,
+ MTK_WED_DUMMY_CR_WO_STATUS,
+};
+
+#define MT7986_FIRMWARE_WO0 "mediatek/mt7986_wo_0.bin"
+#define MT7986_FIRMWARE_WO1 "mediatek/mt7986_wo_1.bin"
+
+#define MTK_WO_MCU_CFG_LS_BASE 0
+#define MTK_WO_MCU_CFG_LS_HW_VER_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x000)
+#define MTK_WO_MCU_CFG_LS_FW_VER_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x004)
+#define MTK_WO_MCU_CFG_LS_CFG_DBG1_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x00c)
+#define MTK_WO_MCU_CFG_LS_CFG_DBG2_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x010)
+#define MTK_WO_MCU_CFG_LS_WF_MCCR_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x014)
+#define MTK_WO_MCU_CFG_LS_WF_MCCR_SET_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x018)
+#define MTK_WO_MCU_CFG_LS_WF_MCCR_CLR_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x01c)
+#define MTK_WO_MCU_CFG_LS_WF_MCU_CFG_WM_WA_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x050)
+#define MTK_WO_MCU_CFG_LS_WM_BOOT_ADDR_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x060)
+#define MTK_WO_MCU_CFG_LS_WA_BOOT_ADDR_ADDR (MTK_WO_MCU_CFG_LS_BASE + 0x064)
+
+#define MTK_WO_MCU_CFG_LS_WF_WM_WA_WM_CPU_RSTB_MASK BIT(5)
+#define MTK_WO_MCU_CFG_LS_WF_WM_WA_WA_CPU_RSTB_MASK BIT(0)
+
+struct mtk_wed_wo_memory_region {
+ const char *name;
+ void __iomem *addr;
+ phys_addr_t phy_addr;
+ u32 size;
+ bool shared:1;
+ bool consumed:1;
+};
+
+struct mtk_wed_fw_region {
+ __le32 decomp_crc;
+ __le32 decomp_len;
+ __le32 decomp_blk_sz;
+ u8 rsv0[4];
+ __le32 addr;
+ __le32 len;
+ u8 feature_set;
+ u8 rsv1[15];
+} __packed;
+
+struct mtk_wed_fw_trailer {
+ u8 chip_id;
+ u8 eco_code;
+ u8 num_region;
+ u8 format_ver;
+ u8 format_flag;
+ u8 rsv[2];
+ char fw_ver[10];
+ char build_date[15];
+ u32 crc;
+};
+
+struct mtk_wed_wo {
+ struct mtk_wed_hw *hw;
+ struct mtk_wed_wo_memory_region boot;
+
+ struct {
+ struct mutex mutex;
+ int timeout;
+ u16 seq;
+
+ struct sk_buff_head res_q;
+ wait_queue_head_t wait;
+ } mcu;
+};
+
+static inline int
+mtk_wed_mcu_check_msg(struct mtk_wed_wo *wo, struct sk_buff *skb)
+{
+ struct mtk_wed_mcu_hdr *hdr = (struct mtk_wed_mcu_hdr *)skb->data;
+
+ if (hdr->version)
+ return -EINVAL;
+
+ if (skb->len < sizeof(*hdr) || skb->len != le16_to_cpu(hdr->length))
+ return -EINVAL;
+
+ return 0;
+}
+
+void mtk_wed_mcu_rx_event(struct mtk_wed_wo *wo, struct sk_buff *skb);
+void mtk_wed_mcu_rx_unsolicited_event(struct mtk_wed_wo *wo,
+ struct sk_buff *skb);
+int mtk_wed_mcu_send_msg(struct mtk_wed_wo *wo, int id, int cmd,
+ const void *data, int len, bool wait_resp);
+int mtk_wed_mcu_init(struct mtk_wed_wo *wo);
+
+#endif /* __MTK_WED_WO_H */
--- a/include/linux/soc/mediatek/mtk_wed.h
+++ b/include/linux/soc/mediatek/mtk_wed.h
@@ -11,6 +11,35 @@
struct mtk_wed_hw;
struct mtk_wdma_desc;
+enum mtk_wed_wo_cmd {
+ MTK_WED_WO_CMD_WED_CFG,
+ MTK_WED_WO_CMD_WED_RX_STAT,
+ MTK_WED_WO_CMD_RRO_SER,
+ MTK_WED_WO_CMD_DBG_INFO,
+ MTK_WED_WO_CMD_DEV_INFO,
+ MTK_WED_WO_CMD_BSS_INFO,
+ MTK_WED_WO_CMD_STA_REC,
+ MTK_WED_WO_CMD_DEV_INFO_DUMP,
+ MTK_WED_WO_CMD_BSS_INFO_DUMP,
+ MTK_WED_WO_CMD_STA_REC_DUMP,
+ MTK_WED_WO_CMD_BA_INFO_DUMP,
+ MTK_WED_WO_CMD_FBCMD_Q_DUMP,
+ MTK_WED_WO_CMD_FW_LOG_CTRL,
+ MTK_WED_WO_CMD_LOG_FLUSH,
+ MTK_WED_WO_CMD_CHANGE_STATE,
+ MTK_WED_WO_CMD_CPU_STATS_ENABLE,
+ MTK_WED_WO_CMD_CPU_STATS_DUMP,
+ MTK_WED_WO_CMD_EXCEPTION_INIT,
+ MTK_WED_WO_CMD_PROF_CTRL,
+ MTK_WED_WO_CMD_STA_BA_DUMP,
+ MTK_WED_WO_CMD_BA_CTRL_DUMP,
+ MTK_WED_WO_CMD_RXCNT_CTRL,
+ MTK_WED_WO_CMD_RXCNT_INFO,
+ MTK_WED_WO_CMD_SET_CAP,
+ MTK_WED_WO_CMD_CCIF_RING_DUMP,
+ MTK_WED_WO_CMD_WED_END
+};
+
enum mtk_wed_bus_tye {
MTK_WED_BUS_PCIE,
MTK_WED_BUS_AXI,

737
target/linux/generic/backport-6.1/729-02-v6.1-net-ethernet-mtk_wed-introduce-wed-wo-support.patch Normal file
View File

@ -0,0 +1,737 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 5 Nov 2022 23:36:19 +0100
Subject: [PATCH] net: ethernet: mtk_wed: introduce wed wo support
Introduce WO chip support to mtk wed driver. MTK WED WO is used to
implement RX Wireless Ethernet Dispatch and offload traffic received by
wlan nic to the wired interface.
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
create mode 100644 drivers/net/ethernet/mediatek/mtk_wed_wo.c
--- a/drivers/net/ethernet/mediatek/Makefile
+++ b/drivers/net/ethernet/mediatek/Makefile
@@ -5,7 +5,7 @@
obj-$(CONFIG_NET_MEDIATEK_SOC) += mtk_eth.o
mtk_eth-y := mtk_eth_soc.o mtk_sgmii.o mtk_eth_path.o mtk_ppe.o mtk_ppe_debugfs.o mtk_ppe_offload.o
-mtk_eth-$(CONFIG_NET_MEDIATEK_SOC_WED) += mtk_wed.o mtk_wed_mcu.o
+mtk_eth-$(CONFIG_NET_MEDIATEK_SOC_WED) += mtk_wed.o mtk_wed_mcu.o mtk_wed_wo.o
ifdef CONFIG_DEBUG_FS
mtk_eth-$(CONFIG_NET_MEDIATEK_SOC_WED) += mtk_wed_debugfs.o
endif
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -16,6 +16,7 @@
#include "mtk_wed_regs.h"
#include "mtk_wed.h"
#include "mtk_ppe.h"
+#include "mtk_wed_wo.h"
#define MTK_PCIE_BASE(n) (0x1a143000 + (n) * 0x2000)
@@ -355,6 +356,8 @@ mtk_wed_detach(struct mtk_wed_device *de
mtk_wed_free_buffer(dev);
mtk_wed_free_tx_rings(dev);
+ if (hw->version != 1)
+ mtk_wed_wo_deinit(hw);
if (dev->wlan.bus_type == MTK_WED_BUS_PCIE) {
struct device_node *wlan_node;
@@ -878,9 +881,11 @@ mtk_wed_attach(struct mtk_wed_device *de
}
mtk_wed_hw_init_early(dev);
- if (hw->hifsys)
+ if (hw->version == 1)
regmap_update_bits(hw->hifsys, HIFSYS_DMA_AG_MAP,
BIT(hw->index), 0);
+ else
+ ret = mtk_wed_wo_init(hw);
out:
mutex_unlock(&hw_lock);
--- a/drivers/net/ethernet/mediatek/mtk_wed.h
+++ b/drivers/net/ethernet/mediatek/mtk_wed.h
@@ -10,6 +10,7 @@
#include <linux/netdevice.h>
struct mtk_eth;
+struct mtk_wed_wo;
struct mtk_wed_hw {
struct device_node *node;
@@ -22,6 +23,7 @@ struct mtk_wed_hw {
struct regmap *mirror;
struct dentry *debugfs_dir;
struct mtk_wed_device *wed_dev;
+ struct mtk_wed_wo *wed_wo;
u32 debugfs_reg;
u32 num_flows;
u8 version;
--- a/drivers/net/ethernet/mediatek/mtk_wed_mcu.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed_mcu.c
@@ -122,8 +122,7 @@ mtk_wed_mcu_skb_send_msg(struct mtk_wed_
if (id == MTK_WED_MODULE_ID_WO)
hdr->flag |= cpu_to_le16(MTK_WED_WARP_CMD_FLAG_FROM_TO_WO);
- dev_kfree_skb(skb);
- return 0;
+ return mtk_wed_wo_queue_tx_skb(wo, &wo->q_tx, skb);
}
static int
--- /dev/null
+++ b/drivers/net/ethernet/mediatek/mtk_wed_wo.c
@@ -0,0 +1,508 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2022 MediaTek Inc.
+ *
+ * Author: Lorenzo Bianconi <lorenzo@kernel.org>
+ * Sujuan Chen <sujuan.chen@mediatek.com>
+ */
+
+#include <linux/kernel.h>
+#include <linux/dma-mapping.h>
+#include <linux/of_platform.h>
+#include <linux/interrupt.h>
+#include <linux/of_address.h>
+#include <linux/mfd/syscon.h>
+#include <linux/of_irq.h>
+#include <linux/bitfield.h>
+
+#include "mtk_wed.h"
+#include "mtk_wed_regs.h"
+#include "mtk_wed_wo.h"
+
+static u32
+mtk_wed_mmio_r32(struct mtk_wed_wo *wo, u32 reg)
+{
+ u32 val;
+
+ if (regmap_read(wo->mmio.regs, reg, &val))
+ val = ~0;
+
+ return val;
+}
+
+static void
+mtk_wed_mmio_w32(struct mtk_wed_wo *wo, u32 reg, u32 val)
+{
+ regmap_write(wo->mmio.regs, reg, val);
+}
+
+static u32
+mtk_wed_wo_get_isr(struct mtk_wed_wo *wo)
+{
+ u32 val = mtk_wed_mmio_r32(wo, MTK_WED_WO_CCIF_RCHNUM);
+
+ return val & MTK_WED_WO_CCIF_RCHNUM_MASK;
+}
+
+static void
+mtk_wed_wo_set_isr(struct mtk_wed_wo *wo, u32 mask)
+{
+ mtk_wed_mmio_w32(wo, MTK_WED_WO_CCIF_IRQ0_MASK, mask);
+}
+
+static void
+mtk_wed_wo_set_ack(struct mtk_wed_wo *wo, u32 mask)
+{
+ mtk_wed_mmio_w32(wo, MTK_WED_WO_CCIF_ACK, mask);
+}
+
+static void
+mtk_wed_wo_set_isr_mask(struct mtk_wed_wo *wo, u32 mask, u32 val, bool set)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&wo->mmio.lock, flags);
+ wo->mmio.irq_mask &= ~mask;
+ wo->mmio.irq_mask |= val;
+ if (set)
+ mtk_wed_wo_set_isr(wo, wo->mmio.irq_mask);
+ spin_unlock_irqrestore(&wo->mmio.lock, flags);
+}
+
+static void
+mtk_wed_wo_irq_enable(struct mtk_wed_wo *wo, u32 mask)
+{
+ mtk_wed_wo_set_isr_mask(wo, 0, mask, false);
+ tasklet_schedule(&wo->mmio.irq_tasklet);
+}
+
+static void
+mtk_wed_wo_irq_disable(struct mtk_wed_wo *wo, u32 mask)
+{
+ mtk_wed_wo_set_isr_mask(wo, mask, 0, true);
+}
+
+static void
+mtk_wed_wo_kickout(struct mtk_wed_wo *wo)
+{
+ mtk_wed_mmio_w32(wo, MTK_WED_WO_CCIF_BUSY, 1 << MTK_WED_WO_TXCH_NUM);
+ mtk_wed_mmio_w32(wo, MTK_WED_WO_CCIF_TCHNUM, MTK_WED_WO_TXCH_NUM);
+}
+
+static void
+mtk_wed_wo_queue_kick(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q,
+ u32 val)
+{
+ wmb();
+ mtk_wed_mmio_w32(wo, q->regs.cpu_idx, val);
+}
+
+static void *
+mtk_wed_wo_dequeue(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q, u32 *len,
+ bool flush)
+{
+ int buf_len = SKB_WITH_OVERHEAD(q->buf_size);
+ int index = (q->tail + 1) % q->n_desc;
+ struct mtk_wed_wo_queue_entry *entry;
+ struct mtk_wed_wo_queue_desc *desc;
+ void *buf;
+
+ if (!q->queued)
+ return NULL;
+
+ if (flush)
+ q->desc[index].ctrl |= cpu_to_le32(MTK_WED_WO_CTL_DMA_DONE);
+ else if (!(q->desc[index].ctrl & cpu_to_le32(MTK_WED_WO_CTL_DMA_DONE)))
+ return NULL;
+
+ q->tail = index;
+ q->queued--;
+
+ desc = &q->desc[index];
+ entry = &q->entry[index];
+ buf = entry->buf;
+ if (len)
+ *len = FIELD_GET(MTK_WED_WO_CTL_SD_LEN0,
+ le32_to_cpu(READ_ONCE(desc->ctrl)));
+ if (buf)
+ dma_unmap_single(wo->hw->dev, entry->addr, buf_len,
+ DMA_FROM_DEVICE);
+ entry->buf = NULL;
+
+ return buf;
+}
+
+static int
+mtk_wed_wo_queue_refill(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q,
+ gfp_t gfp, bool rx)
+{
+ enum dma_data_direction dir = rx ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
+ int n_buf = 0;
+
+ spin_lock_bh(&q->lock);
+ while (q->queued < q->n_desc) {
+ void *buf = page_frag_alloc(&q->cache, q->buf_size, gfp);
+ struct mtk_wed_wo_queue_entry *entry;
+ dma_addr_t addr;
+
+ if (!buf)
+ break;
+
+ addr = dma_map_single(wo->hw->dev, buf, q->buf_size, dir);
+ if (unlikely(dma_mapping_error(wo->hw->dev, addr))) {
+ skb_free_frag(buf);
+ break;
+ }
+
+ q->head = (q->head + 1) % q->n_desc;
+ entry = &q->entry[q->head];
+ entry->addr = addr;
+ entry->len = q->buf_size;
+ q->entry[q->head].buf = buf;
+
+ if (rx) {
+ struct mtk_wed_wo_queue_desc *desc = &q->desc[q->head];
+ u32 ctrl = MTK_WED_WO_CTL_LAST_SEC0 |
+ FIELD_PREP(MTK_WED_WO_CTL_SD_LEN0,
+ entry->len);
+
+ WRITE_ONCE(desc->buf0, cpu_to_le32(addr));
+ WRITE_ONCE(desc->ctrl, cpu_to_le32(ctrl));
+ }
+ q->queued++;
+ n_buf++;
+ }
+ spin_unlock_bh(&q->lock);
+
+ return n_buf;
+}
+
+static void
+mtk_wed_wo_rx_complete(struct mtk_wed_wo *wo)
+{
+ mtk_wed_wo_set_ack(wo, MTK_WED_WO_RXCH_INT_MASK);
+ mtk_wed_wo_irq_enable(wo, MTK_WED_WO_RXCH_INT_MASK);
+}
+
+static void
+mtk_wed_wo_rx_run_queue(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q)
+{
+ for (;;) {
+ struct mtk_wed_mcu_hdr *hdr;
+ struct sk_buff *skb;
+ void *data;
+ u32 len;
+
+ data = mtk_wed_wo_dequeue(wo, q, &len, false);
+ if (!data)
+ break;
+
+ skb = build_skb(data, q->buf_size);
+ if (!skb) {
+ skb_free_frag(data);
+ continue;
+ }
+
+ __skb_put(skb, len);
+ if (mtk_wed_mcu_check_msg(wo, skb)) {
+ dev_kfree_skb(skb);
+ continue;
+ }
+
+ hdr = (struct mtk_wed_mcu_hdr *)skb->data;
+ if (hdr->flag & cpu_to_le16(MTK_WED_WARP_CMD_FLAG_RSP))
+ mtk_wed_mcu_rx_event(wo, skb);
+ else
+ mtk_wed_mcu_rx_unsolicited_event(wo, skb);
+ }
+
+ if (mtk_wed_wo_queue_refill(wo, q, GFP_ATOMIC, true)) {
+ u32 index = (q->head - 1) % q->n_desc;
+
+ mtk_wed_wo_queue_kick(wo, q, index);
+ }
+}
+
+static irqreturn_t
+mtk_wed_wo_irq_handler(int irq, void *data)
+{
+ struct mtk_wed_wo *wo = data;
+
+ mtk_wed_wo_set_isr(wo, 0);
+ tasklet_schedule(&wo->mmio.irq_tasklet);
+
+ return IRQ_HANDLED;
+}
+
+static void mtk_wed_wo_irq_tasklet(struct tasklet_struct *t)
+{
+ struct mtk_wed_wo *wo = from_tasklet(wo, t, mmio.irq_tasklet);
+ u32 intr, mask;
+
+ /* disable interrupts */
+ mtk_wed_wo_set_isr(wo, 0);
+
+ intr = mtk_wed_wo_get_isr(wo);
+ intr &= wo->mmio.irq_mask;
+ mask = intr & (MTK_WED_WO_RXCH_INT_MASK | MTK_WED_WO_EXCEPTION_INT_MASK);
+ mtk_wed_wo_irq_disable(wo, mask);
+
+ if (intr & MTK_WED_WO_RXCH_INT_MASK) {
+ mtk_wed_wo_rx_run_queue(wo, &wo->q_rx);
+ mtk_wed_wo_rx_complete(wo);
+ }
+}
+
+/* mtk wed wo hw queues */
+
+static int
+mtk_wed_wo_queue_alloc(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q,
+ int n_desc, int buf_size, int index,
+ struct mtk_wed_wo_queue_regs *regs)
+{
+ spin_lock_init(&q->lock);
+ q->regs = *regs;
+ q->n_desc = n_desc;
+ q->buf_size = buf_size;
+
+ q->desc = dmam_alloc_coherent(wo->hw->dev, n_desc * sizeof(*q->desc),
+ &q->desc_dma, GFP_KERNEL);
+ if (!q->desc)
+ return -ENOMEM;
+
+ q->entry = devm_kzalloc(wo->hw->dev, n_desc * sizeof(*q->entry),
+ GFP_KERNEL);
+ if (!q->entry)
+ return -ENOMEM;
+
+ return 0;
+}
+
+static void
+mtk_wed_wo_queue_free(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q)
+{
+ mtk_wed_mmio_w32(wo, q->regs.cpu_idx, 0);
+ dma_free_coherent(wo->hw->dev, q->n_desc * sizeof(*q->desc), q->desc,
+ q->desc_dma);
+}
+
+static void
+mtk_wed_wo_queue_tx_clean(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q)
+{
+ struct page *page;
+ int i;
+
+ spin_lock_bh(&q->lock);
+ for (i = 0; i < q->n_desc; i++) {
+ struct mtk_wed_wo_queue_entry *entry = &q->entry[i];
+
+ dma_unmap_single(wo->hw->dev, entry->addr, entry->len,
+ DMA_TO_DEVICE);
+ skb_free_frag(entry->buf);
+ entry->buf = NULL;
+ }
+ spin_unlock_bh(&q->lock);
+
+ if (!q->cache.va)
+ return;
+
+ page = virt_to_page(q->cache.va);
+ __page_frag_cache_drain(page, q->cache.pagecnt_bias);
+ memset(&q->cache, 0, sizeof(q->cache));
+}
+
+static void
+mtk_wed_wo_queue_rx_clean(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q)
+{
+ struct page *page;
+
+ spin_lock_bh(&q->lock);
+ for (;;) {
+ void *buf = mtk_wed_wo_dequeue(wo, q, NULL, true);
+
+ if (!buf)
+ break;
+
+ skb_free_frag(buf);
+ }
+ spin_unlock_bh(&q->lock);
+
+ if (!q->cache.va)
+ return;
+
+ page = virt_to_page(q->cache.va);
+ __page_frag_cache_drain(page, q->cache.pagecnt_bias);
+ memset(&q->cache, 0, sizeof(q->cache));
+}
+
+static void
+mtk_wed_wo_queue_reset(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q)
+{
+ mtk_wed_mmio_w32(wo, q->regs.cpu_idx, 0);
+ mtk_wed_mmio_w32(wo, q->regs.desc_base, q->desc_dma);
+ mtk_wed_mmio_w32(wo, q->regs.ring_size, q->n_desc);
+}
+
+int mtk_wed_wo_queue_tx_skb(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q,
+ struct sk_buff *skb)
+{
+ struct mtk_wed_wo_queue_entry *entry;
+ struct mtk_wed_wo_queue_desc *desc;
+ int ret = 0, index;
+ u32 ctrl;
+
+ spin_lock_bh(&q->lock);
+
+ q->tail = mtk_wed_mmio_r32(wo, q->regs.dma_idx);
+ index = (q->head + 1) % q->n_desc;
+ if (q->tail == index) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ entry = &q->entry[index];
+ if (skb->len > entry->len) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ desc = &q->desc[index];
+ q->head = index;
+
+ dma_sync_single_for_cpu(wo->hw->dev, entry->addr, skb->len,
+ DMA_TO_DEVICE);
+ memcpy(entry->buf, skb->data, skb->len);
+ dma_sync_single_for_device(wo->hw->dev, entry->addr, skb->len,
+ DMA_TO_DEVICE);
+
+ ctrl = FIELD_PREP(MTK_WED_WO_CTL_SD_LEN0, skb->len) |
+ MTK_WED_WO_CTL_LAST_SEC0 | MTK_WED_WO_CTL_DMA_DONE;
+ WRITE_ONCE(desc->buf0, cpu_to_le32(entry->addr));
+ WRITE_ONCE(desc->ctrl, cpu_to_le32(ctrl));
+
+ mtk_wed_wo_queue_kick(wo, q, q->head);
+ mtk_wed_wo_kickout(wo);
+out:
+ spin_unlock_bh(&q->lock);
+
+ dev_kfree_skb(skb);
+
+ return ret;
+}
+
+static int
+mtk_wed_wo_exception_init(struct mtk_wed_wo *wo)
+{
+ return 0;
+}
+
+static int
+mtk_wed_wo_hardware_init(struct mtk_wed_wo *wo)
+{
+ struct mtk_wed_wo_queue_regs regs;
+ struct device_node *np;
+ int ret;
+
+ np = of_parse_phandle(wo->hw->node, "mediatek,wo-ccif", 0);
+ if (!np)
+ return -ENODEV;
+
+ wo->mmio.regs = syscon_regmap_lookup_by_phandle(np, NULL);
+ if (IS_ERR_OR_NULL(wo->mmio.regs))
+ return PTR_ERR(wo->mmio.regs);
+
+ wo->mmio.irq = irq_of_parse_and_map(np, 0);
+ wo->mmio.irq_mask = MTK_WED_WO_ALL_INT_MASK;
+ spin_lock_init(&wo->mmio.lock);
+ tasklet_setup(&wo->mmio.irq_tasklet, mtk_wed_wo_irq_tasklet);
+
+ ret = devm_request_irq(wo->hw->dev, wo->mmio.irq,
+ mtk_wed_wo_irq_handler, IRQF_TRIGGER_HIGH,
+ KBUILD_MODNAME, wo);
+ if (ret)
+ goto error;
+
+ regs.desc_base = MTK_WED_WO_CCIF_DUMMY1;
+ regs.ring_size = MTK_WED_WO_CCIF_DUMMY2;
+ regs.dma_idx = MTK_WED_WO_CCIF_SHADOW4;
+ regs.cpu_idx = MTK_WED_WO_CCIF_DUMMY3;
+
+ ret = mtk_wed_wo_queue_alloc(wo, &wo->q_tx, MTK_WED_WO_RING_SIZE,
+ MTK_WED_WO_CMD_LEN, MTK_WED_WO_TXCH_NUM,
+ &regs);
+ if (ret)
+ goto error;
+
+ mtk_wed_wo_queue_refill(wo, &wo->q_tx, GFP_KERNEL, false);
+ mtk_wed_wo_queue_reset(wo, &wo->q_tx);
+
+ regs.desc_base = MTK_WED_WO_CCIF_DUMMY5;
+ regs.ring_size = MTK_WED_WO_CCIF_DUMMY6;
+ regs.dma_idx = MTK_WED_WO_CCIF_SHADOW8;
+ regs.cpu_idx = MTK_WED_WO_CCIF_DUMMY7;
+
+ ret = mtk_wed_wo_queue_alloc(wo, &wo->q_rx, MTK_WED_WO_RING_SIZE,
+ MTK_WED_WO_CMD_LEN, MTK_WED_WO_RXCH_NUM,
+ &regs);
+ if (ret)
+ goto error;
+
+ mtk_wed_wo_queue_refill(wo, &wo->q_rx, GFP_KERNEL, true);
+ mtk_wed_wo_queue_reset(wo, &wo->q_rx);
+
+ /* rx queue irqmask */
+ mtk_wed_wo_set_isr(wo, wo->mmio.irq_mask);
+
+ return 0;
+
+error:
+ devm_free_irq(wo->hw->dev, wo->mmio.irq, wo);
+
+ return ret;
+}
+
+static void
+mtk_wed_wo_hw_deinit(struct mtk_wed_wo *wo)
+{
+ /* disable interrupts */
+ mtk_wed_wo_set_isr(wo, 0);
+
+ tasklet_disable(&wo->mmio.irq_tasklet);
+
+ disable_irq(wo->mmio.irq);
+ devm_free_irq(wo->hw->dev, wo->mmio.irq, wo);
+
+ mtk_wed_wo_queue_tx_clean(wo, &wo->q_tx);
+ mtk_wed_wo_queue_rx_clean(wo, &wo->q_rx);
+ mtk_wed_wo_queue_free(wo, &wo->q_tx);
+ mtk_wed_wo_queue_free(wo, &wo->q_rx);
+}
+
+int mtk_wed_wo_init(struct mtk_wed_hw *hw)
+{
+ struct mtk_wed_wo *wo;
+ int ret;
+
+ wo = devm_kzalloc(hw->dev, sizeof(*wo), GFP_KERNEL);
+ if (!wo)
+ return -ENOMEM;
+
+ hw->wed_wo = wo;
+ wo->hw = hw;
+
+ ret = mtk_wed_wo_hardware_init(wo);
+ if (ret)
+ return ret;
+
+ ret = mtk_wed_mcu_init(wo);
+ if (ret)
+ return ret;
+
+ return mtk_wed_wo_exception_init(wo);
+}
+
+void mtk_wed_wo_deinit(struct mtk_wed_hw *hw)
+{
+ struct mtk_wed_wo *wo = hw->wed_wo;
+
+ mtk_wed_wo_hw_deinit(wo);
+}
--- a/drivers/net/ethernet/mediatek/mtk_wed_wo.h
+++ b/drivers/net/ethernet/mediatek/mtk_wed_wo.h
@@ -80,6 +80,54 @@ enum mtk_wed_dummy_cr_idx {
#define MTK_WO_MCU_CFG_LS_WF_WM_WA_WM_CPU_RSTB_MASK BIT(5)
#define MTK_WO_MCU_CFG_LS_WF_WM_WA_WA_CPU_RSTB_MASK BIT(0)
+#define MTK_WED_WO_RING_SIZE 256
+#define MTK_WED_WO_CMD_LEN 1504
+
+#define MTK_WED_WO_TXCH_NUM 0
+#define MTK_WED_WO_RXCH_NUM 1
+#define MTK_WED_WO_RXCH_WO_EXCEPTION 7
+
+#define MTK_WED_WO_TXCH_INT_MASK BIT(0)
+#define MTK_WED_WO_RXCH_INT_MASK BIT(1)
+#define MTK_WED_WO_EXCEPTION_INT_MASK BIT(7)
+#define MTK_WED_WO_ALL_INT_MASK (MTK_WED_WO_RXCH_INT_MASK | \
+ MTK_WED_WO_EXCEPTION_INT_MASK)
+
+#define MTK_WED_WO_CCIF_BUSY 0x004
+#define MTK_WED_WO_CCIF_START 0x008
+#define MTK_WED_WO_CCIF_TCHNUM 0x00c
+#define MTK_WED_WO_CCIF_RCHNUM 0x010
+#define MTK_WED_WO_CCIF_RCHNUM_MASK GENMASK(7, 0)
+
+#define MTK_WED_WO_CCIF_ACK 0x014
+#define MTK_WED_WO_CCIF_IRQ0_MASK 0x018
+#define MTK_WED_WO_CCIF_IRQ1_MASK 0x01c
+#define MTK_WED_WO_CCIF_DUMMY1 0x020
+#define MTK_WED_WO_CCIF_DUMMY2 0x024
+#define MTK_WED_WO_CCIF_DUMMY3 0x028
+#define MTK_WED_WO_CCIF_DUMMY4 0x02c
+#define MTK_WED_WO_CCIF_SHADOW1 0x030
+#define MTK_WED_WO_CCIF_SHADOW2 0x034
+#define MTK_WED_WO_CCIF_SHADOW3 0x038
+#define MTK_WED_WO_CCIF_SHADOW4 0x03c
+#define MTK_WED_WO_CCIF_DUMMY5 0x050
+#define MTK_WED_WO_CCIF_DUMMY6 0x054
+#define MTK_WED_WO_CCIF_DUMMY7 0x058
+#define MTK_WED_WO_CCIF_DUMMY8 0x05c
+#define MTK_WED_WO_CCIF_SHADOW5 0x060
+#define MTK_WED_WO_CCIF_SHADOW6 0x064
+#define MTK_WED_WO_CCIF_SHADOW7 0x068
+#define MTK_WED_WO_CCIF_SHADOW8 0x06c
+
+#define MTK_WED_WO_CTL_SD_LEN1 GENMASK(13, 0)
+#define MTK_WED_WO_CTL_LAST_SEC1 BIT(14)
+#define MTK_WED_WO_CTL_BURST BIT(15)
+#define MTK_WED_WO_CTL_SD_LEN0_SHIFT 16
+#define MTK_WED_WO_CTL_SD_LEN0 GENMASK(29, 16)
+#define MTK_WED_WO_CTL_LAST_SEC0 BIT(30)
+#define MTK_WED_WO_CTL_DMA_DONE BIT(31)
+#define MTK_WED_WO_INFO_WINFO GENMASK(15, 0)
+
struct mtk_wed_wo_memory_region {
const char *name;
void __iomem *addr;
@@ -112,10 +160,53 @@ struct mtk_wed_fw_trailer {
u32 crc;
};
+struct mtk_wed_wo_queue_regs {
+ u32 desc_base;
+ u32 ring_size;
+ u32 cpu_idx;
+ u32 dma_idx;
+};
+
+struct mtk_wed_wo_queue_desc {
+ __le32 buf0;
+ __le32 ctrl;
+ __le32 buf1;
+ __le32 info;
+ __le32 reserved[4];
+} __packed __aligned(32);
+
+struct mtk_wed_wo_queue_entry {
+ dma_addr_t addr;
+ void *buf;
+ u32 len;
+};
+
+struct mtk_wed_wo_queue {
+ struct mtk_wed_wo_queue_regs regs;
+
+ struct page_frag_cache cache;
+ spinlock_t lock;
+
+ struct mtk_wed_wo_queue_desc *desc;
+ dma_addr_t desc_dma;
+
+ struct mtk_wed_wo_queue_entry *entry;
+
+ u16 head;
+ u16 tail;
+ int n_desc;
+ int queued;
+ int buf_size;
+
+};
+
struct mtk_wed_wo {
struct mtk_wed_hw *hw;
struct mtk_wed_wo_memory_region boot;
+ struct mtk_wed_wo_queue q_tx;
+ struct mtk_wed_wo_queue q_rx;
+
struct {
struct mutex mutex;
int timeout;
@@ -124,6 +215,15 @@ struct mtk_wed_wo {
struct sk_buff_head res_q;
wait_queue_head_t wait;
} mcu;
+
+ struct {
+ struct regmap *regs;
+
+ spinlock_t lock;
+ struct tasklet_struct irq_tasklet;
+ int irq;
+ u32 irq_mask;
+ } mmio;
};
static inline int
@@ -146,5 +246,9 @@ void mtk_wed_mcu_rx_unsolicited_event(st
int mtk_wed_mcu_send_msg(struct mtk_wed_wo *wo, int id, int cmd,
const void *data, int len, bool wait_resp);
int mtk_wed_mcu_init(struct mtk_wed_wo *wo);
+int mtk_wed_wo_init(struct mtk_wed_hw *hw);
+void mtk_wed_wo_deinit(struct mtk_wed_hw *hw);
+int mtk_wed_wo_queue_tx_skb(struct mtk_wed_wo *dev, struct mtk_wed_wo_queue *q,
+ struct sk_buff *skb);
#endif /* __MTK_WED_WO_H */

79
target/linux/generic/backport-6.1/729-03-v6.1-net-ethernet-mtk_wed-rename-tx_wdma-array-in-rx_wdma.patch Normal file
View File

@ -0,0 +1,79 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 5 Nov 2022 23:36:20 +0100
Subject: [PATCH] net: ethernet: mtk_wed: rename tx_wdma array in rx_wdma
Rename tx_wdma queue array in rx_wdma since this is rx side of wdma soc.
Moreover rename mtk_wed_wdma_ring_setup routine in
mtk_wed_wdma_rx_ring_setup()
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -253,8 +253,8 @@ mtk_wed_free_tx_rings(struct mtk_wed_dev
for (i = 0; i < ARRAY_SIZE(dev->tx_ring); i++)
mtk_wed_free_ring(dev, &dev->tx_ring[i]);
- for (i = 0; i < ARRAY_SIZE(dev->tx_wdma); i++)
- mtk_wed_free_ring(dev, &dev->tx_wdma[i]);
+ for (i = 0; i < ARRAY_SIZE(dev->rx_wdma); i++)
+ mtk_wed_free_ring(dev, &dev->rx_wdma[i]);
}
static void
@@ -688,10 +688,10 @@ mtk_wed_ring_alloc(struct mtk_wed_device
}
static int
-mtk_wed_wdma_ring_setup(struct mtk_wed_device *dev, int idx, int size)
+mtk_wed_wdma_rx_ring_setup(struct mtk_wed_device *dev, int idx, int size)
{
u32 desc_size = sizeof(struct mtk_wdma_desc) * dev->hw->version;
- struct mtk_wed_ring *wdma = &dev->tx_wdma[idx];
+ struct mtk_wed_ring *wdma = &dev->rx_wdma[idx];
if (mtk_wed_ring_alloc(dev, wdma, MTK_WED_WDMA_RING_SIZE, desc_size))
return -ENOMEM;
@@ -805,9 +805,9 @@ mtk_wed_start(struct mtk_wed_device *dev
{
int i;
- for (i = 0; i < ARRAY_SIZE(dev->tx_wdma); i++)
- if (!dev->tx_wdma[i].desc)
- mtk_wed_wdma_ring_setup(dev, i, 16);
+ for (i = 0; i < ARRAY_SIZE(dev->rx_wdma); i++)
+ if (!dev->rx_wdma[i].desc)
+ mtk_wed_wdma_rx_ring_setup(dev, i, 16);
mtk_wed_hw_init(dev);
mtk_wed_configure_irq(dev, irq_mask);
@@ -916,7 +916,7 @@ mtk_wed_tx_ring_setup(struct mtk_wed_dev
sizeof(*ring->desc)))
return -ENOMEM;
- if (mtk_wed_wdma_ring_setup(dev, idx, MTK_WED_WDMA_RING_SIZE))
+ if (mtk_wed_wdma_rx_ring_setup(dev, idx, MTK_WED_WDMA_RING_SIZE))
return -ENOMEM;
ring->reg_base = MTK_WED_RING_TX(idx);
--- a/include/linux/soc/mediatek/mtk_wed.h
+++ b/include/linux/soc/mediatek/mtk_wed.h
@@ -7,6 +7,7 @@
#include <linux/pci.h>
#define MTK_WED_TX_QUEUES 2
+#define MTK_WED_RX_QUEUES 2
struct mtk_wed_hw;
struct mtk_wdma_desc;
@@ -66,7 +67,7 @@ struct mtk_wed_device {
struct mtk_wed_ring tx_ring[MTK_WED_TX_QUEUES];
struct mtk_wed_ring txfree_ring;
- struct mtk_wed_ring tx_wdma[MTK_WED_TX_QUEUES];
+ struct mtk_wed_ring rx_wdma[MTK_WED_RX_QUEUES];
struct {
int size;

1521
target/linux/generic/backport-6.1/729-04-v6.1-net-ethernet-mtk_wed-add-configure-wed-wo-support.patch Normal file
View File

File diff suppressed because it is too large Load Diff

149
target/linux/generic/backport-6.1/729-05-v6.1-net-ethernet-mtk_wed-add-rx-mib-counters.patch Normal file
View File

@ -0,0 +1,149 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 5 Nov 2022 23:36:22 +0100
Subject: [PATCH] net: ethernet: mtk_wed: add rx mib counters
Introduce WED RX MIB counters support available on MT7986a SoC.
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed_debugfs.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed_debugfs.c
@@ -2,6 +2,7 @@
/* Copyright (C) 2021 Felix Fietkau <nbd@nbd.name> */
#include <linux/seq_file.h>
+#include <linux/soc/mediatek/mtk_wed.h>
#include "mtk_wed.h"
#include "mtk_wed_regs.h"
@@ -18,6 +19,8 @@ enum {
DUMP_TYPE_WDMA,
DUMP_TYPE_WPDMA_TX,
DUMP_TYPE_WPDMA_TXFREE,
+ DUMP_TYPE_WPDMA_RX,
+ DUMP_TYPE_WED_RRO,
};
#define DUMP_STR(_str) { _str, 0, DUMP_TYPE_STRING }
@@ -36,6 +39,9 @@ enum {
#define DUMP_WPDMA_TX_RING(_n) DUMP_RING("WPDMA_TX" #_n, 0, DUMP_TYPE_WPDMA_TX, _n)
#define DUMP_WPDMA_TXFREE_RING DUMP_RING("WPDMA_RX1", 0, DUMP_TYPE_WPDMA_TXFREE)
+#define DUMP_WPDMA_RX_RING(_n) DUMP_RING("WPDMA_RX" #_n, 0, DUMP_TYPE_WPDMA_RX, _n)
+#define DUMP_WED_RRO_RING(_base)DUMP_RING("WED_RRO_MIOD", MTK_##_base, DUMP_TYPE_WED_RRO)
+#define DUMP_WED_RRO_FDBK(_base)DUMP_RING("WED_RRO_FDBK", MTK_##_base, DUMP_TYPE_WED_RRO)
static void
print_reg_val(struct seq_file *s, const char *name, u32 val)
@@ -57,6 +63,7 @@ dump_wed_regs(struct seq_file *s, struct
cur > regs ? "\n" : "",
cur->name);
continue;
+ case DUMP_TYPE_WED_RRO:
case DUMP_TYPE_WED:
val = wed_r32(dev, cur->offset);
break;
@@ -69,6 +76,9 @@ dump_wed_regs(struct seq_file *s, struct
case DUMP_TYPE_WPDMA_TXFREE:
val = wpdma_txfree_r32(dev, cur->offset);
break;
+ case DUMP_TYPE_WPDMA_RX:
+ val = wpdma_rx_r32(dev, cur->base, cur->offset);
+ break;
}
print_reg_val(s, cur->name, val);
}
@@ -132,6 +142,80 @@ wed_txinfo_show(struct seq_file *s, void
}
DEFINE_SHOW_ATTRIBUTE(wed_txinfo);
+static int
+wed_rxinfo_show(struct seq_file *s, void *data)
+{
+ static const struct reg_dump regs[] = {
+ DUMP_STR("WPDMA RX"),
+ DUMP_WPDMA_RX_RING(0),
+ DUMP_WPDMA_RX_RING(1),
+
+ DUMP_STR("WPDMA RX"),
+ DUMP_WED(WED_WPDMA_RX_D_MIB(0)),
+ DUMP_WED_RING(WED_WPDMA_RING_RX_DATA(0)),
+ DUMP_WED(WED_WPDMA_RX_D_PROCESSED_MIB(0)),
+ DUMP_WED(WED_WPDMA_RX_D_MIB(1)),
+ DUMP_WED_RING(WED_WPDMA_RING_RX_DATA(1)),
+ DUMP_WED(WED_WPDMA_RX_D_PROCESSED_MIB(1)),
+ DUMP_WED(WED_WPDMA_RX_D_COHERENT_MIB),
+
+ DUMP_STR("WED RX"),
+ DUMP_WED_RING(WED_RING_RX_DATA(0)),
+ DUMP_WED_RING(WED_RING_RX_DATA(1)),
+
+ DUMP_STR("WED RRO"),
+ DUMP_WED_RRO_RING(WED_RROQM_MIOD_CTRL0),
+ DUMP_WED(WED_RROQM_MID_MIB),
+ DUMP_WED(WED_RROQM_MOD_MIB),
+ DUMP_WED(WED_RROQM_MOD_COHERENT_MIB),
+ DUMP_WED_RRO_FDBK(WED_RROQM_FDBK_CTRL0),
+ DUMP_WED(WED_RROQM_FDBK_IND_MIB),
+ DUMP_WED(WED_RROQM_FDBK_ENQ_MIB),
+ DUMP_WED(WED_RROQM_FDBK_ANC_MIB),
+ DUMP_WED(WED_RROQM_FDBK_ANC2H_MIB),
+
+ DUMP_STR("WED Route QM"),
+ DUMP_WED(WED_RTQM_R2H_MIB(0)),
+ DUMP_WED(WED_RTQM_R2Q_MIB(0)),
+ DUMP_WED(WED_RTQM_Q2H_MIB(0)),
+ DUMP_WED(WED_RTQM_R2H_MIB(1)),
+ DUMP_WED(WED_RTQM_R2Q_MIB(1)),
+ DUMP_WED(WED_RTQM_Q2H_MIB(1)),
+ DUMP_WED(WED_RTQM_Q2N_MIB),
+ DUMP_WED(WED_RTQM_Q2B_MIB),
+ DUMP_WED(WED_RTQM_PFDBK_MIB),
+
+ DUMP_STR("WED WDMA TX"),
+ DUMP_WED(WED_WDMA_TX_MIB),
+ DUMP_WED_RING(WED_WDMA_RING_TX),
+
+ DUMP_STR("WDMA TX"),
+ DUMP_WDMA(WDMA_GLO_CFG),
+ DUMP_WDMA_RING(WDMA_RING_TX(0)),
+ DUMP_WDMA_RING(WDMA_RING_TX(1)),
+
+ DUMP_STR("WED RX BM"),
+ DUMP_WED(WED_RX_BM_BASE),
+ DUMP_WED(WED_RX_BM_RX_DMAD),
+ DUMP_WED(WED_RX_BM_PTR),
+ DUMP_WED(WED_RX_BM_TKID_MIB),
+ DUMP_WED(WED_RX_BM_BLEN),
+ DUMP_WED(WED_RX_BM_STS),
+ DUMP_WED(WED_RX_BM_INTF2),
+ DUMP_WED(WED_RX_BM_INTF),
+ DUMP_WED(WED_RX_BM_ERR_STS),
+ };
+ struct mtk_wed_hw *hw = s->private;
+ struct mtk_wed_device *dev = hw->wed_dev;
+
+ if (!dev)
+ return 0;
+
+ dump_wed_regs(s, dev, regs, ARRAY_SIZE(regs));
+
+ return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(wed_rxinfo);
static int
mtk_wed_reg_set(void *data, u64 val)
@@ -175,4 +259,7 @@ void mtk_wed_hw_add_debugfs(struct mtk_w
debugfs_create_u32("regidx", 0600, dir, &hw->debugfs_reg);
debugfs_create_file_unsafe("regval", 0600, dir, hw, &fops_regval);
debugfs_create_file_unsafe("txinfo", 0400, dir, hw, &wed_txinfo_fops);
+ if (hw->version != 1)
+ debugfs_create_file_unsafe("rxinfo", 0400, dir, hw,
+ &wed_rxinfo_fops);
}

36
target/linux/generic/backport-6.1/729-07-v6.1-net-ethernet-mtk_eth_soc-remove-cpu_relax-in-mtk_pen.patch Normal file
View File

@ -0,0 +1,36 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 17 Nov 2022 00:58:46 +0100
Subject: [PATCH] net: ethernet: mtk_eth_soc: remove cpu_relax in
mtk_pending_work
Get rid of cpu_relax in mtk_pending_work routine since MTK_RESETTING is
set only in mtk_pending_work() and it runs holding rtnl lock
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -3495,11 +3495,8 @@ static void mtk_pending_work(struct work
rtnl_lock();
dev_dbg(eth->dev, "[%s][%d] reset\n", __func__, __LINE__);
+ set_bit(MTK_RESETTING, &eth->state);
- while (test_and_set_bit_lock(MTK_RESETTING, &eth->state))
- cpu_relax();
-
- dev_dbg(eth->dev, "[%s][%d] mtk_stop starts\n", __func__, __LINE__);
/* stop all devices to make sure that dma is properly shut down */
for (i = 0; i < MTK_MAC_COUNT; i++) {
if (!eth->netdev[i])
@@ -3533,7 +3530,7 @@ static void mtk_pending_work(struct work
dev_dbg(eth->dev, "[%s][%d] reset done\n", __func__, __LINE__);
- clear_bit_unlock(MTK_RESETTING, &eth->state);
+ clear_bit(MTK_RESETTING, &eth->state);
rtnl_unlock();
}

80
target/linux/generic/backport-6.1/729-09-v6.2-net-ethernet-mtk_wed-add-wcid-overwritten-support-fo.patch Normal file
View File

@ -0,0 +1,80 @@
From: Sujuan Chen <sujuan.chen@mediatek.com>
Date: Thu, 24 Nov 2022 11:18:14 +0800
Subject: [PATCH] net: ethernet: mtk_wed: add wcid overwritten support for wed
v1
All wed versions should enable the wcid overwritten feature,
since the wcid size is controlled by the wlan driver.
Tested-by: Sujuan Chen <sujuan.chen@mediatek.com>
Co-developed-by: Bo Jiao <bo.jiao@mediatek.com>
Signed-off-by: Bo Jiao <bo.jiao@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -526,9 +526,9 @@ mtk_wed_dma_disable(struct mtk_wed_devic
MTK_WED_WPDMA_RX_D_RX_DRV_EN);
wed_clr(dev, MTK_WED_WDMA_GLO_CFG,
MTK_WED_WDMA_GLO_CFG_TX_DDONE_CHK);
-
- mtk_wed_set_512_support(dev, false);
}
+
+ mtk_wed_set_512_support(dev, false);
}
static void
@@ -1290,9 +1290,10 @@ mtk_wed_start(struct mtk_wed_device *dev
if (mtk_wed_rro_cfg(dev))
return;
- mtk_wed_set_512_support(dev, dev->wlan.wcid_512);
}
+ mtk_wed_set_512_support(dev, dev->wlan.wcid_512);
+
mtk_wed_dma_enable(dev);
dev->running = true;
}
@@ -1358,11 +1359,13 @@ mtk_wed_attach(struct mtk_wed_device *de
}
mtk_wed_hw_init_early(dev);
- if (hw->version == 1)
+ if (hw->version == 1) {
regmap_update_bits(hw->hifsys, HIFSYS_DMA_AG_MAP,
BIT(hw->index), 0);
- else
+ } else {
+ dev->rev_id = wed_r32(dev, MTK_WED_REV_ID);
ret = mtk_wed_wo_init(hw);
+ }
out:
if (ret)
mtk_wed_detach(dev);
--- a/drivers/net/ethernet/mediatek/mtk_wed_regs.h
+++ b/drivers/net/ethernet/mediatek/mtk_wed_regs.h
@@ -20,6 +20,8 @@ struct mtk_wdma_desc {
__le32 info;
} __packed __aligned(4);
+#define MTK_WED_REV_ID 0x004
+
#define MTK_WED_RESET 0x008
#define MTK_WED_RESET_TX_BM BIT(0)
#define MTK_WED_RESET_TX_FREE_AGENT BIT(4)
--- a/include/linux/soc/mediatek/mtk_wed.h
+++ b/include/linux/soc/mediatek/mtk_wed.h
@@ -85,6 +85,9 @@ struct mtk_wed_device {
int irq;
u8 version;
+ /* used by wlan driver */
+ u32 rev_id;
+
struct mtk_wed_ring tx_ring[MTK_WED_TX_QUEUES];
struct mtk_wed_ring rx_ring[MTK_WED_RX_QUEUES];
struct mtk_wed_ring txfree_ring;

85
target/linux/generic/backport-6.1/729-10-v6.2-net-ethernet-mtk_wed-return-status-value-in-mtk_wdma.patch Normal file
View File

@ -0,0 +1,85 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 24 Nov 2022 16:22:51 +0100
Subject: [PATCH] net: ethernet: mtk_wed: return status value in
mtk_wdma_rx_reset
Move MTK_WDMA_RESET_IDX configuration in mtk_wdma_rx_reset routine.
Increase poll timeout to 10ms in order to be aligned with vendor sdk.
This is a preliminary patch to add Wireless Ethernet Dispatcher reset
support.
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -101,17 +101,21 @@ mtk_wdma_read_reset(struct mtk_wed_devic
return wdma_r32(dev, MTK_WDMA_GLO_CFG);
}
-static void
+static int
mtk_wdma_rx_reset(struct mtk_wed_device *dev)
{
u32 status, mask = MTK_WDMA_GLO_CFG_RX_DMA_BUSY;
- int i;
+ int i, ret;
wdma_clr(dev, MTK_WDMA_GLO_CFG, MTK_WDMA_GLO_CFG_RX_DMA_EN);
- if (readx_poll_timeout(mtk_wdma_read_reset, dev, status,
- !(status & mask), 0, 1000))
+ ret = readx_poll_timeout(mtk_wdma_read_reset, dev, status,
+ !(status & mask), 0, 10000);
+ if (ret)
dev_err(dev->hw->dev, "rx reset failed\n");
+ wdma_w32(dev, MTK_WDMA_RESET_IDX, MTK_WDMA_RESET_IDX_RX);
+ wdma_w32(dev, MTK_WDMA_RESET_IDX, 0);
+
for (i = 0; i < ARRAY_SIZE(dev->rx_wdma); i++) {
if (dev->rx_wdma[i].desc)
continue;
@@ -119,6 +123,8 @@ mtk_wdma_rx_reset(struct mtk_wed_device
wdma_w32(dev,
MTK_WDMA_RING_RX(i) + MTK_WED_RING_OFS_CPU_IDX, 0);
}
+
+ return ret;
}
static void
@@ -565,9 +571,7 @@ mtk_wed_detach(struct mtk_wed_device *de
mtk_wed_stop(dev);
- wdma_w32(dev, MTK_WDMA_RESET_IDX, MTK_WDMA_RESET_IDX_RX);
- wdma_w32(dev, MTK_WDMA_RESET_IDX, 0);
-
+ mtk_wdma_rx_reset(dev);
mtk_wed_reset(dev, MTK_WED_RESET_WED);
if (mtk_wed_get_rx_capa(dev)) {
wdma_clr(dev, MTK_WDMA_GLO_CFG, MTK_WDMA_GLO_CFG_TX_DMA_EN);
@@ -582,7 +586,6 @@ mtk_wed_detach(struct mtk_wed_device *de
mtk_wed_wo_reset(dev);
mtk_wed_free_rx_rings(dev);
mtk_wed_wo_deinit(hw);
- mtk_wdma_rx_reset(dev);
}
if (dev->wlan.bus_type == MTK_WED_BUS_PCIE) {
@@ -999,11 +1002,7 @@ mtk_wed_reset_dma(struct mtk_wed_device
wed_w32(dev, MTK_WED_RESET_IDX, 0);
}
- wdma_w32(dev, MTK_WDMA_RESET_IDX, MTK_WDMA_RESET_IDX_RX);
- wdma_w32(dev, MTK_WDMA_RESET_IDX, 0);
-
- if (mtk_wed_get_rx_capa(dev))
- mtk_wdma_rx_reset(dev);
+ mtk_wdma_rx_reset(dev);
if (busy) {
mtk_wed_reset(dev, MTK_WED_RESET_WDMA_INT_AGENT);

52
target/linux/generic/backport-6.1/729-11-v6.2-net-ethernet-mtk_wed-move-MTK_WDMA_RESET_IDX_TX-conf.patch Normal file
View File

@ -0,0 +1,52 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 24 Nov 2022 16:22:52 +0100
Subject: [PATCH] net: ethernet: mtk_wed: move MTK_WDMA_RESET_IDX_TX
configuration in mtk_wdma_tx_reset
Remove duplicated code. Increase poll timeout to 10ms in order to be
aligned with vendor sdk.
This is a preliminary patch to add Wireless Ethernet Dispatcher reset
support.
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -135,16 +135,15 @@ mtk_wdma_tx_reset(struct mtk_wed_device
wdma_clr(dev, MTK_WDMA_GLO_CFG, MTK_WDMA_GLO_CFG_TX_DMA_EN);
if (readx_poll_timeout(mtk_wdma_read_reset, dev, status,
- !(status & mask), 0, 1000))
+ !(status & mask), 0, 10000))
dev_err(dev->hw->dev, "tx reset failed\n");
- for (i = 0; i < ARRAY_SIZE(dev->tx_wdma); i++) {
- if (dev->tx_wdma[i].desc)
- continue;
+ wdma_w32(dev, MTK_WDMA_RESET_IDX, MTK_WDMA_RESET_IDX_TX);
+ wdma_w32(dev, MTK_WDMA_RESET_IDX, 0);
+ for (i = 0; i < ARRAY_SIZE(dev->tx_wdma); i++)
wdma_w32(dev,
MTK_WDMA_RING_TX(i) + MTK_WED_RING_OFS_CPU_IDX, 0);
- }
}
static void
@@ -573,12 +572,6 @@ mtk_wed_detach(struct mtk_wed_device *de
mtk_wdma_rx_reset(dev);
mtk_wed_reset(dev, MTK_WED_RESET_WED);
- if (mtk_wed_get_rx_capa(dev)) {
- wdma_clr(dev, MTK_WDMA_GLO_CFG, MTK_WDMA_GLO_CFG_TX_DMA_EN);
- wdma_w32(dev, MTK_WDMA_RESET_IDX, MTK_WDMA_RESET_IDX_TX);
- wdma_w32(dev, MTK_WDMA_RESET_IDX, 0);
- }
-
mtk_wed_free_tx_buffer(dev);
mtk_wed_free_tx_rings(dev);

98
target/linux/generic/backport-6.1/729-12-v6.2-net-ethernet-mtk_wed-update-mtk_wed_stop.patch Normal file
View File

@ -0,0 +1,98 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 24 Nov 2022 16:22:53 +0100
Subject: [PATCH] net: ethernet: mtk_wed: update mtk_wed_stop
Update mtk_wed_stop routine and rename old mtk_wed_stop() to
mtk_wed_deinit(). This is a preliminary patch to add Wireless Ethernet
Dispatcher reset support.
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -539,14 +539,8 @@ mtk_wed_dma_disable(struct mtk_wed_devic
static void
mtk_wed_stop(struct mtk_wed_device *dev)
{
- mtk_wed_dma_disable(dev);
mtk_wed_set_ext_int(dev, false);
- wed_clr(dev, MTK_WED_CTRL,
- MTK_WED_CTRL_WDMA_INT_AGENT_EN |
- MTK_WED_CTRL_WPDMA_INT_AGENT_EN |
- MTK_WED_CTRL_WED_TX_BM_EN |
- MTK_WED_CTRL_WED_TX_FREE_AGENT_EN);
wed_w32(dev, MTK_WED_WPDMA_INT_TRIGGER, 0);
wed_w32(dev, MTK_WED_WDMA_INT_TRIGGER, 0);
wdma_w32(dev, MTK_WDMA_INT_MASK, 0);
@@ -558,7 +552,27 @@ mtk_wed_stop(struct mtk_wed_device *dev)
wed_w32(dev, MTK_WED_EXT_INT_MASK1, 0);
wed_w32(dev, MTK_WED_EXT_INT_MASK2, 0);
- wed_clr(dev, MTK_WED_CTRL, MTK_WED_CTRL_WED_RX_BM_EN);
+}
+
+static void
+mtk_wed_deinit(struct mtk_wed_device *dev)
+{
+ mtk_wed_stop(dev);
+ mtk_wed_dma_disable(dev);
+
+ wed_clr(dev, MTK_WED_CTRL,
+ MTK_WED_CTRL_WDMA_INT_AGENT_EN |
+ MTK_WED_CTRL_WPDMA_INT_AGENT_EN |
+ MTK_WED_CTRL_WED_TX_BM_EN |
+ MTK_WED_CTRL_WED_TX_FREE_AGENT_EN);
+
+ if (dev->hw->version == 1)
+ return;
+
+ wed_clr(dev, MTK_WED_CTRL,
+ MTK_WED_CTRL_RX_ROUTE_QM_EN |
+ MTK_WED_CTRL_WED_RX_BM_EN |
+ MTK_WED_CTRL_RX_RRO_QM_EN);
}
static void
@@ -568,7 +582,7 @@ mtk_wed_detach(struct mtk_wed_device *de
mutex_lock(&hw_lock);
- mtk_wed_stop(dev);
+ mtk_wed_deinit(dev);
mtk_wdma_rx_reset(dev);
mtk_wed_reset(dev, MTK_WED_RESET_WED);
@@ -670,7 +684,7 @@ mtk_wed_hw_init_early(struct mtk_wed_dev
{
u32 mask, set;
- mtk_wed_stop(dev);
+ mtk_wed_deinit(dev);
mtk_wed_reset(dev, MTK_WED_RESET_WED);
mtk_wed_set_wpdma(dev);
--- a/include/linux/soc/mediatek/mtk_wed.h
+++ b/include/linux/soc/mediatek/mtk_wed.h
@@ -234,6 +234,8 @@ mtk_wed_get_rx_capa(struct mtk_wed_devic
(_dev)->ops->ppe_check(_dev, _skb, _reason, _hash)
#define mtk_wed_device_update_msg(_dev, _id, _msg, _len) \
(_dev)->ops->msg_update(_dev, _id, _msg, _len)
+#define mtk_wed_device_stop(_dev) (_dev)->ops->stop(_dev)
+#define mtk_wed_device_dma_reset(_dev) (_dev)->ops->reset_dma(_dev)
#else
static inline bool mtk_wed_device_active(struct mtk_wed_device *dev)
{
@@ -250,6 +252,8 @@ static inline bool mtk_wed_device_active
#define mtk_wed_device_rx_ring_setup(_dev, _ring, _regs) -ENODEV
#define mtk_wed_device_ppe_check(_dev, _skb, _reason, _hash) do {} while (0)
#define mtk_wed_device_update_msg(_dev, _id, _msg, _len) -ENODEV
+#define mtk_wed_device_stop(_dev) do {} while (0)
+#define mtk_wed_device_dma_reset(_dev) do {} while (0)
#endif
#endif

309
target/linux/generic/backport-6.1/729-13-v6.2-net-ethernet-mtk_wed-add-mtk_wed_rx_reset-routine.patch Normal file
View File

@ -0,0 +1,309 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 24 Nov 2022 16:22:54 +0100
Subject: [PATCH] net: ethernet: mtk_wed: add mtk_wed_rx_reset routine
Introduce mtk_wed_rx_reset routine in order to reset rx DMA for Wireless
Ethernet Dispatcher available on MT7986 SoC.
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -944,42 +944,130 @@ mtk_wed_ring_reset(struct mtk_wed_ring *
}
static u32
-mtk_wed_check_busy(struct mtk_wed_device *dev)
+mtk_wed_check_busy(struct mtk_wed_device *dev, u32 reg, u32 mask)
{
- if (wed_r32(dev, MTK_WED_GLO_CFG) & MTK_WED_GLO_CFG_TX_DMA_BUSY)
- return true;
-
- if (wed_r32(dev, MTK_WED_WPDMA_GLO_CFG) &
- MTK_WED_WPDMA_GLO_CFG_TX_DRV_BUSY)
- return true;
-
- if (wed_r32(dev, MTK_WED_CTRL) & MTK_WED_CTRL_WDMA_INT_AGENT_BUSY)
- return true;
-
- if (wed_r32(dev, MTK_WED_WDMA_GLO_CFG) &
- MTK_WED_WDMA_GLO_CFG_RX_DRV_BUSY)
- return true;
-
- if (wdma_r32(dev, MTK_WDMA_GLO_CFG) &
- MTK_WED_WDMA_GLO_CFG_RX_DRV_BUSY)
- return true;
-
- if (wed_r32(dev, MTK_WED_CTRL) &
- (MTK_WED_CTRL_WED_TX_BM_BUSY | MTK_WED_CTRL_WED_TX_FREE_AGENT_BUSY))
- return true;
-
- return false;
+ return !!(wed_r32(dev, reg) & mask);
}
static int
-mtk_wed_poll_busy(struct mtk_wed_device *dev)
+mtk_wed_poll_busy(struct mtk_wed_device *dev, u32 reg, u32 mask)
{
int sleep = 15000;
int timeout = 100 * sleep;
u32 val;
return read_poll_timeout(mtk_wed_check_busy, val, !val, sleep,
- timeout, false, dev);
+ timeout, false, dev, reg, mask);
+}
+
+static int
+mtk_wed_rx_reset(struct mtk_wed_device *dev)
+{
+ struct mtk_wed_wo *wo = dev->hw->wed_wo;
+ u8 val = MTK_WED_WO_STATE_SER_RESET;
+ int i, ret;
+
+ ret = mtk_wed_mcu_send_msg(wo, MTK_WED_MODULE_ID_WO,
+ MTK_WED_WO_CMD_CHANGE_STATE, &val,
+ sizeof(val), true);
+ if (ret)
+ return ret;
+
+ wed_clr(dev, MTK_WED_WPDMA_RX_D_GLO_CFG, MTK_WED_WPDMA_RX_D_RX_DRV_EN);
+ ret = mtk_wed_poll_busy(dev, MTK_WED_WPDMA_RX_D_GLO_CFG,
+ MTK_WED_WPDMA_RX_D_RX_DRV_BUSY);
+ if (ret) {
+ mtk_wed_reset(dev, MTK_WED_RESET_WPDMA_INT_AGENT);
+ mtk_wed_reset(dev, MTK_WED_RESET_WPDMA_RX_D_DRV);
+ } else {
+ wed_w32(dev, MTK_WED_WPDMA_RX_D_RST_IDX,
+ MTK_WED_WPDMA_RX_D_RST_CRX_IDX |
+ MTK_WED_WPDMA_RX_D_RST_DRV_IDX);
+
+ wed_set(dev, MTK_WED_WPDMA_RX_D_GLO_CFG,
+ MTK_WED_WPDMA_RX_D_RST_INIT_COMPLETE |
+ MTK_WED_WPDMA_RX_D_FSM_RETURN_IDLE);
+ wed_clr(dev, MTK_WED_WPDMA_RX_D_GLO_CFG,
+ MTK_WED_WPDMA_RX_D_RST_INIT_COMPLETE |
+ MTK_WED_WPDMA_RX_D_FSM_RETURN_IDLE);
+
+ wed_w32(dev, MTK_WED_WPDMA_RX_D_RST_IDX, 0);
+ }
+
+ /* reset rro qm */
+ wed_clr(dev, MTK_WED_CTRL, MTK_WED_CTRL_RX_RRO_QM_EN);
+ ret = mtk_wed_poll_busy(dev, MTK_WED_CTRL,
+ MTK_WED_CTRL_RX_RRO_QM_BUSY);
+ if (ret) {
+ mtk_wed_reset(dev, MTK_WED_RESET_RX_RRO_QM);
+ } else {
+ wed_set(dev, MTK_WED_RROQM_RST_IDX,
+ MTK_WED_RROQM_RST_IDX_MIOD |
+ MTK_WED_RROQM_RST_IDX_FDBK);
+ wed_w32(dev, MTK_WED_RROQM_RST_IDX, 0);
+ }
+
+ /* reset route qm */
+ wed_clr(dev, MTK_WED_CTRL, MTK_WED_CTRL_RX_ROUTE_QM_EN);
+ ret = mtk_wed_poll_busy(dev, MTK_WED_CTRL,
+ MTK_WED_CTRL_RX_ROUTE_QM_BUSY);
+ if (ret)
+ mtk_wed_reset(dev, MTK_WED_RESET_RX_ROUTE_QM);
+ else
+ wed_set(dev, MTK_WED_RTQM_GLO_CFG,
+ MTK_WED_RTQM_Q_RST);
+
+ /* reset tx wdma */
+ mtk_wdma_tx_reset(dev);
+
+ /* reset tx wdma drv */
+ wed_clr(dev, MTK_WED_WDMA_GLO_CFG, MTK_WED_WDMA_GLO_CFG_TX_DRV_EN);
+ mtk_wed_poll_busy(dev, MTK_WED_CTRL,
+ MTK_WED_CTRL_WDMA_INT_AGENT_BUSY);
+ mtk_wed_reset(dev, MTK_WED_RESET_WDMA_TX_DRV);
+
+ /* reset wed rx dma */
+ ret = mtk_wed_poll_busy(dev, MTK_WED_GLO_CFG,
+ MTK_WED_GLO_CFG_RX_DMA_BUSY);
+ wed_clr(dev, MTK_WED_GLO_CFG, MTK_WED_GLO_CFG_RX_DMA_EN);
+ if (ret) {
+ mtk_wed_reset(dev, MTK_WED_RESET_WED_RX_DMA);
+ } else {
+ struct mtk_eth *eth = dev->hw->eth;
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2))
+ wed_set(dev, MTK_WED_RESET_IDX,
+ MTK_WED_RESET_IDX_RX_V2);
+ else
+ wed_set(dev, MTK_WED_RESET_IDX, MTK_WED_RESET_IDX_RX);
+ wed_w32(dev, MTK_WED_RESET_IDX, 0);
+ }
+
+ /* reset rx bm */
+ wed_clr(dev, MTK_WED_CTRL, MTK_WED_CTRL_WED_RX_BM_EN);
+ mtk_wed_poll_busy(dev, MTK_WED_CTRL,
+ MTK_WED_CTRL_WED_RX_BM_BUSY);
+ mtk_wed_reset(dev, MTK_WED_RESET_RX_BM);
+
+ /* wo change to enable state */
+ val = MTK_WED_WO_STATE_ENABLE;
+ ret = mtk_wed_mcu_send_msg(wo, MTK_WED_MODULE_ID_WO,
+ MTK_WED_WO_CMD_CHANGE_STATE, &val,
+ sizeof(val), true);
+ if (ret)
+ return ret;
+
+ /* wed_rx_ring_reset */
+ for (i = 0; i < ARRAY_SIZE(dev->rx_ring); i++) {
+ if (!dev->rx_ring[i].desc)
+ continue;
+
+ mtk_wed_ring_reset(&dev->rx_ring[i], MTK_WED_RX_RING_SIZE,
+ false);
+ }
+ mtk_wed_free_rx_buffer(dev);
+
+ return 0;
}
static void
@@ -997,19 +1085,23 @@ mtk_wed_reset_dma(struct mtk_wed_device
true);
}
- if (mtk_wed_poll_busy(dev))
- busy = mtk_wed_check_busy(dev);
-
+ /* 1. reset WED tx DMA */
+ wed_clr(dev, MTK_WED_GLO_CFG, MTK_WED_GLO_CFG_TX_DMA_EN);
+ busy = mtk_wed_poll_busy(dev, MTK_WED_GLO_CFG,
+ MTK_WED_GLO_CFG_TX_DMA_BUSY);
if (busy) {
mtk_wed_reset(dev, MTK_WED_RESET_WED_TX_DMA);
} else {
- wed_w32(dev, MTK_WED_RESET_IDX,
- MTK_WED_RESET_IDX_TX |
- MTK_WED_RESET_IDX_RX);
+ wed_w32(dev, MTK_WED_RESET_IDX, MTK_WED_RESET_IDX_TX);
wed_w32(dev, MTK_WED_RESET_IDX, 0);
}
- mtk_wdma_rx_reset(dev);
+ /* 2. reset WDMA rx DMA */
+ busy = !!mtk_wdma_rx_reset(dev);
+ wed_clr(dev, MTK_WED_WDMA_GLO_CFG, MTK_WED_WDMA_GLO_CFG_RX_DRV_EN);
+ if (!busy)
+ busy = mtk_wed_poll_busy(dev, MTK_WED_WDMA_GLO_CFG,
+ MTK_WED_WDMA_GLO_CFG_RX_DRV_BUSY);
if (busy) {
mtk_wed_reset(dev, MTK_WED_RESET_WDMA_INT_AGENT);
@@ -1026,6 +1118,9 @@ mtk_wed_reset_dma(struct mtk_wed_device
MTK_WED_WDMA_GLO_CFG_RST_INIT_COMPLETE);
}
+ /* 3. reset WED WPDMA tx */
+ wed_clr(dev, MTK_WED_CTRL, MTK_WED_CTRL_WED_TX_FREE_AGENT_EN);
+
for (i = 0; i < 100; i++) {
val = wed_r32(dev, MTK_WED_TX_BM_INTF);
if (FIELD_GET(MTK_WED_TX_BM_INTF_TKFIFO_FDEP, val) == 0x40)
@@ -1033,8 +1128,19 @@ mtk_wed_reset_dma(struct mtk_wed_device
}
mtk_wed_reset(dev, MTK_WED_RESET_TX_FREE_AGENT);
+ wed_clr(dev, MTK_WED_CTRL, MTK_WED_CTRL_WED_TX_BM_EN);
mtk_wed_reset(dev, MTK_WED_RESET_TX_BM);
+ /* 4. reset WED WPDMA tx */
+ busy = mtk_wed_poll_busy(dev, MTK_WED_WPDMA_GLO_CFG,
+ MTK_WED_WPDMA_GLO_CFG_TX_DRV_BUSY);
+ wed_clr(dev, MTK_WED_WPDMA_GLO_CFG,
+ MTK_WED_WPDMA_GLO_CFG_TX_DRV_EN |
+ MTK_WED_WPDMA_GLO_CFG_RX_DRV_EN);
+ if (!busy)
+ busy = mtk_wed_poll_busy(dev, MTK_WED_WPDMA_GLO_CFG,
+ MTK_WED_WPDMA_GLO_CFG_RX_DRV_BUSY);
+
if (busy) {
mtk_wed_reset(dev, MTK_WED_RESET_WPDMA_INT_AGENT);
mtk_wed_reset(dev, MTK_WED_RESET_WPDMA_TX_DRV);
@@ -1045,6 +1151,17 @@ mtk_wed_reset_dma(struct mtk_wed_device
MTK_WED_WPDMA_RESET_IDX_RX);
wed_w32(dev, MTK_WED_WPDMA_RESET_IDX, 0);
}
+
+ dev->init_done = false;
+ if (dev->hw->version == 1)
+ return;
+
+ if (!busy) {
+ wed_w32(dev, MTK_WED_RESET_IDX, MTK_WED_RESET_WPDMA_IDX_RX);
+ wed_w32(dev, MTK_WED_RESET_IDX, 0);
+ }
+
+ mtk_wed_rx_reset(dev);
}
static int
@@ -1267,6 +1384,9 @@ mtk_wed_start(struct mtk_wed_device *dev
{
int i;
+ if (mtk_wed_get_rx_capa(dev) && mtk_wed_rx_buffer_alloc(dev))
+ return;
+
for (i = 0; i < ARRAY_SIZE(dev->rx_wdma); i++)
if (!dev->rx_wdma[i].desc)
mtk_wed_wdma_rx_ring_setup(dev, i, 16);
@@ -1355,10 +1475,6 @@ mtk_wed_attach(struct mtk_wed_device *de
goto out;
if (mtk_wed_get_rx_capa(dev)) {
- ret = mtk_wed_rx_buffer_alloc(dev);
- if (ret)
- goto out;
-
ret = mtk_wed_rro_alloc(dev);
if (ret)
goto out;
--- a/drivers/net/ethernet/mediatek/mtk_wed_regs.h
+++ b/drivers/net/ethernet/mediatek/mtk_wed_regs.h
@@ -24,11 +24,15 @@ struct mtk_wdma_desc {
#define MTK_WED_RESET 0x008
#define MTK_WED_RESET_TX_BM BIT(0)
+#define MTK_WED_RESET_RX_BM BIT(1)
#define MTK_WED_RESET_TX_FREE_AGENT BIT(4)
#define MTK_WED_RESET_WPDMA_TX_DRV BIT(8)
#define MTK_WED_RESET_WPDMA_RX_DRV BIT(9)
+#define MTK_WED_RESET_WPDMA_RX_D_DRV BIT(10)
#define MTK_WED_RESET_WPDMA_INT_AGENT BIT(11)
#define MTK_WED_RESET_WED_TX_DMA BIT(12)
+#define MTK_WED_RESET_WED_RX_DMA BIT(13)
+#define MTK_WED_RESET_WDMA_TX_DRV BIT(16)
#define MTK_WED_RESET_WDMA_RX_DRV BIT(17)
#define MTK_WED_RESET_WDMA_INT_AGENT BIT(19)
#define MTK_WED_RESET_RX_RRO_QM BIT(20)
@@ -158,6 +162,8 @@ struct mtk_wdma_desc {
#define MTK_WED_RESET_IDX 0x20c
#define MTK_WED_RESET_IDX_TX GENMASK(3, 0)
#define MTK_WED_RESET_IDX_RX GENMASK(17, 16)
+#define MTK_WED_RESET_IDX_RX_V2 GENMASK(7, 6)
+#define MTK_WED_RESET_WPDMA_IDX_RX GENMASK(31, 30)
#define MTK_WED_TX_MIB(_n) (0x2a0 + (_n) * 4)
#define MTK_WED_RX_MIB(_n) (0x2e0 + (_n) * 4)
@@ -267,6 +273,9 @@ struct mtk_wdma_desc {
#define MTK_WED_WPDMA_RX_D_GLO_CFG 0x75c
#define MTK_WED_WPDMA_RX_D_RX_DRV_EN BIT(0)
+#define MTK_WED_WPDMA_RX_D_RX_DRV_BUSY BIT(1)
+#define MTK_WED_WPDMA_RX_D_FSM_RETURN_IDLE BIT(3)
+#define MTK_WED_WPDMA_RX_D_RST_INIT_COMPLETE BIT(4)
#define MTK_WED_WPDMA_RX_D_INIT_PHASE_RXEN_SEL GENMASK(11, 7)
#define MTK_WED_WPDMA_RX_D_RXD_READ_LEN GENMASK(31, 24)

103
target/linux/generic/backport-6.1/729-14-v6.2-net-ethernet-mtk_wed-add-reset-to-tx_ring_setup-call.patch Normal file
View File

@ -0,0 +1,103 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 24 Nov 2022 16:22:55 +0100
Subject: [PATCH] net: ethernet: mtk_wed: add reset to tx_ring_setup callback
Introduce reset parameter to mtk_wed_tx_ring_setup signature.
This is a preliminary patch to add Wireless Ethernet Dispatcher reset
support.
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -1181,7 +1181,8 @@ mtk_wed_ring_alloc(struct mtk_wed_device
}
static int
-mtk_wed_wdma_rx_ring_setup(struct mtk_wed_device *dev, int idx, int size)
+mtk_wed_wdma_rx_ring_setup(struct mtk_wed_device *dev, int idx, int size,
+ bool reset)
{
u32 desc_size = sizeof(struct mtk_wdma_desc) * dev->hw->version;
struct mtk_wed_ring *wdma;
@@ -1190,8 +1191,8 @@ mtk_wed_wdma_rx_ring_setup(struct mtk_we
return -EINVAL;
wdma = &dev->rx_wdma[idx];
- if (mtk_wed_ring_alloc(dev, wdma, MTK_WED_WDMA_RING_SIZE, desc_size,
- true))
+ if (!reset && mtk_wed_ring_alloc(dev, wdma, MTK_WED_WDMA_RING_SIZE,
+ desc_size, true))
return -ENOMEM;
wdma_w32(dev, MTK_WDMA_RING_RX(idx) + MTK_WED_RING_OFS_BASE,
@@ -1389,7 +1390,7 @@ mtk_wed_start(struct mtk_wed_device *dev
for (i = 0; i < ARRAY_SIZE(dev->rx_wdma); i++)
if (!dev->rx_wdma[i].desc)
- mtk_wed_wdma_rx_ring_setup(dev, i, 16);
+ mtk_wed_wdma_rx_ring_setup(dev, i, 16, false);
mtk_wed_hw_init(dev);
mtk_wed_configure_irq(dev, irq_mask);
@@ -1498,7 +1499,8 @@ unlock:
}
static int
-mtk_wed_tx_ring_setup(struct mtk_wed_device *dev, int idx, void __iomem *regs)
+mtk_wed_tx_ring_setup(struct mtk_wed_device *dev, int idx, void __iomem *regs,
+ bool reset)
{
struct mtk_wed_ring *ring = &dev->tx_ring[idx];
@@ -1517,11 +1519,12 @@ mtk_wed_tx_ring_setup(struct mtk_wed_dev
if (WARN_ON(idx >= ARRAY_SIZE(dev->tx_ring)))
return -EINVAL;
- if (mtk_wed_ring_alloc(dev, ring, MTK_WED_TX_RING_SIZE,
- sizeof(*ring->desc), true))
+ if (!reset && mtk_wed_ring_alloc(dev, ring, MTK_WED_TX_RING_SIZE,
+ sizeof(*ring->desc), true))
return -ENOMEM;
- if (mtk_wed_wdma_rx_ring_setup(dev, idx, MTK_WED_WDMA_RING_SIZE))
+ if (mtk_wed_wdma_rx_ring_setup(dev, idx, MTK_WED_WDMA_RING_SIZE,
+ reset))
return -ENOMEM;
ring->reg_base = MTK_WED_RING_TX(idx);
--- a/include/linux/soc/mediatek/mtk_wed.h
+++ b/include/linux/soc/mediatek/mtk_wed.h
@@ -158,7 +158,7 @@ struct mtk_wed_device {
struct mtk_wed_ops {
int (*attach)(struct mtk_wed_device *dev);
int (*tx_ring_setup)(struct mtk_wed_device *dev, int ring,
- void __iomem *regs);
+ void __iomem *regs, bool reset);
int (*rx_ring_setup)(struct mtk_wed_device *dev, int ring,
void __iomem *regs);
int (*txfree_ring_setup)(struct mtk_wed_device *dev,
@@ -216,8 +216,8 @@ mtk_wed_get_rx_capa(struct mtk_wed_devic
#define mtk_wed_device_active(_dev) !!(_dev)->ops
#define mtk_wed_device_detach(_dev) (_dev)->ops->detach(_dev)
#define mtk_wed_device_start(_dev, _mask) (_dev)->ops->start(_dev, _mask)
-#define mtk_wed_device_tx_ring_setup(_dev, _ring, _regs) \
- (_dev)->ops->tx_ring_setup(_dev, _ring, _regs)
+#define mtk_wed_device_tx_ring_setup(_dev, _ring, _regs, _reset) \
+ (_dev)->ops->tx_ring_setup(_dev, _ring, _regs, _reset)
#define mtk_wed_device_txfree_ring_setup(_dev, _regs) \
(_dev)->ops->txfree_ring_setup(_dev, _regs)
#define mtk_wed_device_reg_read(_dev, _reg) \
@@ -243,7 +243,7 @@ static inline bool mtk_wed_device_active
}
#define mtk_wed_device_detach(_dev) do {} while (0)
#define mtk_wed_device_start(_dev, _mask) do {} while (0)
-#define mtk_wed_device_tx_ring_setup(_dev, _ring, _regs) -ENODEV
+#define mtk_wed_device_tx_ring_setup(_dev, _ring, _regs, _reset) -ENODEV
#define mtk_wed_device_txfree_ring_setup(_dev, _ring, _regs) -ENODEV
#define mtk_wed_device_reg_read(_dev, _reg) 0
#define mtk_wed_device_reg_write(_dev, _reg, _val) do {} while (0)

103
target/linux/generic/backport-6.1/729-15-v6.2-net-ethernet-mtk_wed-fix-sleep-while-atomic-in-mtk_w.patch Normal file
View File

@ -0,0 +1,103 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 1 Dec 2022 16:26:53 +0100
Subject: [PATCH] net: ethernet: mtk_wed: fix sleep while atomic in
mtk_wed_wo_queue_refill
In order to fix the following sleep while atomic bug always alloc pages
with GFP_ATOMIC in mtk_wed_wo_queue_refill since page_frag_alloc runs in
spin_lock critical section.
[ 9.049719] Hardware name: MediaTek MT7986a RFB (DT)
[ 9.054665] Call trace:
[ 9.057096] dump_backtrace+0x0/0x154
[ 9.060751] show_stack+0x14/0x1c
[ 9.064052] dump_stack_lvl+0x64/0x7c
[ 9.067702] dump_stack+0x14/0x2c
[ 9.071001] ___might_sleep+0xec/0x120
[ 9.074736] __might_sleep+0x4c/0x9c
[ 9.078296] __alloc_pages+0x184/0x2e4
[ 9.082030] page_frag_alloc_align+0x98/0x1ac
[ 9.086369] mtk_wed_wo_queue_refill+0x134/0x234
[ 9.090974] mtk_wed_wo_init+0x174/0x2c0
[ 9.094881] mtk_wed_attach+0x7c8/0x7e0
[ 9.098701] mt7915_mmio_wed_init+0x1f0/0x3a0 [mt7915e]
[ 9.103940] mt7915_pci_probe+0xec/0x3bc [mt7915e]
[ 9.108727] pci_device_probe+0xac/0x13c
[ 9.112638] really_probe.part.0+0x98/0x2f4
[ 9.116807] __driver_probe_device+0x94/0x13c
[ 9.121147] driver_probe_device+0x40/0x114
[ 9.125314] __driver_attach+0x7c/0x180
[ 9.129133] bus_for_each_dev+0x5c/0x90
[ 9.132953] driver_attach+0x20/0x2c
[ 9.136513] bus_add_driver+0x104/0x1fc
[ 9.140333] driver_register+0x74/0x120
[ 9.144153] __pci_register_driver+0x40/0x50
[ 9.148407] mt7915_init+0x5c/0x1000 [mt7915e]
[ 9.152848] do_one_initcall+0x40/0x25c
[ 9.156669] do_init_module+0x44/0x230
[ 9.160403] load_module+0x1f30/0x2750
[ 9.164135] __do_sys_init_module+0x150/0x200
[ 9.168475] __arm64_sys_init_module+0x18/0x20
[ 9.172901] invoke_syscall.constprop.0+0x4c/0xe0
[ 9.177589] do_el0_svc+0x48/0xe0
[ 9.180889] el0_svc+0x14/0x50
[ 9.183929] el0t_64_sync_handler+0x9c/0x120
[ 9.188183] el0t_64_sync+0x158/0x15c
Fixes: 799684448e3e ("net: ethernet: mtk_wed: introduce wed wo support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Link: https://lore.kernel.org/r/67ca94bdd3d9eaeb86e52b3050fbca0bcf7bb02f.1669908312.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed_wo.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed_wo.c
@@ -133,17 +133,18 @@ mtk_wed_wo_dequeue(struct mtk_wed_wo *wo
static int
mtk_wed_wo_queue_refill(struct mtk_wed_wo *wo, struct mtk_wed_wo_queue *q,
- gfp_t gfp, bool rx)
+ bool rx)
{
enum dma_data_direction dir = rx ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
int n_buf = 0;
spin_lock_bh(&q->lock);
while (q->queued < q->n_desc) {
- void *buf = page_frag_alloc(&q->cache, q->buf_size, gfp);
struct mtk_wed_wo_queue_entry *entry;
dma_addr_t addr;
+ void *buf;
+ buf = page_frag_alloc(&q->cache, q->buf_size, GFP_ATOMIC);
if (!buf)
break;
@@ -215,7 +216,7 @@ mtk_wed_wo_rx_run_queue(struct mtk_wed_w
mtk_wed_mcu_rx_unsolicited_event(wo, skb);
}
- if (mtk_wed_wo_queue_refill(wo, q, GFP_ATOMIC, true)) {
+ if (mtk_wed_wo_queue_refill(wo, q, true)) {
u32 index = (q->head - 1) % q->n_desc;
mtk_wed_wo_queue_kick(wo, q, index);
@@ -432,7 +433,7 @@ mtk_wed_wo_hardware_init(struct mtk_wed_
if (ret)
goto error;
- mtk_wed_wo_queue_refill(wo, &wo->q_tx, GFP_KERNEL, false);
+ mtk_wed_wo_queue_refill(wo, &wo->q_tx, false);
mtk_wed_wo_queue_reset(wo, &wo->q_tx);
regs.desc_base = MTK_WED_WO_CCIF_DUMMY5;
@@ -446,7 +447,7 @@ mtk_wed_wo_hardware_init(struct mtk_wed_
if (ret)
goto error;
- mtk_wed_wo_queue_refill(wo, &wo->q_rx, GFP_KERNEL, true);
+ mtk_wed_wo_queue_refill(wo, &wo->q_rx, true);
mtk_wed_wo_queue_reset(wo, &wo->q_rx);
/* rx queue irqmask */

52
target/linux/generic/backport-6.1/729-16-v6.3-net-ethernet-mtk_wed-get-rid-of-queue-lock-for-rx-qu.patch Normal file
View File

@ -0,0 +1,52 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Tue, 10 Jan 2023 10:31:26 +0100
Subject: [PATCH] net: ethernet: mtk_wed: get rid of queue lock for rx queue
Queue spinlock is currently held in mtk_wed_wo_queue_rx_clean and
mtk_wed_wo_queue_refill routines for MTK Wireless Ethernet Dispatcher
MCU rx queue. mtk_wed_wo_queue_refill() is running during initialization
and in rx tasklet while mtk_wed_wo_queue_rx_clean() is running in
mtk_wed_wo_hw_deinit() during hw de-init phase after rx tasklet has been
disabled. Since mtk_wed_wo_queue_rx_clean and mtk_wed_wo_queue_refill
routines can't run concurrently get rid of spinlock for mcu rx queue.
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Link: https://lore.kernel.org/r/36ec3b729542ea60898471d890796f745479ba32.1673342990.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed_wo.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed_wo.c
@@ -138,7 +138,6 @@ mtk_wed_wo_queue_refill(struct mtk_wed_w
enum dma_data_direction dir = rx ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
int n_buf = 0;
- spin_lock_bh(&q->lock);
while (q->queued < q->n_desc) {
struct mtk_wed_wo_queue_entry *entry;
dma_addr_t addr;
@@ -172,7 +171,6 @@ mtk_wed_wo_queue_refill(struct mtk_wed_w
q->queued++;
n_buf++;
}
- spin_unlock_bh(&q->lock);
return n_buf;
}
@@ -316,7 +314,6 @@ mtk_wed_wo_queue_rx_clean(struct mtk_wed
{
struct page *page;
- spin_lock_bh(&q->lock);
for (;;) {
void *buf = mtk_wed_wo_dequeue(wo, q, NULL, true);
@@ -325,7 +322,6 @@ mtk_wed_wo_queue_rx_clean(struct mtk_wed
skb_free_frag(buf);
}
- spin_unlock_bh(&q->lock);
if (!q->cache.va)
return;

75
target/linux/generic/backport-6.1/729-17-v6.3-net-ethernet-mtk_wed-get-rid-of-queue-lock-for-tx-qu.patch Normal file
View File

@ -0,0 +1,75 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 12 Jan 2023 10:21:29 +0100
Subject: [PATCH] net: ethernet: mtk_wed: get rid of queue lock for tx queue
Similar to MTK Wireless Ethernet Dispatcher (WED) MCU rx queue,
we do not need to protect WED MCU tx queue with a spin lock since
the tx queue is accessed in the two following routines:
- mtk_wed_wo_queue_tx_skb():
it is run at initialization and during mt7915 normal operation.
Moreover MCU messages are serialized through MCU mutex.
- mtk_wed_wo_queue_tx_clean():
it runs just at mt7915 driver module unload when no more messages
are sent to the MCU.
Remove tx queue spinlock.
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Link: https://lore.kernel.org/r/7bd0337b2a13ab1a63673b7c03fd35206b3b284e.1673515140.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed_wo.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed_wo.c
@@ -258,7 +258,6 @@ mtk_wed_wo_queue_alloc(struct mtk_wed_wo
int n_desc, int buf_size, int index,
struct mtk_wed_wo_queue_regs *regs)
{
- spin_lock_init(&q->lock);
q->regs = *regs;
q->n_desc = n_desc;
q->buf_size = buf_size;
@@ -290,7 +289,6 @@ mtk_wed_wo_queue_tx_clean(struct mtk_wed
struct page *page;
int i;
- spin_lock_bh(&q->lock);
for (i = 0; i < q->n_desc; i++) {
struct mtk_wed_wo_queue_entry *entry = &q->entry[i];
@@ -299,7 +297,6 @@ mtk_wed_wo_queue_tx_clean(struct mtk_wed
skb_free_frag(entry->buf);
entry->buf = NULL;
}
- spin_unlock_bh(&q->lock);
if (!q->cache.va)
return;
@@ -347,8 +344,6 @@ int mtk_wed_wo_queue_tx_skb(struct mtk_w
int ret = 0, index;
u32 ctrl;
- spin_lock_bh(&q->lock);
-
q->tail = mtk_wed_mmio_r32(wo, q->regs.dma_idx);
index = (q->head + 1) % q->n_desc;
if (q->tail == index) {
@@ -379,8 +374,6 @@ int mtk_wed_wo_queue_tx_skb(struct mtk_w
mtk_wed_wo_queue_kick(wo, q, q->head);
mtk_wed_wo_kickout(wo);
out:
- spin_unlock_bh(&q->lock);
-
dev_kfree_skb(skb);
return ret;
--- a/drivers/net/ethernet/mediatek/mtk_wed_wo.h
+++ b/drivers/net/ethernet/mediatek/mtk_wed_wo.h
@@ -211,7 +211,6 @@ struct mtk_wed_wo_queue {
struct mtk_wed_wo_queue_regs regs;
struct page_frag_cache cache;
- spinlock_t lock;
struct mtk_wed_wo_queue_desc *desc;
dma_addr_t desc_dma;

70
target/linux/generic/backport-6.1/729-18-v6.3-net-ethernet-mtk_eth_soc-introduce-mtk_hw_reset-util.patch Normal file
View File

@ -0,0 +1,70 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 14 Jan 2023 18:01:28 +0100
Subject: [PATCH] net: ethernet: mtk_eth_soc: introduce mtk_hw_reset utility
routine
This is a preliminary patch to add Wireless Ethernet Dispatcher reset
support.
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -3254,6 +3254,27 @@ static void mtk_set_mcr_max_rx(struct mt
mtk_w32(mac->hw, mcr_new, MTK_MAC_MCR(mac->id));
}
+static void mtk_hw_reset(struct mtk_eth *eth)
+{
+ u32 val;
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
+ regmap_write(eth->ethsys, ETHSYS_FE_RST_CHK_IDLE_EN, 0);
+ val = RSTCTRL_PPE0_V2;
+ } else {
+ val = RSTCTRL_PPE0;
+ }
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_RSTCTRL_PPE1))
+ val |= RSTCTRL_PPE1;
+
+ ethsys_reset(eth, RSTCTRL_ETH | RSTCTRL_FE | val);
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2))
+ regmap_write(eth->ethsys, ETHSYS_FE_RST_CHK_IDLE_EN,
+ 0x3ffffff);
+}
+
static int mtk_hw_init(struct mtk_eth *eth)
{
u32 dma_mask = ETHSYS_DMA_AG_MAP_PDMA | ETHSYS_DMA_AG_MAP_QDMA |
@@ -3293,22 +3314,9 @@ static int mtk_hw_init(struct mtk_eth *e
return 0;
}
- if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
- regmap_write(eth->ethsys, ETHSYS_FE_RST_CHK_IDLE_EN, 0);
- val = RSTCTRL_PPE0_V2;
- } else {
- val = RSTCTRL_PPE0;
- }
-
- if (MTK_HAS_CAPS(eth->soc->caps, MTK_RSTCTRL_PPE1))
- val |= RSTCTRL_PPE1;
-
- ethsys_reset(eth, RSTCTRL_ETH | RSTCTRL_FE | val);
+ mtk_hw_reset(eth);
if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
- regmap_write(eth->ethsys, ETHSYS_FE_RST_CHK_IDLE_EN,
- 0x3ffffff);
-
/* Set FE to PDMAv2 if necessary */
val = mtk_r32(eth, MTK_FE_GLO_MISC);
mtk_w32(eth, val | BIT(4), MTK_FE_GLO_MISC);

107
target/linux/generic/backport-6.1/729-19-v6.3-net-ethernet-mtk_eth_soc-introduce-mtk_hw_warm_reset.patch Normal file
View File

@ -0,0 +1,107 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 14 Jan 2023 18:01:29 +0100
Subject: [PATCH] net: ethernet: mtk_eth_soc: introduce mtk_hw_warm_reset
support
Introduce mtk_hw_warm_reset utility routine. This is a preliminary patch
to align reset procedure to vendor sdk and avoid to power down the chip
during hw reset.
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -3275,7 +3275,54 @@ static void mtk_hw_reset(struct mtk_eth
0x3ffffff);
}
-static int mtk_hw_init(struct mtk_eth *eth)
+static u32 mtk_hw_reset_read(struct mtk_eth *eth)
+{
+ u32 val;
+
+ regmap_read(eth->ethsys, ETHSYS_RSTCTRL, &val);
+ return val;
+}
+
+static void mtk_hw_warm_reset(struct mtk_eth *eth)
+{
+ u32 rst_mask, val;
+
+ regmap_update_bits(eth->ethsys, ETHSYS_RSTCTRL, RSTCTRL_FE,
+ RSTCTRL_FE);
+ if (readx_poll_timeout_atomic(mtk_hw_reset_read, eth, val,
+ val & RSTCTRL_FE, 1, 1000)) {
+ dev_err(eth->dev, "warm reset failed\n");
+ mtk_hw_reset(eth);
+ return;
+ }
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2))
+ rst_mask = RSTCTRL_ETH | RSTCTRL_PPE0_V2;
+ else
+ rst_mask = RSTCTRL_ETH | RSTCTRL_PPE0;
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_RSTCTRL_PPE1))
+ rst_mask |= RSTCTRL_PPE1;
+
+ regmap_update_bits(eth->ethsys, ETHSYS_RSTCTRL, rst_mask, rst_mask);
+
+ udelay(1);
+ val = mtk_hw_reset_read(eth);
+ if (!(val & rst_mask))
+ dev_err(eth->dev, "warm reset stage0 failed %08x (%08x)\n",
+ val, rst_mask);
+
+ rst_mask |= RSTCTRL_FE;
+ regmap_update_bits(eth->ethsys, ETHSYS_RSTCTRL, rst_mask, ~rst_mask);
+
+ udelay(1);
+ val = mtk_hw_reset_read(eth);
+ if (val & rst_mask)
+ dev_err(eth->dev, "warm reset stage1 failed %08x (%08x)\n",
+ val, rst_mask);
+}
+
+static int mtk_hw_init(struct mtk_eth *eth, bool reset)
{
u32 dma_mask = ETHSYS_DMA_AG_MAP_PDMA | ETHSYS_DMA_AG_MAP_QDMA |
ETHSYS_DMA_AG_MAP_PPE;
@@ -3314,7 +3361,12 @@ static int mtk_hw_init(struct mtk_eth *e
return 0;
}
- mtk_hw_reset(eth);
+ msleep(100);
+
+ if (reset)
+ mtk_hw_warm_reset(eth);
+ else
+ mtk_hw_reset(eth);
if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
/* Set FE to PDMAv2 if necessary */
@@ -3522,7 +3574,7 @@ static void mtk_pending_work(struct work
if (eth->dev->pins)
pinctrl_select_state(eth->dev->pins->p,
eth->dev->pins->default_state);
- mtk_hw_init(eth);
+ mtk_hw_init(eth, true);
/* restart DMA and enable IRQs */
for (i = 0; i < MTK_MAC_COUNT; i++) {
@@ -4114,7 +4166,7 @@ static int mtk_probe(struct platform_dev
eth->msg_enable = netif_msg_init(mtk_msg_level, MTK_DEFAULT_MSG_ENABLE);
INIT_WORK(&eth->pending_work, mtk_pending_work);
- err = mtk_hw_init(eth);
+ err = mtk_hw_init(eth, false);
if (err)
goto err_wed_exit;

262
target/linux/generic/backport-6.1/729-20-v6.3-net-ethernet-mtk_eth_soc-align-reset-procedure-to-ve.patch Normal file
View File

@ -0,0 +1,262 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 14 Jan 2023 18:01:30 +0100
Subject: [PATCH] net: ethernet: mtk_eth_soc: align reset procedure to vendor
sdk
Avoid to power-down the ethernet chip during hw reset and align reset
procedure to vendor sdk.
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -2842,14 +2842,29 @@ static void mtk_dma_free(struct mtk_eth
kfree(eth->scratch_head);
}
+static bool mtk_hw_reset_check(struct mtk_eth *eth)
+{
+ u32 val = mtk_r32(eth, MTK_INT_STATUS2);
+
+ return (val & MTK_FE_INT_FQ_EMPTY) || (val & MTK_FE_INT_RFIFO_UF) ||
+ (val & MTK_FE_INT_RFIFO_OV) || (val & MTK_FE_INT_TSO_FAIL) ||
+ (val & MTK_FE_INT_TSO_ALIGN) || (val & MTK_FE_INT_TSO_ILLEGAL);
+}
+
static void mtk_tx_timeout(struct net_device *dev, unsigned int txqueue)
{
struct mtk_mac *mac = netdev_priv(dev);
struct mtk_eth *eth = mac->hw;
+ if (test_bit(MTK_RESETTING, &eth->state))
+ return;
+
+ if (!mtk_hw_reset_check(eth))
+ return;
+
eth->netdev[mac->id]->stats.tx_errors++;
- netif_err(eth, tx_err, dev,
- "transmit timed out\n");
+ netif_err(eth, tx_err, dev, "transmit timed out\n");
+
schedule_work(&eth->pending_work);
}
@@ -3329,15 +3344,17 @@ static int mtk_hw_init(struct mtk_eth *e
const struct mtk_reg_map *reg_map = eth->soc->reg_map;
int i, val, ret;
- if (test_and_set_bit(MTK_HW_INIT, &eth->state))
+ if (!reset && test_and_set_bit(MTK_HW_INIT, &eth->state))
return 0;
- pm_runtime_enable(eth->dev);
- pm_runtime_get_sync(eth->dev);
+ if (!reset) {
+ pm_runtime_enable(eth->dev);
+ pm_runtime_get_sync(eth->dev);
- ret = mtk_clk_enable(eth);
- if (ret)
- goto err_disable_pm;
+ ret = mtk_clk_enable(eth);
+ if (ret)
+ goto err_disable_pm;
+ }
if (eth->ethsys)
regmap_update_bits(eth->ethsys, ETHSYS_DMA_AG_MAP, dma_mask,
@@ -3466,8 +3483,10 @@ static int mtk_hw_init(struct mtk_eth *e
return 0;
err_disable_pm:
- pm_runtime_put_sync(eth->dev);
- pm_runtime_disable(eth->dev);
+ if (!reset) {
+ pm_runtime_put_sync(eth->dev);
+ pm_runtime_disable(eth->dev);
+ }
return ret;
}
@@ -3546,30 +3565,53 @@ static int mtk_do_ioctl(struct net_devic
return -EOPNOTSUPP;
}
+static void mtk_prepare_for_reset(struct mtk_eth *eth)
+{
+ u32 val;
+ int i;
+
+ /* disabe FE P3 and P4 */
+ val = mtk_r32(eth, MTK_FE_GLO_CFG) | MTK_FE_LINK_DOWN_P3;
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_RSTCTRL_PPE1))
+ val |= MTK_FE_LINK_DOWN_P4;
+ mtk_w32(eth, val, MTK_FE_GLO_CFG);
+
+ /* adjust PPE configurations to prepare for reset */
+ for (i = 0; i < ARRAY_SIZE(eth->ppe); i++)
+ mtk_ppe_prepare_reset(eth->ppe[i]);
+
+ /* disable NETSYS interrupts */
+ mtk_w32(eth, 0, MTK_FE_INT_ENABLE);
+
+ /* force link down GMAC */
+ for (i = 0; i < 2; i++) {
+ val = mtk_r32(eth, MTK_MAC_MCR(i)) & ~MAC_MCR_FORCE_LINK;
+ mtk_w32(eth, val, MTK_MAC_MCR(i));
+ }
+}
+
static void mtk_pending_work(struct work_struct *work)
{
struct mtk_eth *eth = container_of(work, struct mtk_eth, pending_work);
- int err, i;
unsigned long restart = 0;
+ u32 val;
+ int i;
rtnl_lock();
-
- dev_dbg(eth->dev, "[%s][%d] reset\n", __func__, __LINE__);
set_bit(MTK_RESETTING, &eth->state);
+ mtk_prepare_for_reset(eth);
+
/* stop all devices to make sure that dma is properly shut down */
for (i = 0; i < MTK_MAC_COUNT; i++) {
- if (!eth->netdev[i])
+ if (!eth->netdev[i] || !netif_running(eth->netdev[i]))
continue;
+
mtk_stop(eth->netdev[i]);
__set_bit(i, &restart);
}
- dev_dbg(eth->dev, "[%s][%d] mtk_stop ends\n", __func__, __LINE__);
- /* restart underlying hardware such as power, clock, pin mux
- * and the connected phy
- */
- mtk_hw_deinit(eth);
+ usleep_range(15000, 16000);
if (eth->dev->pins)
pinctrl_select_state(eth->dev->pins->p,
@@ -3580,15 +3622,19 @@ static void mtk_pending_work(struct work
for (i = 0; i < MTK_MAC_COUNT; i++) {
if (!test_bit(i, &restart))
continue;
- err = mtk_open(eth->netdev[i]);
- if (err) {
+
+ if (mtk_open(eth->netdev[i])) {
netif_alert(eth, ifup, eth->netdev[i],
- "Driver up/down cycle failed, closing device.\n");
+ "Driver up/down cycle failed\n");
dev_close(eth->netdev[i]);
}
}
- dev_dbg(eth->dev, "[%s][%d] reset done\n", __func__, __LINE__);
+ /* enabe FE P3 and P4 */
+ val = mtk_r32(eth, MTK_FE_GLO_CFG) & ~MTK_FE_LINK_DOWN_P3;
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_RSTCTRL_PPE1))
+ val &= ~MTK_FE_LINK_DOWN_P4;
+ mtk_w32(eth, val, MTK_FE_GLO_CFG);
clear_bit(MTK_RESETTING, &eth->state);
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -72,12 +72,24 @@
#define MTK_HW_LRO_REPLACE_DELTA 1000
#define MTK_HW_LRO_SDL_REMAIN_ROOM 1522
+/* Frame Engine Global Configuration */
+#define MTK_FE_GLO_CFG 0x00
+#define MTK_FE_LINK_DOWN_P3 BIT(11)
+#define MTK_FE_LINK_DOWN_P4 BIT(12)
+
/* Frame Engine Global Reset Register */
#define MTK_RST_GL 0x04
#define RST_GL_PSE BIT(0)
/* Frame Engine Interrupt Status Register */
#define MTK_INT_STATUS2 0x08
+#define MTK_FE_INT_ENABLE 0x0c
+#define MTK_FE_INT_FQ_EMPTY BIT(8)
+#define MTK_FE_INT_TSO_FAIL BIT(12)
+#define MTK_FE_INT_TSO_ILLEGAL BIT(13)
+#define MTK_FE_INT_TSO_ALIGN BIT(14)
+#define MTK_FE_INT_RFIFO_OV BIT(18)
+#define MTK_FE_INT_RFIFO_UF BIT(19)
#define MTK_GDM1_AF BIT(28)
#define MTK_GDM2_AF BIT(29)
--- a/drivers/net/ethernet/mediatek/mtk_ppe.c
+++ b/drivers/net/ethernet/mediatek/mtk_ppe.c
@@ -710,6 +710,33 @@ int mtk_foe_entry_idle_time(struct mtk_p
return __mtk_foe_entry_idle_time(ppe, entry->data.ib1);
}
+int mtk_ppe_prepare_reset(struct mtk_ppe *ppe)
+{
+ if (!ppe)
+ return -EINVAL;
+
+ /* disable KA */
+ ppe_clear(ppe, MTK_PPE_TB_CFG, MTK_PPE_TB_CFG_KEEPALIVE);
+ ppe_clear(ppe, MTK_PPE_BIND_LMT1, MTK_PPE_NTU_KEEPALIVE);
+ ppe_w32(ppe, MTK_PPE_KEEPALIVE, 0);
+ usleep_range(10000, 11000);
+
+ /* set KA timer to maximum */
+ ppe_set(ppe, MTK_PPE_BIND_LMT1, MTK_PPE_NTU_KEEPALIVE);
+ ppe_w32(ppe, MTK_PPE_KEEPALIVE, 0xffffffff);
+
+ /* set KA tick select */
+ ppe_set(ppe, MTK_PPE_TB_CFG, MTK_PPE_TB_TICK_SEL);
+ ppe_set(ppe, MTK_PPE_TB_CFG, MTK_PPE_TB_CFG_KEEPALIVE);
+ usleep_range(10000, 11000);
+
+ /* disable scan mode */
+ ppe_clear(ppe, MTK_PPE_TB_CFG, MTK_PPE_TB_CFG_SCAN_MODE);
+ usleep_range(10000, 11000);
+
+ return mtk_ppe_wait_busy(ppe);
+}
+
struct mtk_ppe *mtk_ppe_init(struct mtk_eth *eth, void __iomem *base,
int version, int index)
{
--- a/drivers/net/ethernet/mediatek/mtk_ppe.h
+++ b/drivers/net/ethernet/mediatek/mtk_ppe.h
@@ -306,6 +306,7 @@ struct mtk_ppe *mtk_ppe_init(struct mtk_
void mtk_ppe_deinit(struct mtk_eth *eth);
void mtk_ppe_start(struct mtk_ppe *ppe);
int mtk_ppe_stop(struct mtk_ppe *ppe);
+int mtk_ppe_prepare_reset(struct mtk_ppe *ppe);
void __mtk_ppe_check_skb(struct mtk_ppe *ppe, struct sk_buff *skb, u16 hash);
--- a/drivers/net/ethernet/mediatek/mtk_ppe_regs.h
+++ b/drivers/net/ethernet/mediatek/mtk_ppe_regs.h
@@ -58,6 +58,12 @@
#define MTK_PPE_TB_CFG_SCAN_MODE GENMASK(17, 16)
#define MTK_PPE_TB_CFG_HASH_DEBUG GENMASK(19, 18)
#define MTK_PPE_TB_CFG_INFO_SEL BIT(20)
+#define MTK_PPE_TB_TICK_SEL BIT(24)
+
+#define MTK_PPE_BIND_LMT1 0x230
+#define MTK_PPE_NTU_KEEPALIVE GENMASK(23, 16)
+
+#define MTK_PPE_KEEPALIVE 0x234
enum {
MTK_PPE_SCAN_MODE_DISABLED,

249
target/linux/generic/backport-6.1/729-21-v6.3-net-ethernet-mtk_eth_soc-add-dma-checks-to-mtk_hw_re.patch Normal file
View File

@ -0,0 +1,249 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 14 Jan 2023 18:01:31 +0100
Subject: [PATCH] net: ethernet: mtk_eth_soc: add dma checks to
mtk_hw_reset_check
Introduce mtk_hw_check_dma_hang routine to monitor possible dma hangs.
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -50,6 +50,7 @@ static const struct mtk_reg_map mtk_reg_
.delay_irq = 0x0a0c,
.irq_status = 0x0a20,
.irq_mask = 0x0a28,
+ .adma_rx_dbg0 = 0x0a38,
.int_grp = 0x0a50,
},
.qdma = {
@@ -79,6 +80,8 @@ static const struct mtk_reg_map mtk_reg_
[0] = 0x2800,
[1] = 0x2c00,
},
+ .pse_iq_sta = 0x0110,
+ .pse_oq_sta = 0x0118,
};
static const struct mtk_reg_map mt7628_reg_map = {
@@ -109,6 +112,7 @@ static const struct mtk_reg_map mt7986_r
.delay_irq = 0x620c,
.irq_status = 0x6220,
.irq_mask = 0x6228,
+ .adma_rx_dbg0 = 0x6238,
.int_grp = 0x6250,
},
.qdma = {
@@ -138,6 +142,8 @@ static const struct mtk_reg_map mt7986_r
[0] = 0x4800,
[1] = 0x4c00,
},
+ .pse_iq_sta = 0x0180,
+ .pse_oq_sta = 0x01a0,
};
/* strings used by ethtool */
@@ -3337,6 +3343,102 @@ static void mtk_hw_warm_reset(struct mtk
val, rst_mask);
}
+static bool mtk_hw_check_dma_hang(struct mtk_eth *eth)
+{
+ const struct mtk_reg_map *reg_map = eth->soc->reg_map;
+ bool gmac1_tx, gmac2_tx, gdm1_tx, gdm2_tx;
+ bool oq_hang, cdm1_busy, adma_busy;
+ bool wtx_busy, cdm_full, oq_free;
+ u32 wdidx, val, gdm1_fc, gdm2_fc;
+ bool qfsm_hang, qfwd_hang;
+ bool ret = false;
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_SOC_MT7628))
+ return false;
+
+ /* WDMA sanity checks */
+ wdidx = mtk_r32(eth, reg_map->wdma_base[0] + 0xc);
+
+ val = mtk_r32(eth, reg_map->wdma_base[0] + 0x204);
+ wtx_busy = FIELD_GET(MTK_TX_DMA_BUSY, val);
+
+ val = mtk_r32(eth, reg_map->wdma_base[0] + 0x230);
+ cdm_full = !FIELD_GET(MTK_CDM_TXFIFO_RDY, val);
+
+ oq_free = (!(mtk_r32(eth, reg_map->pse_oq_sta) & GENMASK(24, 16)) &&
+ !(mtk_r32(eth, reg_map->pse_oq_sta + 0x4) & GENMASK(8, 0)) &&
+ !(mtk_r32(eth, reg_map->pse_oq_sta + 0x10) & GENMASK(24, 16)));
+
+ if (wdidx == eth->reset.wdidx && wtx_busy && cdm_full && oq_free) {
+ if (++eth->reset.wdma_hang_count > 2) {
+ eth->reset.wdma_hang_count = 0;
+ ret = true;
+ }
+ goto out;
+ }
+
+ /* QDMA sanity checks */
+ qfsm_hang = !!mtk_r32(eth, reg_map->qdma.qtx_cfg + 0x234);
+ qfwd_hang = !mtk_r32(eth, reg_map->qdma.qtx_cfg + 0x308);
+
+ gdm1_tx = FIELD_GET(GENMASK(31, 16), mtk_r32(eth, MTK_FE_GDM1_FSM)) > 0;
+ gdm2_tx = FIELD_GET(GENMASK(31, 16), mtk_r32(eth, MTK_FE_GDM2_FSM)) > 0;
+ gmac1_tx = FIELD_GET(GENMASK(31, 24), mtk_r32(eth, MTK_MAC_FSM(0))) != 1;
+ gmac2_tx = FIELD_GET(GENMASK(31, 24), mtk_r32(eth, MTK_MAC_FSM(1))) != 1;
+ gdm1_fc = mtk_r32(eth, reg_map->gdm1_cnt + 0x24);
+ gdm2_fc = mtk_r32(eth, reg_map->gdm1_cnt + 0x64);
+
+ if (qfsm_hang && qfwd_hang &&
+ ((gdm1_tx && gmac1_tx && gdm1_fc < 1) ||
+ (gdm2_tx && gmac2_tx && gdm2_fc < 1))) {
+ if (++eth->reset.qdma_hang_count > 2) {
+ eth->reset.qdma_hang_count = 0;
+ ret = true;
+ }
+ goto out;
+ }
+
+ /* ADMA sanity checks */
+ oq_hang = !!(mtk_r32(eth, reg_map->pse_oq_sta) & GENMASK(8, 0));
+ cdm1_busy = !!(mtk_r32(eth, MTK_FE_CDM1_FSM) & GENMASK(31, 16));
+ adma_busy = !(mtk_r32(eth, reg_map->pdma.adma_rx_dbg0) & GENMASK(4, 0)) &&
+ !(mtk_r32(eth, reg_map->pdma.adma_rx_dbg0) & BIT(6));
+
+ if (oq_hang && cdm1_busy && adma_busy) {
+ if (++eth->reset.adma_hang_count > 2) {
+ eth->reset.adma_hang_count = 0;
+ ret = true;
+ }
+ goto out;
+ }
+
+ eth->reset.wdma_hang_count = 0;
+ eth->reset.qdma_hang_count = 0;
+ eth->reset.adma_hang_count = 0;
+out:
+ eth->reset.wdidx = wdidx;
+
+ return ret;
+}
+
+static void mtk_hw_reset_monitor_work(struct work_struct *work)
+{
+ struct delayed_work *del_work = to_delayed_work(work);
+ struct mtk_eth *eth = container_of(del_work, struct mtk_eth,
+ reset.monitor_work);
+
+ if (test_bit(MTK_RESETTING, &eth->state))
+ goto out;
+
+ /* DMA stuck checks */
+ if (mtk_hw_check_dma_hang(eth))
+ schedule_work(&eth->pending_work);
+
+out:
+ schedule_delayed_work(&eth->reset.monitor_work,
+ MTK_DMA_MONITOR_TIMEOUT);
+}
+
static int mtk_hw_init(struct mtk_eth *eth, bool reset)
{
u32 dma_mask = ETHSYS_DMA_AG_MAP_PDMA | ETHSYS_DMA_AG_MAP_QDMA |
@@ -3672,6 +3774,7 @@ static int mtk_cleanup(struct mtk_eth *e
mtk_unreg_dev(eth);
mtk_free_dev(eth);
cancel_work_sync(&eth->pending_work);
+ cancel_delayed_work_sync(&eth->reset.monitor_work);
return 0;
}
@@ -4099,6 +4202,7 @@ static int mtk_probe(struct platform_dev
eth->rx_dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE;
INIT_WORK(&eth->rx_dim.work, mtk_dim_rx);
+ INIT_DELAYED_WORK(&eth->reset.monitor_work, mtk_hw_reset_monitor_work);
eth->tx_dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE;
INIT_WORK(&eth->tx_dim.work, mtk_dim_tx);
@@ -4301,6 +4405,8 @@ static int mtk_probe(struct platform_dev
netif_napi_add(&eth->dummy_dev, &eth->rx_napi, mtk_napi_rx);
platform_set_drvdata(pdev, eth);
+ schedule_delayed_work(&eth->reset.monitor_work,
+ MTK_DMA_MONITOR_TIMEOUT);
return 0;
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -257,6 +257,8 @@
#define MTK_RX_DONE_INT_V2 BIT(14)
+#define MTK_CDM_TXFIFO_RDY BIT(7)
+
/* QDMA Interrupt grouping registers */
#define MTK_RLS_DONE_INT BIT(0)
@@ -542,6 +544,17 @@
#define MT7628_SDM_RBCNT (MT7628_SDM_OFFSET + 0x10c)
#define MT7628_SDM_CS_ERR (MT7628_SDM_OFFSET + 0x110)
+#define MTK_FE_CDM1_FSM 0x220
+#define MTK_FE_CDM2_FSM 0x224
+#define MTK_FE_CDM3_FSM 0x238
+#define MTK_FE_CDM4_FSM 0x298
+#define MTK_FE_CDM5_FSM 0x318
+#define MTK_FE_CDM6_FSM 0x328
+#define MTK_FE_GDM1_FSM 0x228
+#define MTK_FE_GDM2_FSM 0x22C
+
+#define MTK_MAC_FSM(x) (0x1010C + ((x) * 0x100))
+
struct mtk_rx_dma {
unsigned int rxd1;
unsigned int rxd2;
@@ -938,6 +951,7 @@ struct mtk_reg_map {
u32 delay_irq; /* delay interrupt */
u32 irq_status; /* interrupt status */
u32 irq_mask; /* interrupt mask */
+ u32 adma_rx_dbg0;
u32 int_grp;
} pdma;
struct {
@@ -964,6 +978,8 @@ struct mtk_reg_map {
u32 gdma_to_ppe;
u32 ppe_base;
u32 wdma_base[2];
+ u32 pse_iq_sta;
+ u32 pse_oq_sta;
};
/* struct mtk_eth_data - This is the structure holding all differences
@@ -1006,6 +1022,8 @@ struct mtk_soc_data {
} txrx;
};
+#define MTK_DMA_MONITOR_TIMEOUT msecs_to_jiffies(1000)
+
/* currently no SoC has more than 2 macs */
#define MTK_MAX_DEVS 2
@@ -1128,6 +1146,14 @@ struct mtk_eth {
struct rhashtable flow_table;
struct bpf_prog __rcu *prog;
+
+ struct {
+ struct delayed_work monitor_work;
+ u32 wdidx;
+ u8 wdma_hang_count;
+ u8 qdma_hang_count;
+ u8 adma_hang_count;
+ } reset;
};
/* struct mtk_mac - the structure that holds the info about the MACs of the

124
target/linux/generic/backport-6.1/729-22-v6.3-net-ethernet-mtk_wed-add-reset-reset_complete-callba.patch Normal file
View File

@ -0,0 +1,124 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Sat, 14 Jan 2023 18:01:32 +0100
Subject: [PATCH] net: ethernet: mtk_wed: add reset/reset_complete callbacks
Introduce reset and reset_complete wlan callback to schedule WLAN driver
reset when ethernet/wed driver is resetting.
Tested-by: Daniel Golle <daniel@makrotopia.org>
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -3703,6 +3703,11 @@ static void mtk_pending_work(struct work
set_bit(MTK_RESETTING, &eth->state);
mtk_prepare_for_reset(eth);
+ mtk_wed_fe_reset();
+ /* Run again reset preliminary configuration in order to avoid any
+ * possible race during FE reset since it can run releasing RTNL lock.
+ */
+ mtk_prepare_for_reset(eth);
/* stop all devices to make sure that dma is properly shut down */
for (i = 0; i < MTK_MAC_COUNT; i++) {
@@ -3740,6 +3745,8 @@ static void mtk_pending_work(struct work
clear_bit(MTK_RESETTING, &eth->state);
+ mtk_wed_fe_reset_complete();
+
rtnl_unlock();
}
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -205,6 +205,48 @@ mtk_wed_wo_reset(struct mtk_wed_device *
iounmap(reg);
}
+void mtk_wed_fe_reset(void)
+{
+ int i;
+
+ mutex_lock(&hw_lock);
+
+ for (i = 0; i < ARRAY_SIZE(hw_list); i++) {
+ struct mtk_wed_hw *hw = hw_list[i];
+ struct mtk_wed_device *dev = hw->wed_dev;
+ int err;
+
+ if (!dev || !dev->wlan.reset)
+ continue;
+
+ /* reset callback blocks until WLAN reset is completed */
+ err = dev->wlan.reset(dev);
+ if (err)
+ dev_err(dev->dev, "wlan reset failed: %d\n", err);
+ }
+
+ mutex_unlock(&hw_lock);
+}
+
+void mtk_wed_fe_reset_complete(void)
+{
+ int i;
+
+ mutex_lock(&hw_lock);
+
+ for (i = 0; i < ARRAY_SIZE(hw_list); i++) {
+ struct mtk_wed_hw *hw = hw_list[i];
+ struct mtk_wed_device *dev = hw->wed_dev;
+
+ if (!dev || !dev->wlan.reset_complete)
+ continue;
+
+ dev->wlan.reset_complete(dev);
+ }
+
+ mutex_unlock(&hw_lock);
+}
+
static struct mtk_wed_hw *
mtk_wed_assign(struct mtk_wed_device *dev)
{
--- a/drivers/net/ethernet/mediatek/mtk_wed.h
+++ b/drivers/net/ethernet/mediatek/mtk_wed.h
@@ -128,6 +128,8 @@ void mtk_wed_add_hw(struct device_node *
void mtk_wed_exit(void);
int mtk_wed_flow_add(int index);
void mtk_wed_flow_remove(int index);
+void mtk_wed_fe_reset(void);
+void mtk_wed_fe_reset_complete(void);
#else
static inline void
mtk_wed_add_hw(struct device_node *np, struct mtk_eth *eth,
@@ -147,6 +149,13 @@ static inline void mtk_wed_flow_remove(i
{
}
+static inline void mtk_wed_fe_reset(void)
+{
+}
+
+static inline void mtk_wed_fe_reset_complete(void)
+{
+}
#endif
#ifdef CONFIG_DEBUG_FS
--- a/include/linux/soc/mediatek/mtk_wed.h
+++ b/include/linux/soc/mediatek/mtk_wed.h
@@ -151,6 +151,8 @@ struct mtk_wed_device {
void (*release_rx_buf)(struct mtk_wed_device *wed);
void (*update_wo_rx_stats)(struct mtk_wed_device *wed,
struct mtk_wed_wo_rx_stats *stats);
+ int (*reset)(struct mtk_wed_device *wed);
+ void (*reset_complete)(struct mtk_wed_device *wed);
} wlan;
#endif
};

106
target/linux/generic/backport-6.1/729-23-v6.3-net-ethernet-mtk_wed-add-reset-to-rx_ring_setup-call.patch Normal file
View File

@ -0,0 +1,106 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Mon, 5 Dec 2022 12:34:42 +0100
Subject: [PATCH] net: ethernet: mtk_wed: add reset to rx_ring_setup callback
This patch adds reset parameter to mtk_wed_rx_ring_setup signature
in order to align rx_ring_setup callback to tx_ring_setup one introduced
in 'commit 23dca7a90017 ("net: ethernet: mtk_wed: add reset to
tx_ring_setup callback")'
Co-developed-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Sujuan Chen <sujuan.chen@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/29c6e7a5469e784406cf3e2920351d1207713d05.1670239984.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -1252,7 +1252,8 @@ mtk_wed_wdma_rx_ring_setup(struct mtk_we
}
static int
-mtk_wed_wdma_tx_ring_setup(struct mtk_wed_device *dev, int idx, int size)
+mtk_wed_wdma_tx_ring_setup(struct mtk_wed_device *dev, int idx, int size,
+ bool reset)
{
u32 desc_size = sizeof(struct mtk_wdma_desc) * dev->hw->version;
struct mtk_wed_ring *wdma;
@@ -1261,8 +1262,8 @@ mtk_wed_wdma_tx_ring_setup(struct mtk_we
return -EINVAL;
wdma = &dev->tx_wdma[idx];
- if (mtk_wed_ring_alloc(dev, wdma, MTK_WED_WDMA_RING_SIZE, desc_size,
- true))
+ if (!reset && mtk_wed_ring_alloc(dev, wdma, MTK_WED_WDMA_RING_SIZE,
+ desc_size, true))
return -ENOMEM;
wdma_w32(dev, MTK_WDMA_RING_TX(idx) + MTK_WED_RING_OFS_BASE,
@@ -1272,6 +1273,9 @@ mtk_wed_wdma_tx_ring_setup(struct mtk_we
wdma_w32(dev, MTK_WDMA_RING_TX(idx) + MTK_WED_RING_OFS_CPU_IDX, 0);
wdma_w32(dev, MTK_WDMA_RING_TX(idx) + MTK_WED_RING_OFS_DMA_IDX, 0);
+ if (reset)
+ mtk_wed_ring_reset(wdma, MTK_WED_WDMA_RING_SIZE, true);
+
if (!idx) {
wed_w32(dev, MTK_WED_WDMA_RING_TX + MTK_WED_RING_OFS_BASE,
wdma->desc_phys);
@@ -1611,18 +1615,20 @@ mtk_wed_txfree_ring_setup(struct mtk_wed
}
static int
-mtk_wed_rx_ring_setup(struct mtk_wed_device *dev, int idx, void __iomem *regs)
+mtk_wed_rx_ring_setup(struct mtk_wed_device *dev, int idx, void __iomem *regs,
+ bool reset)
{
struct mtk_wed_ring *ring = &dev->rx_ring[idx];
if (WARN_ON(idx >= ARRAY_SIZE(dev->rx_ring)))
return -EINVAL;
- if (mtk_wed_ring_alloc(dev, ring, MTK_WED_RX_RING_SIZE,
- sizeof(*ring->desc), false))
+ if (!reset && mtk_wed_ring_alloc(dev, ring, MTK_WED_RX_RING_SIZE,
+ sizeof(*ring->desc), false))
return -ENOMEM;
- if (mtk_wed_wdma_tx_ring_setup(dev, idx, MTK_WED_WDMA_RING_SIZE))
+ if (mtk_wed_wdma_tx_ring_setup(dev, idx, MTK_WED_WDMA_RING_SIZE,
+ reset))
return -ENOMEM;
ring->reg_base = MTK_WED_RING_RX_DATA(idx);
--- a/include/linux/soc/mediatek/mtk_wed.h
+++ b/include/linux/soc/mediatek/mtk_wed.h
@@ -162,7 +162,7 @@ struct mtk_wed_ops {
int (*tx_ring_setup)(struct mtk_wed_device *dev, int ring,
void __iomem *regs, bool reset);
int (*rx_ring_setup)(struct mtk_wed_device *dev, int ring,
- void __iomem *regs);
+ void __iomem *regs, bool reset);
int (*txfree_ring_setup)(struct mtk_wed_device *dev,
void __iomem *regs);
int (*msg_update)(struct mtk_wed_device *dev, int cmd_id,
@@ -230,8 +230,8 @@ mtk_wed_get_rx_capa(struct mtk_wed_devic
(_dev)->ops->irq_get(_dev, _mask)
#define mtk_wed_device_irq_set_mask(_dev, _mask) \
(_dev)->ops->irq_set_mask(_dev, _mask)
-#define mtk_wed_device_rx_ring_setup(_dev, _ring, _regs) \
- (_dev)->ops->rx_ring_setup(_dev, _ring, _regs)
+#define mtk_wed_device_rx_ring_setup(_dev, _ring, _regs, _reset) \
+ (_dev)->ops->rx_ring_setup(_dev, _ring, _regs, _reset)
#define mtk_wed_device_ppe_check(_dev, _skb, _reason, _hash) \
(_dev)->ops->ppe_check(_dev, _skb, _reason, _hash)
#define mtk_wed_device_update_msg(_dev, _id, _msg, _len) \
@@ -251,7 +251,7 @@ static inline bool mtk_wed_device_active
#define mtk_wed_device_reg_write(_dev, _reg, _val) do {} while (0)
#define mtk_wed_device_irq_get(_dev, _mask) 0
#define mtk_wed_device_irq_set_mask(_dev, _mask) do {} while (0)
-#define mtk_wed_device_rx_ring_setup(_dev, _ring, _regs) -ENODEV
+#define mtk_wed_device_rx_ring_setup(_dev, _ring, _regs, _reset) -ENODEV
#define mtk_wed_device_ppe_check(_dev, _skb, _reason, _hash) do {} while (0)
#define mtk_wed_device_update_msg(_dev, _id, _msg, _len) -ENODEV
#define mtk_wed_device_stop(_dev) do {} while (0)

22
target/linux/generic/backport-6.1/730-01-v6.3-net-ethernet-mtk_eth_soc-account-for-vlan-in-rx-head.patch Normal file
View File

@ -0,0 +1,22 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Thu, 27 Oct 2022 19:50:31 +0200
Subject: [PATCH] net: ethernet: mtk_eth_soc: account for vlan in rx
header length
The network stack assumes that devices can handle an extra VLAN tag without
increasing the MTU
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -29,7 +29,7 @@
#define MTK_TX_DMA_BUF_LEN_V2 0xffff
#define MTK_DMA_SIZE 512
#define MTK_MAC_COUNT 2
-#define MTK_RX_ETH_HLEN (ETH_HLEN + ETH_FCS_LEN)
+#define MTK_RX_ETH_HLEN (VLAN_ETH_HLEN + ETH_FCS_LEN)
#define MTK_RX_HLEN (NET_SKB_PAD + MTK_RX_ETH_HLEN + NET_IP_ALIGN)
#define MTK_DMA_DUMMY_DESC 0xffffffff
#define MTK_DEFAULT_MSG_ENABLE (NETIF_MSG_DRV | \

143
target/linux/generic/backport-6.1/730-02-v6.3-net-ethernet-mtk_eth_soc-increase-tx-ring-side-for-Q.patch Normal file
View File

@ -0,0 +1,143 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Thu, 27 Oct 2022 19:53:57 +0200
Subject: [PATCH] net: ethernet: mtk_eth_soc: increase tx ring side for
QDMA devices
In order to use the hardware traffic shaper feature, a larger tx ring is
needed, especially for the scratch ring, which the hardware shaper uses to
reorder packets.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -945,7 +945,7 @@ static int mtk_init_fq_dma(struct mtk_et
{
const struct mtk_soc_data *soc = eth->soc;
dma_addr_t phy_ring_tail;
- int cnt = MTK_DMA_SIZE;
+ int cnt = MTK_QDMA_RING_SIZE;
dma_addr_t dma_addr;
int i;
@@ -2209,19 +2209,25 @@ static int mtk_tx_alloc(struct mtk_eth *
struct mtk_tx_ring *ring = &eth->tx_ring;
int i, sz = soc->txrx.txd_size;
struct mtk_tx_dma_v2 *txd;
+ int ring_size;
- ring->buf = kcalloc(MTK_DMA_SIZE, sizeof(*ring->buf),
+ if (MTK_HAS_CAPS(soc->caps, MTK_QDMA))
+ ring_size = MTK_QDMA_RING_SIZE;
+ else
+ ring_size = MTK_DMA_SIZE;
+
+ ring->buf = kcalloc(ring_size, sizeof(*ring->buf),
GFP_KERNEL);
if (!ring->buf)
goto no_tx_mem;
- ring->dma = dma_alloc_coherent(eth->dma_dev, MTK_DMA_SIZE * sz,
+ ring->dma = dma_alloc_coherent(eth->dma_dev, ring_size * sz,
&ring->phys, GFP_KERNEL);
if (!ring->dma)
goto no_tx_mem;
- for (i = 0; i < MTK_DMA_SIZE; i++) {
- int next = (i + 1) % MTK_DMA_SIZE;
+ for (i = 0; i < ring_size; i++) {
+ int next = (i + 1) % ring_size;
u32 next_ptr = ring->phys + next * sz;
txd = ring->dma + i * sz;
@@ -2241,22 +2247,22 @@ static int mtk_tx_alloc(struct mtk_eth *
* descriptors in ring->dma_pdma.
*/
if (!MTK_HAS_CAPS(soc->caps, MTK_QDMA)) {
- ring->dma_pdma = dma_alloc_coherent(eth->dma_dev, MTK_DMA_SIZE * sz,
+ ring->dma_pdma = dma_alloc_coherent(eth->dma_dev, ring_size * sz,
&ring->phys_pdma, GFP_KERNEL);
if (!ring->dma_pdma)
goto no_tx_mem;
- for (i = 0; i < MTK_DMA_SIZE; i++) {
+ for (i = 0; i < ring_size; i++) {
ring->dma_pdma[i].txd2 = TX_DMA_DESP2_DEF;
ring->dma_pdma[i].txd4 = 0;
}
}
- ring->dma_size = MTK_DMA_SIZE;
- atomic_set(&ring->free_count, MTK_DMA_SIZE - 2);
+ ring->dma_size = ring_size;
+ atomic_set(&ring->free_count, ring_size - 2);
ring->next_free = ring->dma;
ring->last_free = (void *)txd;
- ring->last_free_ptr = (u32)(ring->phys + ((MTK_DMA_SIZE - 1) * sz));
+ ring->last_free_ptr = (u32)(ring->phys + ((ring_size - 1) * sz));
ring->thresh = MAX_SKB_FRAGS;
/* make sure that all changes to the dma ring are flushed before we
@@ -2268,14 +2274,14 @@ static int mtk_tx_alloc(struct mtk_eth *
mtk_w32(eth, ring->phys, soc->reg_map->qdma.ctx_ptr);
mtk_w32(eth, ring->phys, soc->reg_map->qdma.dtx_ptr);
mtk_w32(eth,
- ring->phys + ((MTK_DMA_SIZE - 1) * sz),
+ ring->phys + ((ring_size - 1) * sz),
soc->reg_map->qdma.crx_ptr);
mtk_w32(eth, ring->last_free_ptr, soc->reg_map->qdma.drx_ptr);
mtk_w32(eth, (QDMA_RES_THRES << 8) | QDMA_RES_THRES,
soc->reg_map->qdma.qtx_cfg);
} else {
mtk_w32(eth, ring->phys_pdma, MT7628_TX_BASE_PTR0);
- mtk_w32(eth, MTK_DMA_SIZE, MT7628_TX_MAX_CNT0);
+ mtk_w32(eth, ring_size, MT7628_TX_MAX_CNT0);
mtk_w32(eth, 0, MT7628_TX_CTX_IDX0);
mtk_w32(eth, MT7628_PST_DTX_IDX0, soc->reg_map->pdma.rst_idx);
}
@@ -2293,7 +2299,7 @@ static void mtk_tx_clean(struct mtk_eth
int i;
if (ring->buf) {
- for (i = 0; i < MTK_DMA_SIZE; i++)
+ for (i = 0; i < ring->dma_size; i++)
mtk_tx_unmap(eth, &ring->buf[i], NULL, false);
kfree(ring->buf);
ring->buf = NULL;
@@ -2301,14 +2307,14 @@ static void mtk_tx_clean(struct mtk_eth
if (ring->dma) {
dma_free_coherent(eth->dma_dev,
- MTK_DMA_SIZE * soc->txrx.txd_size,
+ ring->dma_size * soc->txrx.txd_size,
ring->dma, ring->phys);
ring->dma = NULL;
}
if (ring->dma_pdma) {
dma_free_coherent(eth->dma_dev,
- MTK_DMA_SIZE * soc->txrx.txd_size,
+ ring->dma_size * soc->txrx.txd_size,
ring->dma_pdma, ring->phys_pdma);
ring->dma_pdma = NULL;
}
@@ -2830,7 +2836,7 @@ static void mtk_dma_free(struct mtk_eth
netdev_reset_queue(eth->netdev[i]);
if (eth->scratch_ring) {
dma_free_coherent(eth->dma_dev,
- MTK_DMA_SIZE * soc->txrx.txd_size,
+ MTK_QDMA_RING_SIZE * soc->txrx.txd_size,
eth->scratch_ring, eth->phy_scratch_ring);
eth->scratch_ring = NULL;
eth->phy_scratch_ring = 0;
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -27,6 +27,7 @@
#define MTK_MAX_RX_LENGTH_2K 2048
#define MTK_TX_DMA_BUF_LEN 0x3fff
#define MTK_TX_DMA_BUF_LEN_V2 0xffff
+#define MTK_QDMA_RING_SIZE 2048
#define MTK_DMA_SIZE 512
#define MTK_MAC_COUNT 2
#define MTK_RX_ETH_HLEN (VLAN_ETH_HLEN + ETH_FCS_LEN)

52
target/linux/generic/backport-6.1/730-03-v6.3-net-ethernet-mtk_eth_soc-avoid-port_mg-assignment-on.patch Normal file
View File

@ -0,0 +1,52 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Fri, 4 Nov 2022 19:49:08 +0100
Subject: [PATCH] net: ethernet: mtk_eth_soc: avoid port_mg assignment on
MT7622 and newer
On newer chips, this field is unused and contains some bits related to queue
assignment. Initialize it to 0 in those cases.
Fix offload_version on MT7621 and MT7623, which still need the previous value.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -4484,7 +4484,7 @@ static const struct mtk_soc_data mt7621_
.hw_features = MTK_HW_FEATURES,
.required_clks = MT7621_CLKS_BITMAP,
.required_pctl = false,
- .offload_version = 2,
+ .offload_version = 1,
.hash_offset = 2,
.foe_entry_size = sizeof(struct mtk_foe_entry) - 16,
.txrx = {
@@ -4523,7 +4523,7 @@ static const struct mtk_soc_data mt7623_
.hw_features = MTK_HW_FEATURES,
.required_clks = MT7623_CLKS_BITMAP,
.required_pctl = true,
- .offload_version = 2,
+ .offload_version = 1,
.hash_offset = 2,
.foe_entry_size = sizeof(struct mtk_foe_entry) - 16,
.txrx = {
--- a/drivers/net/ethernet/mediatek/mtk_ppe.c
+++ b/drivers/net/ethernet/mediatek/mtk_ppe.c
@@ -175,6 +175,8 @@ int mtk_foe_entry_prepare(struct mtk_eth
val = FIELD_PREP(MTK_FOE_IB2_DEST_PORT_V2, pse_port) |
FIELD_PREP(MTK_FOE_IB2_PORT_AG_V2, 0xf);
} else {
+ int port_mg = eth->soc->offload_version > 1 ? 0 : 0x3f;
+
val = FIELD_PREP(MTK_FOE_IB1_STATE, MTK_FOE_STATE_BIND) |
FIELD_PREP(MTK_FOE_IB1_PACKET_TYPE, type) |
FIELD_PREP(MTK_FOE_IB1_UDP, l4proto == IPPROTO_UDP) |
@@ -182,7 +184,7 @@ int mtk_foe_entry_prepare(struct mtk_eth
entry->ib1 = val;
val = FIELD_PREP(MTK_FOE_IB2_DEST_PORT, pse_port) |
- FIELD_PREP(MTK_FOE_IB2_PORT_MG, 0x3f) |
+ FIELD_PREP(MTK_FOE_IB2_PORT_MG, port_mg) |
FIELD_PREP(MTK_FOE_IB2_PORT_AG, 0x1f);
}

654
target/linux/generic/backport-6.1/730-04-v6.3-net-ethernet-mtk_eth_soc-implement-multi-queue-suppo.patch Normal file
View File

@ -0,0 +1,654 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Thu, 27 Oct 2022 20:17:27 +0200
Subject: [PATCH] net: ethernet: mtk_eth_soc: implement multi-queue
support for per-port queues
When sending traffic to multiple ports with different link speeds, queued
packets to one port can drown out tx to other ports.
In order to better handle transmission to multiple ports, use the hardware
shaper feature to implement weighted fair queueing between ports.
Weight and maximum rate are automatically adjusted based on the link speed
of the port.
The first 3 queues are unrestricted and reserved for non-DSA direct tx on
GMAC ports. The following queues are automatically assigned by the MTK DSA
tag driver based on the target port number.
The PPE offload code configures the queues for offloaded traffic in the same
way.
This feature is only supported on devices supporting QDMA. All queues still
share the same DMA ring and descriptor pool.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -55,6 +55,7 @@ static const struct mtk_reg_map mtk_reg_
},
.qdma = {
.qtx_cfg = 0x1800,
+ .qtx_sch = 0x1804,
.rx_ptr = 0x1900,
.rx_cnt_cfg = 0x1904,
.qcrx_ptr = 0x1908,
@@ -62,6 +63,7 @@ static const struct mtk_reg_map mtk_reg_
.rst_idx = 0x1a08,
.delay_irq = 0x1a0c,
.fc_th = 0x1a10,
+ .tx_sch_rate = 0x1a14,
.int_grp = 0x1a20,
.hred = 0x1a44,
.ctx_ptr = 0x1b00,
@@ -117,6 +119,7 @@ static const struct mtk_reg_map mt7986_r
},
.qdma = {
.qtx_cfg = 0x4400,
+ .qtx_sch = 0x4404,
.rx_ptr = 0x4500,
.rx_cnt_cfg = 0x4504,
.qcrx_ptr = 0x4508,
@@ -134,6 +137,7 @@ static const struct mtk_reg_map mt7986_r
.fq_tail = 0x4724,
.fq_count = 0x4728,
.fq_blen = 0x472c,
+ .tx_sch_rate = 0x4798,
},
.gdm1_cnt = 0x1c00,
.gdma_to_ppe = 0x3333,
@@ -620,6 +624,75 @@ static void mtk_mac_link_down(struct phy
mtk_w32(mac->hw, mcr, MTK_MAC_MCR(mac->id));
}
+static void mtk_set_queue_speed(struct mtk_eth *eth, unsigned int idx,
+ int speed)
+{
+ const struct mtk_soc_data *soc = eth->soc;
+ u32 ofs, val;
+
+ if (!MTK_HAS_CAPS(soc->caps, MTK_QDMA))
+ return;
+
+ val = MTK_QTX_SCH_MIN_RATE_EN |
+ /* minimum: 10 Mbps */
+ FIELD_PREP(MTK_QTX_SCH_MIN_RATE_MAN, 1) |
+ FIELD_PREP(MTK_QTX_SCH_MIN_RATE_EXP, 4) |
+ MTK_QTX_SCH_LEAKY_BUCKET_SIZE;
+ if (!MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2))
+ val |= MTK_QTX_SCH_LEAKY_BUCKET_EN;
+
+ if (IS_ENABLED(CONFIG_SOC_MT7621)) {
+ switch (speed) {
+ case SPEED_10:
+ val |= MTK_QTX_SCH_MAX_RATE_EN |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_MAN, 103) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_EXP, 2) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_WEIGHT, 1);
+ break;
+ case SPEED_100:
+ val |= MTK_QTX_SCH_MAX_RATE_EN |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_MAN, 103) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_EXP, 3);
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_WEIGHT, 1);
+ break;
+ case SPEED_1000:
+ val |= MTK_QTX_SCH_MAX_RATE_EN |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_MAN, 105) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_EXP, 4) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_WEIGHT, 10);
+ break;
+ default:
+ break;
+ }
+ } else {
+ switch (speed) {
+ case SPEED_10:
+ val |= MTK_QTX_SCH_MAX_RATE_EN |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_MAN, 1) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_EXP, 4) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_WEIGHT, 1);
+ break;
+ case SPEED_100:
+ val |= MTK_QTX_SCH_MAX_RATE_EN |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_MAN, 1) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_EXP, 5);
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_WEIGHT, 1);
+ break;
+ case SPEED_1000:
+ val |= MTK_QTX_SCH_MAX_RATE_EN |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_MAN, 10) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_EXP, 5) |
+ FIELD_PREP(MTK_QTX_SCH_MAX_RATE_WEIGHT, 10);
+ break;
+ default:
+ break;
+ }
+ }
+
+ ofs = MTK_QTX_OFFSET * idx;
+ mtk_w32(eth, val, soc->reg_map->qdma.qtx_sch + ofs);
+}
+
static void mtk_mac_link_up(struct phylink_config *config,
struct phy_device *phy,
unsigned int mode, phy_interface_t interface,
@@ -645,6 +718,8 @@ static void mtk_mac_link_up(struct phyli
break;
}
+ mtk_set_queue_speed(mac->hw, mac->id, speed);
+
/* Configure duplex */
if (duplex == DUPLEX_FULL)
mcr |= MAC_MCR_FORCE_DPX;
@@ -1106,7 +1181,8 @@ static void mtk_tx_set_dma_desc_v1(struc
WRITE_ONCE(desc->txd1, info->addr);
- data = TX_DMA_SWC | TX_DMA_PLEN0(info->size);
+ data = TX_DMA_SWC | TX_DMA_PLEN0(info->size) |
+ FIELD_PREP(TX_DMA_PQID, info->qid);
if (info->last)
data |= TX_DMA_LS0;
WRITE_ONCE(desc->txd3, data);
@@ -1140,9 +1216,6 @@ static void mtk_tx_set_dma_desc_v2(struc
data |= TX_DMA_LS0;
WRITE_ONCE(desc->txd3, data);
- if (!info->qid && mac->id)
- info->qid = MTK_QDMA_GMAC2_QID;
-
data = (mac->id + 1) << TX_DMA_FPORT_SHIFT_V2; /* forward port */
data |= TX_DMA_SWC_V2 | QID_BITS_V2(info->qid);
WRITE_ONCE(desc->txd4, data);
@@ -1186,11 +1259,12 @@ static int mtk_tx_map(struct sk_buff *sk
.gso = gso,
.csum = skb->ip_summed == CHECKSUM_PARTIAL,
.vlan = skb_vlan_tag_present(skb),
- .qid = skb->mark & MTK_QDMA_TX_MASK,
+ .qid = skb_get_queue_mapping(skb),
.vlan_tci = skb_vlan_tag_get(skb),
.first = true,
.last = !skb_is_nonlinear(skb),
};
+ struct netdev_queue *txq;
struct mtk_mac *mac = netdev_priv(dev);
struct mtk_eth *eth = mac->hw;
const struct mtk_soc_data *soc = eth->soc;
@@ -1198,8 +1272,10 @@ static int mtk_tx_map(struct sk_buff *sk
struct mtk_tx_dma *itxd_pdma, *txd_pdma;
struct mtk_tx_buf *itx_buf, *tx_buf;
int i, n_desc = 1;
+ int queue = skb_get_queue_mapping(skb);
int k = 0;
+ txq = netdev_get_tx_queue(dev, queue);
itxd = ring->next_free;
itxd_pdma = qdma_to_pdma(ring, itxd);
if (itxd == ring->last_free)
@@ -1248,7 +1324,7 @@ static int mtk_tx_map(struct sk_buff *sk
memset(&txd_info, 0, sizeof(struct mtk_tx_dma_desc_info));
txd_info.size = min_t(unsigned int, frag_size,
soc->txrx.dma_max_len);
- txd_info.qid = skb->mark & MTK_QDMA_TX_MASK;
+ txd_info.qid = queue;
txd_info.last = i == skb_shinfo(skb)->nr_frags - 1 &&
!(frag_size - txd_info.size);
txd_info.addr = skb_frag_dma_map(eth->dma_dev, frag,
@@ -1287,7 +1363,7 @@ static int mtk_tx_map(struct sk_buff *sk
txd_pdma->txd2 |= TX_DMA_LS1;
}
- netdev_sent_queue(dev, skb->len);
+ netdev_tx_sent_queue(txq, skb->len);
skb_tx_timestamp(skb);
ring->next_free = mtk_qdma_phys_to_virt(ring, txd->txd2);
@@ -1299,8 +1375,7 @@ static int mtk_tx_map(struct sk_buff *sk
wmb();
if (MTK_HAS_CAPS(soc->caps, MTK_QDMA)) {
- if (netif_xmit_stopped(netdev_get_tx_queue(dev, 0)) ||
- !netdev_xmit_more())
+ if (netif_xmit_stopped(txq) || !netdev_xmit_more())
mtk_w32(eth, txd->txd2, soc->reg_map->qdma.ctx_ptr);
} else {
int next_idx;
@@ -1369,7 +1444,7 @@ static void mtk_wake_queue(struct mtk_et
for (i = 0; i < MTK_MAC_COUNT; i++) {
if (!eth->netdev[i])
continue;
- netif_wake_queue(eth->netdev[i]);
+ netif_tx_wake_all_queues(eth->netdev[i]);
}
}
@@ -1393,7 +1468,7 @@ static netdev_tx_t mtk_start_xmit(struct
tx_num = mtk_cal_txd_req(eth, skb);
if (unlikely(atomic_read(&ring->free_count) <= tx_num)) {
- netif_stop_queue(dev);
+ netif_tx_stop_all_queues(dev);
netif_err(eth, tx_queued, dev,
"Tx Ring full when queue awake!\n");
spin_unlock(&eth->page_lock);
@@ -1419,7 +1494,7 @@ static netdev_tx_t mtk_start_xmit(struct
goto drop;
if (unlikely(atomic_read(&ring->free_count) <= ring->thresh))
- netif_stop_queue(dev);
+ netif_tx_stop_all_queues(dev);
spin_unlock(&eth->page_lock);
@@ -1586,10 +1661,12 @@ static int mtk_xdp_submit_frame(struct m
struct skb_shared_info *sinfo = xdp_get_shared_info_from_frame(xdpf);
const struct mtk_soc_data *soc = eth->soc;
struct mtk_tx_ring *ring = &eth->tx_ring;
+ struct mtk_mac *mac = netdev_priv(dev);
struct mtk_tx_dma_desc_info txd_info = {
.size = xdpf->len,
.first = true,
.last = !xdp_frame_has_frags(xdpf),
+ .qid = mac->id,
};
int err, index = 0, n_desc = 1, nr_frags;
struct mtk_tx_buf *htx_buf, *tx_buf;
@@ -1639,6 +1716,7 @@ static int mtk_xdp_submit_frame(struct m
memset(&txd_info, 0, sizeof(struct mtk_tx_dma_desc_info));
txd_info.size = skb_frag_size(&sinfo->frags[index]);
txd_info.last = index + 1 == nr_frags;
+ txd_info.qid = mac->id;
data = skb_frag_address(&sinfo->frags[index]);
index++;
@@ -1993,8 +2071,46 @@ rx_done:
return done;
}
+struct mtk_poll_state {
+ struct netdev_queue *txq;
+ unsigned int total;
+ unsigned int done;
+ unsigned int bytes;
+};
+
+static void
+mtk_poll_tx_done(struct mtk_eth *eth, struct mtk_poll_state *state, u8 mac,
+ struct sk_buff *skb)
+{
+ struct netdev_queue *txq;
+ struct net_device *dev;
+ unsigned int bytes = skb->len;
+
+ state->total++;
+ eth->tx_packets++;
+ eth->tx_bytes += bytes;
+
+ dev = eth->netdev[mac];
+ if (!dev)
+ return;
+
+ txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb));
+ if (state->txq == txq) {
+ state->done++;
+ state->bytes += bytes;
+ return;
+ }
+
+ if (state->txq)
+ netdev_tx_completed_queue(state->txq, state->done, state->bytes);
+
+ state->txq = txq;
+ state->done = 1;
+ state->bytes = bytes;
+}
+
static int mtk_poll_tx_qdma(struct mtk_eth *eth, int budget,
- unsigned int *done, unsigned int *bytes)
+ struct mtk_poll_state *state)
{
const struct mtk_reg_map *reg_map = eth->soc->reg_map;
struct mtk_tx_ring *ring = &eth->tx_ring;
@@ -2026,12 +2142,9 @@ static int mtk_poll_tx_qdma(struct mtk_e
break;
if (tx_buf->data != (void *)MTK_DMA_DUMMY_DESC) {
- if (tx_buf->type == MTK_TYPE_SKB) {
- struct sk_buff *skb = tx_buf->data;
+ if (tx_buf->type == MTK_TYPE_SKB)
+ mtk_poll_tx_done(eth, state, mac, tx_buf->data);
- bytes[mac] += skb->len;
- done[mac]++;
- }
budget--;
}
mtk_tx_unmap(eth, tx_buf, &bq, true);
@@ -2050,7 +2163,7 @@ static int mtk_poll_tx_qdma(struct mtk_e
}
static int mtk_poll_tx_pdma(struct mtk_eth *eth, int budget,
- unsigned int *done, unsigned int *bytes)
+ struct mtk_poll_state *state)
{
struct mtk_tx_ring *ring = &eth->tx_ring;
struct mtk_tx_buf *tx_buf;
@@ -2068,12 +2181,8 @@ static int mtk_poll_tx_pdma(struct mtk_e
break;
if (tx_buf->data != (void *)MTK_DMA_DUMMY_DESC) {
- if (tx_buf->type == MTK_TYPE_SKB) {
- struct sk_buff *skb = tx_buf->data;
-
- bytes[0] += skb->len;
- done[0]++;
- }
+ if (tx_buf->type == MTK_TYPE_SKB)
+ mtk_poll_tx_done(eth, state, 0, tx_buf->data);
budget--;
}
mtk_tx_unmap(eth, tx_buf, &bq, true);
@@ -2095,26 +2204,15 @@ static int mtk_poll_tx(struct mtk_eth *e
{
struct mtk_tx_ring *ring = &eth->tx_ring;
struct dim_sample dim_sample = {};
- unsigned int done[MTK_MAX_DEVS];
- unsigned int bytes[MTK_MAX_DEVS];
- int total = 0, i;
-
- memset(done, 0, sizeof(done));
- memset(bytes, 0, sizeof(bytes));
+ struct mtk_poll_state state = {};
if (MTK_HAS_CAPS(eth->soc->caps, MTK_QDMA))
- budget = mtk_poll_tx_qdma(eth, budget, done, bytes);
+ budget = mtk_poll_tx_qdma(eth, budget, &state);
else
- budget = mtk_poll_tx_pdma(eth, budget, done, bytes);
+ budget = mtk_poll_tx_pdma(eth, budget, &state);
- for (i = 0; i < MTK_MAC_COUNT; i++) {
- if (!eth->netdev[i] || !done[i])
- continue;
- netdev_completed_queue(eth->netdev[i], done[i], bytes[i]);
- total += done[i];
- eth->tx_packets += done[i];
- eth->tx_bytes += bytes[i];
- }
+ if (state.txq)
+ netdev_tx_completed_queue(state.txq, state.done, state.bytes);
dim_update_sample(eth->tx_events, eth->tx_packets, eth->tx_bytes,
&dim_sample);
@@ -2124,7 +2222,7 @@ static int mtk_poll_tx(struct mtk_eth *e
(atomic_read(&ring->free_count) > ring->thresh))
mtk_wake_queue(eth);
- return total;
+ return state.total;
}
static void mtk_handle_status_irq(struct mtk_eth *eth)
@@ -2210,6 +2308,7 @@ static int mtk_tx_alloc(struct mtk_eth *
int i, sz = soc->txrx.txd_size;
struct mtk_tx_dma_v2 *txd;
int ring_size;
+ u32 ofs, val;
if (MTK_HAS_CAPS(soc->caps, MTK_QDMA))
ring_size = MTK_QDMA_RING_SIZE;
@@ -2277,8 +2376,25 @@ static int mtk_tx_alloc(struct mtk_eth *
ring->phys + ((ring_size - 1) * sz),
soc->reg_map->qdma.crx_ptr);
mtk_w32(eth, ring->last_free_ptr, soc->reg_map->qdma.drx_ptr);
- mtk_w32(eth, (QDMA_RES_THRES << 8) | QDMA_RES_THRES,
- soc->reg_map->qdma.qtx_cfg);
+
+ for (i = 0, ofs = 0; i < MTK_QDMA_NUM_QUEUES; i++) {
+ val = (QDMA_RES_THRES << 8) | QDMA_RES_THRES;
+ mtk_w32(eth, val, soc->reg_map->qdma.qtx_cfg + ofs);
+
+ val = MTK_QTX_SCH_MIN_RATE_EN |
+ /* minimum: 10 Mbps */
+ FIELD_PREP(MTK_QTX_SCH_MIN_RATE_MAN, 1) |
+ FIELD_PREP(MTK_QTX_SCH_MIN_RATE_EXP, 4) |
+ MTK_QTX_SCH_LEAKY_BUCKET_SIZE;
+ if (!MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2))
+ val |= MTK_QTX_SCH_LEAKY_BUCKET_EN;
+ mtk_w32(eth, val, soc->reg_map->qdma.qtx_sch + ofs);
+ ofs += MTK_QTX_OFFSET;
+ }
+ val = MTK_QDMA_TX_SCH_MAX_WFQ | (MTK_QDMA_TX_SCH_MAX_WFQ << 16);
+ mtk_w32(eth, val, soc->reg_map->qdma.tx_sch_rate);
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2))
+ mtk_w32(eth, val, soc->reg_map->qdma.tx_sch_rate + 4);
} else {
mtk_w32(eth, ring->phys_pdma, MT7628_TX_BASE_PTR0);
mtk_w32(eth, ring_size, MT7628_TX_MAX_CNT0);
@@ -2960,7 +3076,7 @@ static int mtk_start_dma(struct mtk_eth
if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2))
val |= MTK_MUTLI_CNT | MTK_RESV_BUF |
MTK_WCOMP_EN | MTK_DMAD_WR_WDONE |
- MTK_CHK_DDONE_EN;
+ MTK_CHK_DDONE_EN | MTK_LEAKY_BUCKET_EN;
else
val |= MTK_RX_BT_32DWORDS;
mtk_w32(eth, val, reg_map->qdma.glo_cfg);
@@ -3006,6 +3122,45 @@ static void mtk_gdm_config(struct mtk_et
mtk_w32(eth, 0, MTK_RST_GL);
}
+static int mtk_device_event(struct notifier_block *n, unsigned long event, void *ptr)
+{
+ struct mtk_mac *mac = container_of(n, struct mtk_mac, device_notifier);
+ struct mtk_eth *eth = mac->hw;
+ struct net_device *dev = netdev_notifier_info_to_dev(ptr);
+ struct ethtool_link_ksettings s;
+ struct net_device *ldev;
+ struct list_head *iter;
+ struct dsa_port *dp;
+
+ if (event != NETDEV_CHANGE)
+ return NOTIFY_DONE;
+
+ netdev_for_each_lower_dev(dev, ldev, iter) {
+ if (netdev_priv(ldev) == mac)
+ goto found;
+ }
+
+ return NOTIFY_DONE;
+
+found:
+ if (!dsa_slave_dev_check(dev))
+ return NOTIFY_DONE;
+
+ if (__ethtool_get_link_ksettings(dev, &s))
+ return NOTIFY_DONE;
+
+ if (s.base.speed == 0 || s.base.speed == ((__u32)-1))
+ return NOTIFY_DONE;
+
+ dp = dsa_port_from_netdev(dev);
+ if (dp->index >= MTK_QDMA_NUM_QUEUES)
+ return NOTIFY_DONE;
+
+ mtk_set_queue_speed(eth, dp->index + 3, s.base.speed);
+
+ return NOTIFY_DONE;
+}
+
static int mtk_open(struct net_device *dev)
{
struct mtk_mac *mac = netdev_priv(dev);
@@ -3048,7 +3203,8 @@ static int mtk_open(struct net_device *d
refcount_inc(&eth->dma_refcnt);
phylink_start(mac->phylink);
- netif_start_queue(dev);
+ netif_tx_start_all_queues(dev);
+
return 0;
}
@@ -3774,8 +3930,12 @@ static int mtk_unreg_dev(struct mtk_eth
int i;
for (i = 0; i < MTK_MAC_COUNT; i++) {
+ struct mtk_mac *mac;
if (!eth->netdev[i])
continue;
+ mac = netdev_priv(eth->netdev[i]);
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_QDMA))
+ unregister_netdevice_notifier(&mac->device_notifier);
unregister_netdev(eth->netdev[i]);
}
@@ -3992,6 +4152,23 @@ static int mtk_set_rxnfc(struct net_devi
return ret;
}
+static u16 mtk_select_queue(struct net_device *dev, struct sk_buff *skb,
+ struct net_device *sb_dev)
+{
+ struct mtk_mac *mac = netdev_priv(dev);
+ unsigned int queue = 0;
+
+ if (netdev_uses_dsa(dev))
+ queue = skb_get_queue_mapping(skb) + 3;
+ else
+ queue = mac->id;
+
+ if (queue >= dev->num_tx_queues)
+ queue = 0;
+
+ return queue;
+}
+
static const struct ethtool_ops mtk_ethtool_ops = {
.get_link_ksettings = mtk_get_link_ksettings,
.set_link_ksettings = mtk_set_link_ksettings,
@@ -4027,6 +4204,7 @@ static const struct net_device_ops mtk_n
.ndo_setup_tc = mtk_eth_setup_tc,
.ndo_bpf = mtk_xdp,
.ndo_xdp_xmit = mtk_xdp_xmit,
+ .ndo_select_queue = mtk_select_queue,
};
static int mtk_add_mac(struct mtk_eth *eth, struct device_node *np)
@@ -4036,6 +4214,7 @@ static int mtk_add_mac(struct mtk_eth *e
struct phylink *phylink;
struct mtk_mac *mac;
int id, err;
+ int txqs = 1;
if (!_id) {
dev_err(eth->dev, "missing mac id\n");
@@ -4053,7 +4232,10 @@ static int mtk_add_mac(struct mtk_eth *e
return -EINVAL;
}
- eth->netdev[id] = alloc_etherdev(sizeof(*mac));
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_QDMA))
+ txqs = MTK_QDMA_NUM_QUEUES;
+
+ eth->netdev[id] = alloc_etherdev_mqs(sizeof(*mac), txqs, 1);
if (!eth->netdev[id]) {
dev_err(eth->dev, "alloc_etherdev failed\n");
return -ENOMEM;
@@ -4150,6 +4332,11 @@ static int mtk_add_mac(struct mtk_eth *e
else
eth->netdev[id]->max_mtu = MTK_MAX_RX_LENGTH_2K - MTK_RX_ETH_HLEN;
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_QDMA)) {
+ mac->device_notifier.notifier_call = mtk_device_event;
+ register_netdevice_notifier(&mac->device_notifier);
+ }
+
return 0;
free_netdev:
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -22,6 +22,7 @@
#include <linux/bpf_trace.h>
#include "mtk_ppe.h"
+#define MTK_QDMA_NUM_QUEUES 16
#define MTK_QDMA_PAGE_SIZE 2048
#define MTK_MAX_RX_LENGTH 1536
#define MTK_MAX_RX_LENGTH_2K 2048
@@ -216,8 +217,26 @@
#define MTK_RING_MAX_AGG_CNT_H ((MTK_HW_LRO_MAX_AGG_CNT >> 6) & 0x3)
/* QDMA TX Queue Configuration Registers */
+#define MTK_QTX_OFFSET 0x10
#define QDMA_RES_THRES 4
+/* QDMA Tx Queue Scheduler Configuration Registers */
+#define MTK_QTX_SCH_TX_SEL BIT(31)
+#define MTK_QTX_SCH_TX_SEL_V2 GENMASK(31, 30)
+
+#define MTK_QTX_SCH_LEAKY_BUCKET_EN BIT(30)
+#define MTK_QTX_SCH_LEAKY_BUCKET_SIZE GENMASK(29, 28)
+#define MTK_QTX_SCH_MIN_RATE_EN BIT(27)
+#define MTK_QTX_SCH_MIN_RATE_MAN GENMASK(26, 20)
+#define MTK_QTX_SCH_MIN_RATE_EXP GENMASK(19, 16)
+#define MTK_QTX_SCH_MAX_RATE_WEIGHT GENMASK(15, 12)
+#define MTK_QTX_SCH_MAX_RATE_EN BIT(11)
+#define MTK_QTX_SCH_MAX_RATE_MAN GENMASK(10, 4)
+#define MTK_QTX_SCH_MAX_RATE_EXP GENMASK(3, 0)
+
+/* QDMA TX Scheduler Rate Control Register */
+#define MTK_QDMA_TX_SCH_MAX_WFQ BIT(15)
+
/* QDMA Global Configuration Register */
#define MTK_RX_2B_OFFSET BIT(31)
#define MTK_RX_BT_32DWORDS (3 << 11)
@@ -236,6 +255,7 @@
#define MTK_WCOMP_EN BIT(24)
#define MTK_RESV_BUF (0x40 << 16)
#define MTK_MUTLI_CNT (0x4 << 12)
+#define MTK_LEAKY_BUCKET_EN BIT(11)
/* QDMA Flow Control Register */
#define FC_THRES_DROP_MODE BIT(20)
@@ -266,8 +286,6 @@
#define MTK_STAT_OFFSET 0x40
/* QDMA TX NUM */
-#define MTK_QDMA_TX_NUM 16
-#define MTK_QDMA_TX_MASK (MTK_QDMA_TX_NUM - 1)
#define QID_BITS_V2(x) (((x) & 0x3f) << 16)
#define MTK_QDMA_GMAC2_QID 8
@@ -297,6 +315,7 @@
#define TX_DMA_PLEN0(x) (((x) & eth->soc->txrx.dma_max_len) << eth->soc->txrx.dma_len_offset)
#define TX_DMA_PLEN1(x) ((x) & eth->soc->txrx.dma_max_len)
#define TX_DMA_SWC BIT(14)
+#define TX_DMA_PQID GENMASK(3, 0)
/* PDMA on MT7628 */
#define TX_DMA_DONE BIT(31)
@@ -957,6 +976,7 @@ struct mtk_reg_map {
} pdma;
struct {
u32 qtx_cfg; /* tx queue configuration */
+ u32 qtx_sch; /* tx queue scheduler configuration */
u32 rx_ptr; /* rx base pointer */
u32 rx_cnt_cfg; /* rx max count configuration */
u32 qcrx_ptr; /* rx cpu pointer */
@@ -974,6 +994,7 @@ struct mtk_reg_map {
u32 fq_tail; /* fq tail pointer */
u32 fq_count; /* fq free page count */
u32 fq_blen; /* fq free page buffer length */
+ u32 tx_sch_rate; /* tx scheduler rate control registers */
} qdma;
u32 gdm1_cnt;
u32 gdma_to_ppe;
@@ -1177,6 +1198,7 @@ struct mtk_mac {
__be32 hwlro_ip[MTK_MAX_LRO_IP_CNT];
int hwlro_ip_cnt;
unsigned int syscfg0;
+ struct notifier_block device_notifier;
};
/* the struct describing the SoC. these are declared in the soc_xyz.c files */

20
target/linux/generic/backport-6.1/730-05-v6.3-net-dsa-tag_mtk-assign-per-port-queues.patch Normal file
View File

@ -0,0 +1,20 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Fri, 28 Oct 2022 18:16:03 +0200
Subject: [PATCH] net: dsa: tag_mtk: assign per-port queues
Keeps traffic sent to the switch within link speed limits
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/net/dsa/tag_mtk.c
+++ b/net/dsa/tag_mtk.c
@@ -25,6 +25,8 @@ static struct sk_buff *mtk_tag_xmit(stru
u8 xmit_tpid;
u8 *mtk_tag;
+ skb_set_queue_mapping(skb, dp->index);
+
/* Build the special tag after the MAC Source Address. If VLAN header
* is present, it's required that VLAN header and special tag is
* being combined. Only in this way we can allow the switch can parse

93
target/linux/generic/backport-6.1/730-06-v6.3-net-ethernet-mediatek-ppe-assign-per-port-queues-for.patch Normal file
View File

@ -0,0 +1,93 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Thu, 3 Nov 2022 17:49:44 +0100
Subject: [PATCH] net: ethernet: mediatek: ppe: assign per-port queues
for offloaded traffic
Keeps traffic sent to the switch within link speed limits
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_ppe.c
+++ b/drivers/net/ethernet/mediatek/mtk_ppe.c
@@ -399,6 +399,24 @@ int mtk_foe_entry_set_wdma(struct mtk_et
return 0;
}
+int mtk_foe_entry_set_queue(struct mtk_eth *eth, struct mtk_foe_entry *entry,
+ unsigned int queue)
+{
+ u32 *ib2 = mtk_foe_entry_ib2(eth, entry);
+
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
+ *ib2 &= ~MTK_FOE_IB2_QID_V2;
+ *ib2 |= FIELD_PREP(MTK_FOE_IB2_QID_V2, queue);
+ *ib2 |= MTK_FOE_IB2_PSE_QOS_V2;
+ } else {
+ *ib2 &= ~MTK_FOE_IB2_QID;
+ *ib2 |= FIELD_PREP(MTK_FOE_IB2_QID, queue);
+ *ib2 |= MTK_FOE_IB2_PSE_QOS;
+ }
+
+ return 0;
+}
+
static bool
mtk_flow_entry_match(struct mtk_eth *eth, struct mtk_flow_entry *entry,
struct mtk_foe_entry *data)
--- a/drivers/net/ethernet/mediatek/mtk_ppe.h
+++ b/drivers/net/ethernet/mediatek/mtk_ppe.h
@@ -68,7 +68,9 @@ enum {
#define MTK_FOE_IB2_DSCP GENMASK(31, 24)
/* CONFIG_MEDIATEK_NETSYS_V2 */
+#define MTK_FOE_IB2_QID_V2 GENMASK(6, 0)
#define MTK_FOE_IB2_PORT_MG_V2 BIT(7)
+#define MTK_FOE_IB2_PSE_QOS_V2 BIT(8)
#define MTK_FOE_IB2_DEST_PORT_V2 GENMASK(12, 9)
#define MTK_FOE_IB2_MULTICAST_V2 BIT(13)
#define MTK_FOE_IB2_WDMA_WINFO_V2 BIT(19)
@@ -351,6 +353,8 @@ int mtk_foe_entry_set_pppoe(struct mtk_e
int sid);
int mtk_foe_entry_set_wdma(struct mtk_eth *eth, struct mtk_foe_entry *entry,
int wdma_idx, int txq, int bss, int wcid);
+int mtk_foe_entry_set_queue(struct mtk_eth *eth, struct mtk_foe_entry *entry,
+ unsigned int queue);
int mtk_foe_entry_commit(struct mtk_ppe *ppe, struct mtk_flow_entry *entry);
void mtk_foe_entry_clear(struct mtk_ppe *ppe, struct mtk_flow_entry *entry);
int mtk_foe_entry_idle_time(struct mtk_ppe *ppe, struct mtk_flow_entry *entry);
--- a/drivers/net/ethernet/mediatek/mtk_ppe_offload.c
+++ b/drivers/net/ethernet/mediatek/mtk_ppe_offload.c
@@ -188,7 +188,7 @@ mtk_flow_set_output_device(struct mtk_et
int *wed_index)
{
struct mtk_wdma_info info = {};
- int pse_port, dsa_port;
+ int pse_port, dsa_port, queue;
if (mtk_flow_get_wdma_info(dev, dest_mac, &info) == 0) {
mtk_foe_entry_set_wdma(eth, foe, info.wdma_idx, info.queue,
@@ -212,8 +212,6 @@ mtk_flow_set_output_device(struct mtk_et
}
dsa_port = mtk_flow_get_dsa_port(&dev);
- if (dsa_port >= 0)
- mtk_foe_entry_set_dsa(eth, foe, dsa_port);
if (dev == eth->netdev[0])
pse_port = 1;
@@ -222,6 +220,14 @@ mtk_flow_set_output_device(struct mtk_et
else
return -EOPNOTSUPP;
+ if (dsa_port >= 0) {
+ mtk_foe_entry_set_dsa(eth, foe, dsa_port);
+ queue = 3 + dsa_port;
+ } else {
+ queue = pse_port - 1;
+ }
+ mtk_foe_entry_set_queue(eth, foe, queue);
+
out:
mtk_foe_entry_set_pse_port(eth, foe, pse_port);

28
target/linux/generic/backport-6.1/730-07-v6.3-net-ethernet-mtk_eth_soc-compile-out-netsys-v2-code-.patch Normal file
View File

@ -0,0 +1,28 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Thu, 27 Oct 2022 23:39:52 +0200
Subject: [PATCH] net: ethernet: mtk_eth_soc: compile out netsys v2 code
on mt7621
Avoid some branches in the hot path on low-end devices with limited CPU power,
and reduce code size
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -921,7 +921,13 @@ enum mkt_eth_capabilities {
#define MTK_MUX_GMAC12_TO_GEPHY_SGMII \
(MTK_ETH_MUX_GMAC12_TO_GEPHY_SGMII | MTK_MUX)
-#define MTK_HAS_CAPS(caps, _x) (((caps) & (_x)) == (_x))
+#ifdef CONFIG_SOC_MT7621
+#define MTK_CAP_MASK MTK_NETSYS_V2
+#else
+#define MTK_CAP_MASK 0
+#endif
+
+#define MTK_HAS_CAPS(caps, _x) (((caps) & (_x) & ~(MTK_CAP_MASK)) == (_x))
#define MT7621_CAPS (MTK_GMAC1_RGMII | MTK_GMAC1_TRGMII | \
MTK_GMAC2_RGMII | MTK_SHARED_INT | \

72
target/linux/generic/backport-6.1/730-08-v6.3-net-dsa-add-support-for-DSA-rx-offloading-via-metada.patch Normal file
View File

@ -0,0 +1,72 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Tue, 8 Nov 2022 15:03:15 +0100
Subject: [PATCH] net: dsa: add support for DSA rx offloading via
metadata dst
If a metadata dst is present with the type METADATA_HW_PORT_MUX on a dsa cpu
port netdev, assume that it carries the port number and that there is no DSA
tag present in the skb data.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -971,12 +971,14 @@ bool __skb_flow_dissect(const struct net
#if IS_ENABLED(CONFIG_NET_DSA)
if (unlikely(skb->dev && netdev_uses_dsa(skb->dev) &&
proto == htons(ETH_P_XDSA))) {
+ struct metadata_dst *md_dst = skb_metadata_dst(skb);
const struct dsa_device_ops *ops;
int offset = 0;
ops = skb->dev->dsa_ptr->tag_ops;
/* Only DSA header taggers break flow dissection */
- if (ops->needed_headroom) {
+ if (ops->needed_headroom &&
+ (!md_dst || md_dst->type != METADATA_HW_PORT_MUX)) {
if (ops->flow_dissect)
ops->flow_dissect(skb, &proto, &offset);
else
--- a/net/dsa/dsa.c
+++ b/net/dsa/dsa.c
@@ -11,6 +11,7 @@
#include <linux/netdevice.h>
#include <linux/sysfs.h>
#include <linux/ptp_classify.h>
+#include <net/dst_metadata.h>
#include "dsa_priv.h"
@@ -216,6 +217,7 @@ static bool dsa_skb_defer_rx_timestamp(s
static int dsa_switch_rcv(struct sk_buff *skb, struct net_device *dev,
struct packet_type *pt, struct net_device *unused)
{
+ struct metadata_dst *md_dst = skb_metadata_dst(skb);
struct dsa_port *cpu_dp = dev->dsa_ptr;
struct sk_buff *nskb = NULL;
struct dsa_slave_priv *p;
@@ -229,7 +231,22 @@ static int dsa_switch_rcv(struct sk_buff
if (!skb)
return 0;
- nskb = cpu_dp->rcv(skb, dev);
+ if (md_dst && md_dst->type == METADATA_HW_PORT_MUX) {
+ unsigned int port = md_dst->u.port_info.port_id;
+
+ skb_dst_drop(skb);
+ if (!skb_has_extensions(skb))
+ skb->slow_gro = 0;
+
+ skb->dev = dsa_master_find_slave(dev, 0, port);
+ if (likely(skb->dev)) {
+ dsa_default_offload_fwd_mark(skb);
+ nskb = skb;
+ }
+ } else {
+ nskb = cpu_dp->rcv(skb, dev);
+ }
+
if (!nskb) {
kfree_skb(skb);
return 0;

192
target/linux/generic/backport-6.1/730-09-v6.3-net-ethernet-mtk_eth_soc-fix-VLAN-rx-hardware-accele.patch Normal file
View File

@ -0,0 +1,192 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Fri, 28 Oct 2022 11:01:12 +0200
Subject: [PATCH] net: ethernet: mtk_eth_soc: fix VLAN rx hardware
acceleration
- enable VLAN untagging for PDMA rx
- make it possible to disable the feature via ethtool
- pass VLAN tag to the DSA driver
- untag special tag on PDMA only if no non-DSA devices are in use
- disable special tag untagging on 7986 for now, since it's not working yet
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -23,6 +23,7 @@
#include <linux/jhash.h>
#include <linux/bitfield.h>
#include <net/dsa.h>
+#include <net/dst_metadata.h>
#include "mtk_eth_soc.h"
#include "mtk_wed.h"
@@ -2022,16 +2023,22 @@ static int mtk_poll_rx(struct napi_struc
htons(RX_DMA_VPID(trxd.rxd4)),
RX_DMA_VID(trxd.rxd4));
} else if (trxd.rxd2 & RX_DMA_VTAG) {
- __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
+ __vlan_hwaccel_put_tag(skb, htons(RX_DMA_VPID(trxd.rxd3)),
RX_DMA_VID(trxd.rxd3));
}
+ }
+
+ /* When using VLAN untagging in combination with DSA, the
+ * hardware treats the MTK special tag as a VLAN and untags it.
+ */
+ if (skb_vlan_tag_present(skb) && netdev_uses_dsa(netdev)) {
+ unsigned int port = ntohs(skb->vlan_proto) & GENMASK(2, 0);
- /* If the device is attached to a dsa switch, the special
- * tag inserted in VLAN field by hw switch can * be offloaded
- * by RX HW VLAN offload. Clear vlan info.
- */
- if (netdev_uses_dsa(netdev))
- __vlan_hwaccel_clear_tag(skb);
+ if (port < ARRAY_SIZE(eth->dsa_meta) &&
+ eth->dsa_meta[port])
+ skb_dst_set_noref(skb, &eth->dsa_meta[port]->dst);
+
+ __vlan_hwaccel_clear_tag(skb);
}
skb_record_rx_queue(skb, 0);
@@ -2856,15 +2863,30 @@ static netdev_features_t mtk_fix_feature
static int mtk_set_features(struct net_device *dev, netdev_features_t features)
{
- int err = 0;
+ struct mtk_mac *mac = netdev_priv(dev);
+ struct mtk_eth *eth = mac->hw;
+ netdev_features_t diff = dev->features ^ features;
+ int i;
+
+ if ((diff & NETIF_F_LRO) && !(features & NETIF_F_LRO))
+ mtk_hwlro_netdev_disable(dev);
- if (!((dev->features ^ features) & NETIF_F_LRO))
+ /* Set RX VLAN offloading */
+ if (!(diff & NETIF_F_HW_VLAN_CTAG_RX))
return 0;
- if (!(features & NETIF_F_LRO))
- mtk_hwlro_netdev_disable(dev);
+ mtk_w32(eth, !!(features & NETIF_F_HW_VLAN_CTAG_RX),
+ MTK_CDMP_EG_CTRL);
- return err;
+ /* sync features with other MAC */
+ for (i = 0; i < MTK_MAC_COUNT; i++) {
+ if (!eth->netdev[i] || eth->netdev[i] == dev)
+ continue;
+ eth->netdev[i]->features &= ~NETIF_F_HW_VLAN_CTAG_RX;
+ eth->netdev[i]->features |= features & NETIF_F_HW_VLAN_CTAG_RX;
+ }
+
+ return 0;
}
/* wait for DMA to finish whatever it is doing before we start using it again */
@@ -3161,11 +3183,45 @@ found:
return NOTIFY_DONE;
}
+static bool mtk_uses_dsa(struct net_device *dev)
+{
+#if IS_ENABLED(CONFIG_NET_DSA)
+ return netdev_uses_dsa(dev) &&
+ dev->dsa_ptr->tag_ops->proto == DSA_TAG_PROTO_MTK;
+#else
+ return false;
+#endif
+}
+
static int mtk_open(struct net_device *dev)
{
struct mtk_mac *mac = netdev_priv(dev);
struct mtk_eth *eth = mac->hw;
- int err;
+ int i, err;
+
+ if (mtk_uses_dsa(dev) && !eth->prog) {
+ for (i = 0; i < ARRAY_SIZE(eth->dsa_meta); i++) {
+ struct metadata_dst *md_dst = eth->dsa_meta[i];
+
+ if (md_dst)
+ continue;
+
+ md_dst = metadata_dst_alloc(0, METADATA_HW_PORT_MUX,
+ GFP_KERNEL);
+ if (!md_dst)
+ return -ENOMEM;
+
+ md_dst->u.port_info.port_id = i;
+ eth->dsa_meta[i] = md_dst;
+ }
+ } else {
+ /* Hardware special tag parsing needs to be disabled if at least
+ * one MAC does not use DSA.
+ */
+ u32 val = mtk_r32(eth, MTK_CDMP_IG_CTRL);
+ val &= ~MTK_CDMP_STAG_EN;
+ mtk_w32(eth, val, MTK_CDMP_IG_CTRL);
+ }
err = phylink_of_phy_connect(mac->phylink, mac->of_node, 0);
if (err) {
@@ -3686,6 +3742,10 @@ static int mtk_hw_init(struct mtk_eth *e
*/
val = mtk_r32(eth, MTK_CDMQ_IG_CTRL);
mtk_w32(eth, val | MTK_CDMQ_STAG_EN, MTK_CDMQ_IG_CTRL);
+ if (!MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
+ val = mtk_r32(eth, MTK_CDMP_IG_CTRL);
+ mtk_w32(eth, val | MTK_CDMP_STAG_EN, MTK_CDMP_IG_CTRL);
+ }
/* Enable RX VLan Offloading */
mtk_w32(eth, 1, MTK_CDMP_EG_CTRL);
@@ -3922,6 +3982,12 @@ static int mtk_free_dev(struct mtk_eth *
free_netdev(eth->netdev[i]);
}
+ for (i = 0; i < ARRAY_SIZE(eth->dsa_meta); i++) {
+ if (!eth->dsa_meta[i])
+ break;
+ metadata_dst_free(eth->dsa_meta[i]);
+ }
+
return 0;
}
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -22,6 +22,9 @@
#include <linux/bpf_trace.h>
#include "mtk_ppe.h"
+#define MTK_MAX_DSA_PORTS 7
+#define MTK_DSA_PORT_MASK GENMASK(2, 0)
+
#define MTK_QDMA_NUM_QUEUES 16
#define MTK_QDMA_PAGE_SIZE 2048
#define MTK_MAX_RX_LENGTH 1536
@@ -105,6 +108,9 @@
#define MTK_CDMQ_IG_CTRL 0x1400
#define MTK_CDMQ_STAG_EN BIT(0)
+/* CDMQ Exgress Control Register */
+#define MTK_CDMQ_EG_CTRL 0x1404
+
/* CDMP Ingress Control Register */
#define MTK_CDMP_IG_CTRL 0x400
#define MTK_CDMP_STAG_EN BIT(0)
@@ -1170,6 +1176,8 @@ struct mtk_eth {
int ip_align;
+ struct metadata_dst *dsa_meta[MTK_MAX_DSA_PORTS];
+
struct mtk_ppe *ppe[2];
struct rhashtable flow_table;

42
target/linux/generic/backport-6.1/730-12-v6.3-net-ethernet-mtk_eth_soc-disable-hardware-DSA-untagg.patch Normal file
View File

@ -0,0 +1,42 @@
From: =?UTF-8?q?Ar=C4=B1n=C3=A7=20=C3=9CNAL?= <arinc.unal@arinc9.com>
Date: Sat, 28 Jan 2023 12:42:32 +0300
Subject: [PATCH] net: ethernet: mtk_eth_soc: disable hardware DSA untagging
for second MAC
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
According to my tests on MT7621AT and MT7623NI SoCs, hardware DSA untagging
won't work on the second MAC. Therefore, disable this feature when the
second MAC of the MT7621 and MT7623 SoCs is being used.
Fixes: 2d7605a72906 ("net: ethernet: mtk_eth_soc: enable hardware DSA untagging")
Link: https://lore.kernel.org/netdev/6249fc14-b38a-c770-36b4-5af6d41c21d3@arinc9.com/
Tested-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Link: https://lore.kernel.org/r/20230128094232.2451947-1-arinc.unal@arinc9.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -3199,7 +3199,8 @@ static int mtk_open(struct net_device *d
struct mtk_eth *eth = mac->hw;
int i, err;
- if (mtk_uses_dsa(dev) && !eth->prog) {
+ if ((mtk_uses_dsa(dev) && !eth->prog) &&
+ !(mac->id == 1 && MTK_HAS_CAPS(eth->soc->caps, MTK_GMAC1_TRGMII))) {
for (i = 0; i < ARRAY_SIZE(eth->dsa_meta); i++) {
struct metadata_dst *md_dst = eth->dsa_meta[i];
@@ -3216,7 +3217,8 @@ static int mtk_open(struct net_device *d
}
} else {
/* Hardware special tag parsing needs to be disabled if at least
- * one MAC does not use DSA.
+ * one MAC does not use DSA, or the second MAC of the MT7621 and
+ * MT7623 SoCs is being used.
*/
u32 val = mtk_r32(eth, MTK_CDMP_IG_CTRL);
val &= ~MTK_CDMP_STAG_EN;

54
target/linux/generic/backport-6.1/730-13-v6.3-net-ethernet-mtk_eth_soc-enable-special-tag-when-any.patch Normal file
View File

@ -0,0 +1,54 @@
From: =?UTF-8?q?Ar=C4=B1n=C3=A7=20=C3=9CNAL?= <arinc.unal@arinc9.com>
Date: Sun, 5 Feb 2023 20:53:31 +0300
Subject: [PATCH] net: ethernet: mtk_eth_soc: enable special tag when any MAC
uses DSA
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The special tag is only enabled when the first MAC uses DSA. However, it
must be enabled when any MAC uses DSA. Change the check accordingly.
This fixes hardware DSA untagging not working on the second MAC of the
MT7621 and MT7623 SoCs, and likely other SoCs too. Therefore, remove the
check that disables hardware DSA untagging for the second MAC of the MT7621
and MT7623 SoCs.
Fixes: a1f47752fd62 ("net: ethernet: mtk_eth_soc: disable hardware DSA untagging for second MAC")
Co-developed-by: Richard van Schagen <richard@routerhints.com>
Signed-off-by: Richard van Schagen <richard@routerhints.com>
Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -3134,7 +3134,7 @@ static void mtk_gdm_config(struct mtk_et
val |= config;
- if (!i && eth->netdev[0] && netdev_uses_dsa(eth->netdev[0]))
+ if (eth->netdev[i] && netdev_uses_dsa(eth->netdev[i]))
val |= MTK_GDMA_SPECIAL_TAG;
mtk_w32(eth, val, MTK_GDMA_FWD_CFG(i));
@@ -3199,8 +3199,7 @@ static int mtk_open(struct net_device *d
struct mtk_eth *eth = mac->hw;
int i, err;
- if ((mtk_uses_dsa(dev) && !eth->prog) &&
- !(mac->id == 1 && MTK_HAS_CAPS(eth->soc->caps, MTK_GMAC1_TRGMII))) {
+ if (mtk_uses_dsa(dev) && !eth->prog) {
for (i = 0; i < ARRAY_SIZE(eth->dsa_meta); i++) {
struct metadata_dst *md_dst = eth->dsa_meta[i];
@@ -3217,8 +3216,7 @@ static int mtk_open(struct net_device *d
}
} else {
/* Hardware special tag parsing needs to be disabled if at least
- * one MAC does not use DSA, or the second MAC of the MT7621 and
- * MT7623 SoCs is being used.
+ * one MAC does not use DSA.
*/
u32 val = mtk_r32(eth, MTK_CDMP_IG_CTRL);
val &= ~MTK_CDMP_STAG_EN;

129
target/linux/generic/backport-6.1/730-14-v6.3-net-ethernet-mtk_eth_soc-fix-DSA-TX-tag-hwaccel-for-.patch Normal file
View File

@ -0,0 +1,129 @@
From: Vladimir Oltean <vladimir.oltean@nxp.com>
Date: Tue, 7 Feb 2023 12:30:27 +0200
Subject: [PATCH] net: ethernet: mtk_eth_soc: fix DSA TX tag hwaccel for switch
port 0
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Arınç reports that on his MT7621AT Unielec U7621-06 board and MT7623NI
Bananapi BPI-R2, packets received by the CPU over mt7530 switch port 0
(of which this driver acts as the DSA master) are not processed
correctly by software. More precisely, they arrive without a DSA tag
(in packet or in the hwaccel area - skb_metadata_dst()), so DSA cannot
demux them towards the switch's interface for port 0. Traffic from other
ports receives a skb_metadata_dst() with the correct port and is demuxed
properly.
Looking at mtk_poll_rx(), it becomes apparent that this driver uses the
skb vlan hwaccel area:
union {
u32 vlan_all;
struct {
__be16 vlan_proto;
__u16 vlan_tci;
};
};
as a temporary storage for the VLAN hwaccel tag, or the DSA hwaccel tag.
If this is a DSA master it's a DSA hwaccel tag, and finally clears up
the skb VLAN hwaccel header.
I'm guessing that the problem is the (mis)use of API.
skb_vlan_tag_present() looks like this:
#define skb_vlan_tag_present(__skb) (!!(__skb)->vlan_all)
So if both vlan_proto and vlan_tci are zeroes, skb_vlan_tag_present()
returns precisely false. I don't know for sure what is the format of the
DSA hwaccel tag, but I surely know that lowermost 3 bits of vlan_proto
are 0 when receiving from port 0:
unsigned int port = vlan_proto & GENMASK(2, 0);
If the RX descriptor has no other bits set to non-zero values in
RX_DMA_VTAG, then the call to __vlan_hwaccel_put_tag() will not, in
fact, make the subsequent skb_vlan_tag_present() return true, because
it's implemented like this:
static inline void __vlan_hwaccel_put_tag(struct sk_buff *skb,
__be16 vlan_proto, u16 vlan_tci)
{
skb->vlan_proto = vlan_proto;
skb->vlan_tci = vlan_tci;
}
What we need to do to fix this problem (assuming this is the problem) is
to stop using skb->vlan_all as temporary storage for driver affairs, and
just create some local variables that serve the same purpose, but
hopefully better. Instead of calling skb_vlan_tag_present(), let's look
at a boolean has_hwaccel_tag which we set to true when the RX DMA
descriptors have something. Disambiguate based on netdev_uses_dsa()
whether this is a VLAN or DSA hwaccel tag, and only call
__vlan_hwaccel_put_tag() if we're certain it's a VLAN tag.
Arınç confirms that the treatment works, so this validates the
assumption.
Link: https://lore.kernel.org/netdev/704f3a72-fc9e-714a-db54-272e17612637@arinc9.com/
Fixes: 2d7605a72906 ("net: ethernet: mtk_eth_soc: enable hardware DSA untagging")
Reported-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Tested-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -1878,7 +1878,9 @@ static int mtk_poll_rx(struct napi_struc
while (done < budget) {
unsigned int pktlen, *rxdcsum;
+ bool has_hwaccel_tag = false;
struct net_device *netdev;
+ u16 vlan_proto, vlan_tci;
dma_addr_t dma_addr;
u32 hash, reason;
int mac = 0;
@@ -2018,27 +2020,29 @@ static int mtk_poll_rx(struct napi_struc
if (netdev->features & NETIF_F_HW_VLAN_CTAG_RX) {
if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
- if (trxd.rxd3 & RX_DMA_VTAG_V2)
- __vlan_hwaccel_put_tag(skb,
- htons(RX_DMA_VPID(trxd.rxd4)),
- RX_DMA_VID(trxd.rxd4));
+ if (trxd.rxd3 & RX_DMA_VTAG_V2) {
+ vlan_proto = RX_DMA_VPID(trxd.rxd4);
+ vlan_tci = RX_DMA_VID(trxd.rxd4);
+ has_hwaccel_tag = true;
+ }
} else if (trxd.rxd2 & RX_DMA_VTAG) {
- __vlan_hwaccel_put_tag(skb, htons(RX_DMA_VPID(trxd.rxd3)),
- RX_DMA_VID(trxd.rxd3));
+ vlan_proto = RX_DMA_VPID(trxd.rxd3);
+ vlan_tci = RX_DMA_VID(trxd.rxd3);
+ has_hwaccel_tag = true;
}
}
/* When using VLAN untagging in combination with DSA, the
* hardware treats the MTK special tag as a VLAN and untags it.
*/
- if (skb_vlan_tag_present(skb) && netdev_uses_dsa(netdev)) {
- unsigned int port = ntohs(skb->vlan_proto) & GENMASK(2, 0);
+ if (has_hwaccel_tag && netdev_uses_dsa(netdev)) {
+ unsigned int port = vlan_proto & GENMASK(2, 0);
if (port < ARRAY_SIZE(eth->dsa_meta) &&
eth->dsa_meta[port])
skb_dst_set_noref(skb, &eth->dsa_meta[port]->dst);
-
- __vlan_hwaccel_clear_tag(skb);
+ } else if (has_hwaccel_tag) {
+ __vlan_hwaccel_put_tag(skb, htons(vlan_proto), vlan_tci);
}
skb_record_rx_queue(skb, 0);

26
target/linux/generic/backport-6.1/730-15-v6.3-net-ethernet-mtk_wed-No-need-to-clear-memory-after-a.patch Normal file
View File

@ -0,0 +1,26 @@
From: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Date: Sun, 12 Feb 2023 07:51:51 +0100
Subject: [PATCH] net: ethernet: mtk_wed: No need to clear memory after a
dma_alloc_coherent() call
dma_alloc_coherent() already clears the allocated memory, there is no need
to explicitly call memset().
Moreover, it is likely that the size in the memset() is incorrect and
should be "size * sizeof(*ring->desc)".
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/d5acce7dd108887832c9719f62c7201b4c83b3fb.1676184599.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -779,7 +779,6 @@ mtk_wed_rro_ring_alloc(struct mtk_wed_de
ring->desc_size = sizeof(*ring->desc);
ring->size = size;
- memset(ring->desc, 0, size);
return 0;
}

61
target/linux/generic/backport-6.1/730-16-v6.3-net-ethernet-mtk_wed-fix-some-possible-NULL-pointer-.patch Normal file
View File

@ -0,0 +1,61 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Wed, 7 Dec 2022 15:04:54 +0100
Subject: [PATCH] net: ethernet: mtk_wed: fix some possible NULL pointer
dereferences
Fix possible NULL pointer dereference in mtk_wed_detach routine checking
wo pointer is properly allocated before running mtk_wed_wo_reset() and
mtk_wed_wo_deinit().
Even if it is just a theoretical issue at the moment check wo pointer is
not NULL in mtk_wed_mcu_msg_update.
Moreover, honor mtk_wed_mcu_send_msg return value in mtk_wed_wo_reset()
Fixes: 799684448e3e ("net: ethernet: mtk_wed: introduce wed wo support")
Fixes: 4c5de09eb0d0 ("net: ethernet: mtk_wed: add configure wed wo support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -174,9 +174,10 @@ mtk_wed_wo_reset(struct mtk_wed_device *
mtk_wdma_tx_reset(dev);
mtk_wed_reset(dev, MTK_WED_RESET_WED);
- mtk_wed_mcu_send_msg(wo, MTK_WED_MODULE_ID_WO,
- MTK_WED_WO_CMD_CHANGE_STATE, &state,
- sizeof(state), false);
+ if (mtk_wed_mcu_send_msg(wo, MTK_WED_MODULE_ID_WO,
+ MTK_WED_WO_CMD_CHANGE_STATE, &state,
+ sizeof(state), false))
+ return;
if (readx_poll_timeout(mtk_wed_wo_read_status, dev, val,
val == MTK_WED_WOIF_DISABLE_DONE,
@@ -632,9 +633,11 @@ mtk_wed_detach(struct mtk_wed_device *de
mtk_wed_free_tx_rings(dev);
if (mtk_wed_get_rx_capa(dev)) {
- mtk_wed_wo_reset(dev);
+ if (hw->wed_wo)
+ mtk_wed_wo_reset(dev);
mtk_wed_free_rx_rings(dev);
- mtk_wed_wo_deinit(hw);
+ if (hw->wed_wo)
+ mtk_wed_wo_deinit(hw);
}
if (dev->wlan.bus_type == MTK_WED_BUS_PCIE) {
--- a/drivers/net/ethernet/mediatek/mtk_wed_mcu.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed_mcu.c
@@ -207,6 +207,9 @@ int mtk_wed_mcu_msg_update(struct mtk_we
if (dev->hw->version == 1)
return 0;
+ if (WARN_ON(!wo))
+ return -ENODEV;
+
return mtk_wed_mcu_send_msg(wo, MTK_WED_MODULE_ID_WO, id, data, len,
true);
}

58
target/linux/generic/backport-6.1/730-17-v6.3-net-ethernet-mtk_wed-fix-possible-deadlock-if-mtk_we.patch Normal file
View File

@ -0,0 +1,58 @@
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Wed, 7 Dec 2022 15:04:55 +0100
Subject: [PATCH] net: ethernet: mtk_wed: fix possible deadlock if
mtk_wed_wo_init fails
Introduce __mtk_wed_detach() in order to avoid a deadlock in
mtk_wed_attach routine if mtk_wed_wo_init fails since both
mtk_wed_attach and mtk_wed_detach run holding hw_lock mutex.
Fixes: 4c5de09eb0d0 ("net: ethernet: mtk_wed: add configure wed wo support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
--- a/drivers/net/ethernet/mediatek/mtk_wed.c
+++ b/drivers/net/ethernet/mediatek/mtk_wed.c
@@ -619,12 +619,10 @@ mtk_wed_deinit(struct mtk_wed_device *de
}
static void
-mtk_wed_detach(struct mtk_wed_device *dev)
+__mtk_wed_detach(struct mtk_wed_device *dev)
{
struct mtk_wed_hw *hw = dev->hw;
- mutex_lock(&hw_lock);
-
mtk_wed_deinit(dev);
mtk_wdma_rx_reset(dev);
@@ -657,6 +655,13 @@ mtk_wed_detach(struct mtk_wed_device *de
module_put(THIS_MODULE);
hw->wed_dev = NULL;
+}
+
+static void
+mtk_wed_detach(struct mtk_wed_device *dev)
+{
+ mutex_lock(&hw_lock);
+ __mtk_wed_detach(dev);
mutex_unlock(&hw_lock);
}
@@ -1538,8 +1543,10 @@ mtk_wed_attach(struct mtk_wed_device *de
ret = mtk_wed_wo_init(hw);
}
out:
- if (ret)
- mtk_wed_detach(dev);
+ if (ret) {
+ dev_err(dev->hw->dev, "failed to attach wed device\n");
+ __mtk_wed_detach(dev);
+ }
unlock:
mutex_unlock(&hw_lock);

31
target/linux/generic/backport-6.1/730-18-v6.3-net-ethernet-mtk_eth_soc-fix-tx-throughput-regressio.patch Normal file
View File

@ -0,0 +1,31 @@
From: Felix Fietkau <nbd@nbd.name>
Date: Fri, 24 Mar 2023 14:56:58 +0100
Subject: [PATCH] net: ethernet: mtk_eth_soc: fix tx throughput regression with
direct 1G links
Using the QDMA tx scheduler to throttle tx to line speed works fine for
switch ports, but apparently caused a regression on non-switch ports.
Based on a number of tests, it seems that this throttling can be safely
dropped without re-introducing the issues on switch ports that the
tx scheduling changes resolved.
Link: https://lore.kernel.org/netdev/trinity-92c3826f-c2c8-40af-8339-bc6d0d3ffea4-1678213958520@3c-app-gmx-bs16/
Fixes: f63959c7eec3 ("net: ethernet: mtk_eth_soc: implement multi-queue support for per-port queues")
Reported-by: Frank Wunderlich <frank-w@public-files.de>
Reported-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
@@ -719,8 +719,6 @@ static void mtk_mac_link_up(struct phyli
break;
}
- mtk_set_queue_speed(mac->hw, mac->id, speed);
-
/* Configure duplex */
if (duplex == DUPLEX_FULL)
mcr |= MAC_MCR_FORCE_DPX;

55
target/linux/generic/backport-6.1/733-v6.2-02-net-mtk_eth_soc-add-definitions-for-PCS.patch Normal file
View File

@ -0,0 +1,55 @@
From b6a709cb51f7bdc55c01cec886098a9753ce8c28 Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:10:42 +0100
Subject: [PATCH 01/10] net: mtk_eth_soc: add definitions for PCS
As a result of help from Frank Wunderlich to investigate and test, we
know a bit more about the PCS on the Mediatek platforms. Update the
definitions from this investigation.
This PCS appears similar, but not identical to the Lynx PCS.
Although not included in this patch, but for future reference, the PHY
ID registers at offset 4 read as 0x4d544950 'MTIP'.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_eth_soc.h | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -504,8 +504,10 @@
#define ETHSYS_DMA_AG_MAP_PPE BIT(2)
/* SGMII subsystem config registers */
-/* Register to auto-negotiation restart */
+/* BMCR (low 16) BMSR (high 16) */
#define SGMSYS_PCS_CONTROL_1 0x0
+#define SGMII_BMCR GENMASK(15, 0)
+#define SGMII_BMSR GENMASK(31, 16)
#define SGMII_AN_RESTART BIT(9)
#define SGMII_ISOLATE BIT(10)
#define SGMII_AN_ENABLE BIT(12)
@@ -515,13 +517,18 @@
#define SGMII_PCS_FAULT BIT(23)
#define SGMII_AN_EXPANSION_CLR BIT(30)
+#define SGMSYS_PCS_ADVERTISE 0x8
+#define SGMII_ADVERTISE GENMASK(15, 0)
+#define SGMII_LPA GENMASK(31, 16)
+
/* Register to programmable link timer, the unit in 2 * 8ns */
#define SGMSYS_PCS_LINK_TIMER 0x18
-#define SGMII_LINK_TIMER_DEFAULT (0x186a0 & GENMASK(19, 0))
+#define SGMII_LINK_TIMER_MASK GENMASK(19, 0)
+#define SGMII_LINK_TIMER_DEFAULT (0x186a0 & SGMII_LINK_TIMER_MASK)
/* Register to control remote fault */
#define SGMSYS_SGMII_MODE 0x20
-#define SGMII_IF_MODE_BIT0 BIT(0)
+#define SGMII_IF_MODE_SGMII BIT(0)
#define SGMII_SPEED_DUPLEX_AN BIT(1)
#define SGMII_SPEED_MASK GENMASK(3, 2)
#define SGMII_SPEED_10 FIELD_PREP(SGMII_SPEED_MASK, 0)

74
target/linux/generic/backport-6.1/733-v6.2-03-net-mtk_eth_soc-eliminate-unnecessary-error-handling.patch Normal file
View File

@ -0,0 +1,74 @@
From 5cf7797526ee81bea0f627bccaa3d887f48f53e0 Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:10:47 +0100
Subject: [PATCH 02/10] net: mtk_eth_soc: eliminate unnecessary error handling
The functions called by the pcs_config() method always return zero, so
there is no point trying to handle an error from these functions. Make
these functions void, eliminate the "err" variable and simply return
zero from the pcs_config() function itself.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 18 ++++++------------
1 file changed, 6 insertions(+), 12 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -20,7 +20,7 @@ static struct mtk_pcs *pcs_to_mtk_pcs(st
}
/* For SGMII interface mode */
-static int mtk_pcs_setup_mode_an(struct mtk_pcs *mpcs)
+static void mtk_pcs_setup_mode_an(struct mtk_pcs *mpcs)
{
unsigned int val;
@@ -39,16 +39,13 @@ static int mtk_pcs_setup_mode_an(struct
regmap_read(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, &val);
val &= ~SGMII_PHYA_PWD;
regmap_write(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, val);
-
- return 0;
-
}
/* For 1000BASE-X and 2500BASE-X interface modes, which operate at a
* fixed speed.
*/
-static int mtk_pcs_setup_mode_force(struct mtk_pcs *mpcs,
- phy_interface_t interface)
+static void mtk_pcs_setup_mode_force(struct mtk_pcs *mpcs,
+ phy_interface_t interface)
{
unsigned int val;
@@ -73,8 +70,6 @@ static int mtk_pcs_setup_mode_force(stru
regmap_read(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, &val);
val &= ~SGMII_PHYA_PWD;
regmap_write(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, val);
-
- return 0;
}
static int mtk_pcs_config(struct phylink_pcs *pcs, unsigned int mode,
@@ -83,15 +78,14 @@ static int mtk_pcs_config(struct phylink
bool permit_pause_to_mac)
{
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
- int err = 0;
/* Setup SGMIISYS with the determined property */
if (interface != PHY_INTERFACE_MODE_SGMII)
- err = mtk_pcs_setup_mode_force(mpcs, interface);
+ mtk_pcs_setup_mode_force(mpcs, interface);
else if (phylink_autoneg_inband(mode))
- err = mtk_pcs_setup_mode_an(mpcs);
+ mtk_pcs_setup_mode_an(mpcs);
- return err;
+ return 0;
}
static void mtk_pcs_restart_an(struct phylink_pcs *pcs)

46
target/linux/generic/backport-6.1/733-v6.2-04-net-mtk_eth_soc-add-pcs_get_state-implementation.patch Normal file
View File

@ -0,0 +1,46 @@
From c000dca098002da193b98099df051c9ead0cacb4 Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:10:52 +0100
Subject: [PATCH 03/10] net: mtk_eth_soc: add pcs_get_state() implementation
Add a pcs_get_state() implementation which uses the advertisements
to compute the resulting link modes, and BMSR contents to determine
negotiation and link status.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 15 +++++++++++++++
1 file changed, 15 insertions(+)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -19,6 +19,20 @@ static struct mtk_pcs *pcs_to_mtk_pcs(st
return container_of(pcs, struct mtk_pcs, pcs);
}
+static void mtk_pcs_get_state(struct phylink_pcs *pcs,
+ struct phylink_link_state *state)
+{
+ struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
+ unsigned int bm, adv;
+
+ /* Read the BMSR and LPA */
+ regmap_read(mpcs->regmap, SGMSYS_PCS_CONTROL_1, &bm);
+ regmap_read(mpcs->regmap, SGMSYS_PCS_ADVERTISE, &adv);
+
+ phylink_mii_c22_pcs_decode_state(state, FIELD_GET(SGMII_BMSR, bm),
+ FIELD_GET(SGMII_LPA, adv));
+}
+
/* For SGMII interface mode */
static void mtk_pcs_setup_mode_an(struct mtk_pcs *mpcs)
{
@@ -117,6 +131,7 @@ static void mtk_pcs_link_up(struct phyli
}
static const struct phylink_pcs_ops mtk_pcs_ops = {
+ .pcs_get_state = mtk_pcs_get_state,
.pcs_config = mtk_pcs_config,
.pcs_an_restart = mtk_pcs_restart_an,
.pcs_link_up = mtk_pcs_link_up,

130
target/linux/generic/backport-6.1/733-v6.2-05-net-mtk_eth_soc-convert-mtk_sgmii-to-use-regmap_upda.patch Normal file
View File

@ -0,0 +1,130 @@
From 0d2351dc2768061689abd4de1529fa206bbd574e Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:10:58 +0100
Subject: [PATCH 04/10] net: mtk_eth_soc: convert mtk_sgmii to use
regmap_update_bits()
mtk_sgmii does a lot of read-modify-write operations, for which there
is a specific regmap function. Use this function instead of open-coding
the operations.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 61 ++++++++++-------------
1 file changed, 26 insertions(+), 35 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -36,23 +36,18 @@ static void mtk_pcs_get_state(struct phy
/* For SGMII interface mode */
static void mtk_pcs_setup_mode_an(struct mtk_pcs *mpcs)
{
- unsigned int val;
-
/* Setup the link timer and QPHY power up inside SGMIISYS */
regmap_write(mpcs->regmap, SGMSYS_PCS_LINK_TIMER,
SGMII_LINK_TIMER_DEFAULT);
- regmap_read(mpcs->regmap, SGMSYS_SGMII_MODE, &val);
- val |= SGMII_REMOTE_FAULT_DIS;
- regmap_write(mpcs->regmap, SGMSYS_SGMII_MODE, val);
-
- regmap_read(mpcs->regmap, SGMSYS_PCS_CONTROL_1, &val);
- val |= SGMII_AN_RESTART;
- regmap_write(mpcs->regmap, SGMSYS_PCS_CONTROL_1, val);
-
- regmap_read(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, &val);
- val &= ~SGMII_PHYA_PWD;
- regmap_write(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, val);
+ regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
+ SGMII_REMOTE_FAULT_DIS, SGMII_REMOTE_FAULT_DIS);
+
+ regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
+ SGMII_AN_RESTART, SGMII_AN_RESTART);
+
+ regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
+ SGMII_PHYA_PWD, 0);
}
/* For 1000BASE-X and 2500BASE-X interface modes, which operate at a
@@ -61,29 +56,26 @@ static void mtk_pcs_setup_mode_an(struct
static void mtk_pcs_setup_mode_force(struct mtk_pcs *mpcs,
phy_interface_t interface)
{
- unsigned int val;
+ unsigned int rgc3;
- regmap_read(mpcs->regmap, mpcs->ana_rgc3, &val);
- val &= ~RG_PHY_SPEED_MASK;
if (interface == PHY_INTERFACE_MODE_2500BASEX)
- val |= RG_PHY_SPEED_3_125G;
- regmap_write(mpcs->regmap, mpcs->ana_rgc3, val);
+ rgc3 = RG_PHY_SPEED_3_125G;
+
+ regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
+ RG_PHY_SPEED_3_125G, rgc3);
/* Disable SGMII AN */
- regmap_read(mpcs->regmap, SGMSYS_PCS_CONTROL_1, &val);
- val &= ~SGMII_AN_ENABLE;
- regmap_write(mpcs->regmap, SGMSYS_PCS_CONTROL_1, val);
+ regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
+ SGMII_AN_ENABLE, 0);
/* Set the speed etc but leave the duplex unchanged */
- regmap_read(mpcs->regmap, SGMSYS_SGMII_MODE, &val);
- val &= SGMII_DUPLEX_FULL | ~SGMII_IF_MODE_MASK;
- val |= SGMII_SPEED_1000;
- regmap_write(mpcs->regmap, SGMSYS_SGMII_MODE, val);
+ regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
+ SGMII_IF_MODE_MASK & ~SGMII_DUPLEX_FULL,
+ SGMII_SPEED_1000);
/* Release PHYA power down state */
- regmap_read(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, &val);
- val &= ~SGMII_PHYA_PWD;
- regmap_write(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, val);
+ regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
+ SGMII_PHYA_PWD, 0);
}
static int mtk_pcs_config(struct phylink_pcs *pcs, unsigned int mode,
@@ -105,29 +97,28 @@ static int mtk_pcs_config(struct phylink
static void mtk_pcs_restart_an(struct phylink_pcs *pcs)
{
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
- unsigned int val;
- regmap_read(mpcs->regmap, SGMSYS_PCS_CONTROL_1, &val);
- val |= SGMII_AN_RESTART;
- regmap_write(mpcs->regmap, SGMSYS_PCS_CONTROL_1, val);
+ regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
+ SGMII_AN_RESTART, SGMII_AN_RESTART);
}
static void mtk_pcs_link_up(struct phylink_pcs *pcs, unsigned int mode,
phy_interface_t interface, int speed, int duplex)
{
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
- unsigned int val;
+ unsigned int sgm_mode;
if (!phy_interface_mode_is_8023z(interface))
return;
/* SGMII force duplex setting */
- regmap_read(mpcs->regmap, SGMSYS_SGMII_MODE, &val);
- val &= ~SGMII_DUPLEX_FULL;
if (duplex == DUPLEX_FULL)
- val |= SGMII_DUPLEX_FULL;
+ sgm_mode = SGMII_DUPLEX_FULL;
+ else
+ sgm_mode = 0;
- regmap_write(mpcs->regmap, SGMSYS_SGMII_MODE, val);
+ regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
+ SGMII_DUPLEX_FULL, sgm_mode);
}
static const struct phylink_pcs_ops mtk_pcs_ops = {

52
target/linux/generic/backport-6.1/733-v6.2-06-net-mtk_eth_soc-add-out-of-band-forcing-of-speed-and.patch Normal file
View File

@ -0,0 +1,52 @@
From 12198c3a410fe69843e335c1bbf6d4c2a4d48e4e Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:11:03 +0100
Subject: [PATCH 05/10] net: mtk_eth_soc: add out of band forcing of speed and
duplex in pcs_link_up
Add support for forcing the link speed and duplex setting in the
pcs_link_up() method for out of band modes, which will be useful when
we finish converting the pcs_config() method. Until then, we still have
to force duplex for 802.3z modes to work correctly.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 28 ++++++++++++++---------
1 file changed, 17 insertions(+), 11 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -108,17 +108,23 @@ static void mtk_pcs_link_up(struct phyli
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
unsigned int sgm_mode;
- if (!phy_interface_mode_is_8023z(interface))
- return;
+ if (!phylink_autoneg_inband(mode) ||
+ phy_interface_mode_is_8023z(interface)) {
+ /* Force the speed and duplex setting */
+ if (speed == SPEED_10)
+ sgm_mode = SGMII_SPEED_10;
+ else if (speed == SPEED_100)
+ sgm_mode = SGMII_SPEED_100;
+ else
+ sgm_mode = SGMII_SPEED_1000;
- /* SGMII force duplex setting */
- if (duplex == DUPLEX_FULL)
- sgm_mode = SGMII_DUPLEX_FULL;
- else
- sgm_mode = 0;
+ if (duplex == DUPLEX_FULL)
+ sgm_mode |= SGMII_DUPLEX_FULL;
- regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
- SGMII_DUPLEX_FULL, sgm_mode);
+ regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
+ SGMII_DUPLEX_FULL | SGMII_SPEED_MASK,
+ sgm_mode);
+ }
}
static const struct phylink_pcs_ops mtk_pcs_ops = {

48
target/linux/generic/backport-6.1/733-v6.2-07-net-mtk_eth_soc-move-PHY-power-up.patch Normal file
View File

@ -0,0 +1,48 @@
From 6f38fffe2179dd29612aea2c67c46ed6682b4e46 Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:11:08 +0100
Subject: [PATCH 06/10] net: mtk_eth_soc: move PHY power up
The PHY power up is common to both configuration paths, so move it into
the parent function. We need to do this for all serdes modes.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 11 ++++-------
1 file changed, 4 insertions(+), 7 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -45,9 +45,6 @@ static void mtk_pcs_setup_mode_an(struct
regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
SGMII_AN_RESTART, SGMII_AN_RESTART);
-
- regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
- SGMII_PHYA_PWD, 0);
}
/* For 1000BASE-X and 2500BASE-X interface modes, which operate at a
@@ -72,10 +69,6 @@ static void mtk_pcs_setup_mode_force(str
regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
SGMII_IF_MODE_MASK & ~SGMII_DUPLEX_FULL,
SGMII_SPEED_1000);
-
- /* Release PHYA power down state */
- regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
- SGMII_PHYA_PWD, 0);
}
static int mtk_pcs_config(struct phylink_pcs *pcs, unsigned int mode,
@@ -91,6 +84,10 @@ static int mtk_pcs_config(struct phylink
else if (phylink_autoneg_inband(mode))
mtk_pcs_setup_mode_an(mpcs);
+ /* Release PHYA power down state */
+ regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
+ SGMII_PHYA_PWD, 0);
+
return 0;
}

48
target/linux/generic/backport-6.1/733-v6.2-08-net-mtk_eth_soc-move-interface-speed-selection.patch Normal file
View File

@ -0,0 +1,48 @@
From f752c0df13dfeb721c11d3debb79f08cf437344f Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:11:13 +0100
Subject: [PATCH 07/10] net: mtk_eth_soc: move interface speed selection
Move the selection of the underlying interface speed to the pcs_config
function, so we always program the interface speed.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 18 ++++++++++--------
1 file changed, 10 insertions(+), 8 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -53,14 +53,6 @@ static void mtk_pcs_setup_mode_an(struct
static void mtk_pcs_setup_mode_force(struct mtk_pcs *mpcs,
phy_interface_t interface)
{
- unsigned int rgc3;
-
- if (interface == PHY_INTERFACE_MODE_2500BASEX)
- rgc3 = RG_PHY_SPEED_3_125G;
-
- regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
- RG_PHY_SPEED_3_125G, rgc3);
-
/* Disable SGMII AN */
regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
SGMII_AN_ENABLE, 0);
@@ -77,6 +69,16 @@ static int mtk_pcs_config(struct phylink
bool permit_pause_to_mac)
{
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
+ unsigned int rgc3;
+
+ if (interface == PHY_INTERFACE_MODE_2500BASEX)
+ rgc3 = RG_PHY_SPEED_3_125G;
+ else
+ rgc3 = 0;
+
+ /* Configure the underlying interface speed */
+ regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
+ RG_PHY_SPEED_3_125G, rgc3);
/* Setup SGMIISYS with the determined property */
if (interface != PHY_INTERFACE_MODE_SGMII)

52
target/linux/generic/backport-6.1/733-v6.2-09-net-mtk_eth_soc-add-advertisement-programming.patch Normal file
View File

@ -0,0 +1,52 @@
From c125c66ea71b9377ae2478c4f1b87b180cc5c6ef Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:11:18 +0100
Subject: [PATCH 08/10] net: mtk_eth_soc: add advertisement programming
Program the advertisement into the mtk PCS block.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 13 ++++++++++++-
1 file changed, 12 insertions(+), 1 deletion(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -70,16 +70,27 @@ static int mtk_pcs_config(struct phylink
{
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
unsigned int rgc3;
+ int advertise;
+ bool changed;
if (interface == PHY_INTERFACE_MODE_2500BASEX)
rgc3 = RG_PHY_SPEED_3_125G;
else
rgc3 = 0;
+ advertise = phylink_mii_c22_pcs_encode_advertisement(interface,
+ advertising);
+ if (advertise < 0)
+ return advertise;
+
/* Configure the underlying interface speed */
regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
RG_PHY_SPEED_3_125G, rgc3);
+ /* Update the advertisement, noting whether it has changed */
+ regmap_update_bits_check(mpcs->regmap, SGMSYS_PCS_ADVERTISE,
+ SGMII_ADVERTISE, advertise, &changed);
+
/* Setup SGMIISYS with the determined property */
if (interface != PHY_INTERFACE_MODE_SGMII)
mtk_pcs_setup_mode_force(mpcs, interface);
@@ -90,7 +101,7 @@ static int mtk_pcs_config(struct phylink
regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
SGMII_PHYA_PWD, 0);
- return 0;
+ return changed;
}
static void mtk_pcs_restart_an(struct phylink_pcs *pcs)

63
target/linux/generic/backport-6.1/733-v6.2-10-net-mtk_eth_soc-move-and-correct-link-timer-programm.patch Normal file
View File

@ -0,0 +1,63 @@
From 3027d89f87707e7f3e5b683e0d37a32afb5bde96 Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:11:23 +0100
Subject: [PATCH 09/10] net: mtk_eth_soc: move and correct link timer
programming
Program the link timer appropriately for the interface mode being
used, using the newly introduced phylink helper that provides the
nanosecond link timer interval.
The intervals are 1.6ms for SGMII based protocols and 10ms for
802.3z based protocols.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -36,10 +36,6 @@ static void mtk_pcs_get_state(struct phy
/* For SGMII interface mode */
static void mtk_pcs_setup_mode_an(struct mtk_pcs *mpcs)
{
- /* Setup the link timer and QPHY power up inside SGMIISYS */
- regmap_write(mpcs->regmap, SGMSYS_PCS_LINK_TIMER,
- SGMII_LINK_TIMER_DEFAULT);
-
regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
SGMII_REMOTE_FAULT_DIS, SGMII_REMOTE_FAULT_DIS);
@@ -69,8 +65,8 @@ static int mtk_pcs_config(struct phylink
bool permit_pause_to_mac)
{
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
+ int advertise, link_timer;
unsigned int rgc3;
- int advertise;
bool changed;
if (interface == PHY_INTERFACE_MODE_2500BASEX)
@@ -83,6 +79,10 @@ static int mtk_pcs_config(struct phylink
if (advertise < 0)
return advertise;
+ link_timer = phylink_get_link_timer_ns(interface);
+ if (link_timer < 0)
+ return link_timer;
+
/* Configure the underlying interface speed */
regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
RG_PHY_SPEED_3_125G, rgc3);
@@ -91,6 +91,9 @@ static int mtk_pcs_config(struct phylink
regmap_update_bits_check(mpcs->regmap, SGMSYS_PCS_ADVERTISE,
SGMII_ADVERTISE, advertise, &changed);
+ /* Setup the link timer and QPHY power up inside SGMIISYS */
+ regmap_write(mpcs->regmap, SGMSYS_PCS_LINK_TIMER, link_timer / 2 / 8);
+
/* Setup SGMIISYS with the determined property */
if (interface != PHY_INTERFACE_MODE_SGMII)
mtk_pcs_setup_mode_force(mpcs, interface);

132
target/linux/generic/backport-6.1/733-v6.2-11-net-mtk_eth_soc-add-support-for-in-band-802.3z-negot.patch Normal file
View File

@ -0,0 +1,132 @@
From 81b0f12a2a8a1699a7d49c3995e5f71e4ec018e6 Mon Sep 17 00:00:00 2001
From: "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk>
Date: Thu, 27 Oct 2022 14:11:28 +0100
Subject: [PATCH 10/10] net: mtk_eth_soc: add support for in-band 802.3z
negotiation
As a result of help from Frank Wunderlich to investigate and test, we
now know how to program this PCS for in-band 802.3z negotiation. Add
support for this by moving the contents of the two functions into the
common mtk_pcs_config() function and adding the register settings for
802.3z negotiation.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_sgmii.c | 77 ++++++++++++-----------
1 file changed, 42 insertions(+), 35 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -33,41 +33,15 @@ static void mtk_pcs_get_state(struct phy
FIELD_GET(SGMII_LPA, adv));
}
-/* For SGMII interface mode */
-static void mtk_pcs_setup_mode_an(struct mtk_pcs *mpcs)
-{
- regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
- SGMII_REMOTE_FAULT_DIS, SGMII_REMOTE_FAULT_DIS);
-
- regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
- SGMII_AN_RESTART, SGMII_AN_RESTART);
-}
-
-/* For 1000BASE-X and 2500BASE-X interface modes, which operate at a
- * fixed speed.
- */
-static void mtk_pcs_setup_mode_force(struct mtk_pcs *mpcs,
- phy_interface_t interface)
-{
- /* Disable SGMII AN */
- regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
- SGMII_AN_ENABLE, 0);
-
- /* Set the speed etc but leave the duplex unchanged */
- regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
- SGMII_IF_MODE_MASK & ~SGMII_DUPLEX_FULL,
- SGMII_SPEED_1000);
-}
-
static int mtk_pcs_config(struct phylink_pcs *pcs, unsigned int mode,
phy_interface_t interface,
const unsigned long *advertising,
bool permit_pause_to_mac)
{
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
+ unsigned int rgc3, sgm_mode, bmcr;
int advertise, link_timer;
- unsigned int rgc3;
- bool changed;
+ bool changed, use_an;
if (interface == PHY_INTERFACE_MODE_2500BASEX)
rgc3 = RG_PHY_SPEED_3_125G;
@@ -83,6 +57,37 @@ static int mtk_pcs_config(struct phylink
if (link_timer < 0)
return link_timer;
+ /* Clearing IF_MODE_BIT0 switches the PCS to BASE-X mode, and
+ * we assume that fixes it's speed at bitrate = line rate (in
+ * other words, 1000Mbps or 2500Mbps).
+ */
+ if (interface == PHY_INTERFACE_MODE_SGMII) {
+ sgm_mode = SGMII_IF_MODE_SGMII;
+ if (phylink_autoneg_inband(mode)) {
+ sgm_mode |= SGMII_REMOTE_FAULT_DIS |
+ SGMII_SPEED_DUPLEX_AN;
+ use_an = true;
+ } else {
+ use_an = false;
+ }
+ } else if (phylink_autoneg_inband(mode)) {
+ /* 1000base-X or 2500base-X autoneg */
+ sgm_mode = SGMII_REMOTE_FAULT_DIS;
+ use_an = linkmode_test_bit(ETHTOOL_LINK_MODE_Autoneg_BIT,
+ advertising);
+ } else {
+ /* 1000base-X or 2500base-X without autoneg */
+ sgm_mode = 0;
+ use_an = false;
+ }
+
+ if (use_an) {
+ /* FIXME: Do we need to set AN_RESTART here? */
+ bmcr = SGMII_AN_RESTART | SGMII_AN_ENABLE;
+ } else {
+ bmcr = 0;
+ }
+
/* Configure the underlying interface speed */
regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
RG_PHY_SPEED_3_125G, rgc3);
@@ -94,11 +99,14 @@ static int mtk_pcs_config(struct phylink
/* Setup the link timer and QPHY power up inside SGMIISYS */
regmap_write(mpcs->regmap, SGMSYS_PCS_LINK_TIMER, link_timer / 2 / 8);
- /* Setup SGMIISYS with the determined property */
- if (interface != PHY_INTERFACE_MODE_SGMII)
- mtk_pcs_setup_mode_force(mpcs, interface);
- else if (phylink_autoneg_inband(mode))
- mtk_pcs_setup_mode_an(mpcs);
+ /* Update the sgmsys mode register */
+ regmap_update_bits(mpcs->regmap, SGMSYS_SGMII_MODE,
+ SGMII_REMOTE_FAULT_DIS | SGMII_SPEED_DUPLEX_AN |
+ SGMII_IF_MODE_SGMII, sgm_mode);
+
+ /* Update the BMCR */
+ regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
+ SGMII_AN_RESTART | SGMII_AN_ENABLE, bmcr);
/* Release PHYA power down state */
regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
@@ -121,8 +129,7 @@ static void mtk_pcs_link_up(struct phyli
struct mtk_pcs *mpcs = pcs_to_mtk_pcs(pcs);
unsigned int sgm_mode;
- if (!phylink_autoneg_inband(mode) ||
- phy_interface_mode_is_8023z(interface)) {
+ if (!phylink_autoneg_inband(mode)) {
/* Force the speed and duplex setting */
if (speed == SPEED_10)
sgm_mode = SGMII_SPEED_10;

119
target/linux/generic/backport-6.1/733-v6.2-12-net-mediatek-sgmii-ensure-the-SGMII-PHY-is-powered-d.patch Normal file
View File

@ -0,0 +1,119 @@
From 7ff82416de8295c61423ef6fd75f052d3837d2f7 Mon Sep 17 00:00:00 2001
From: Alexander Couzens <lynxis@fe80.eu>
Date: Wed, 1 Feb 2023 19:23:29 +0100
Subject: [PATCH 11/13] net: mediatek: sgmii: ensure the SGMII PHY is powered
down on configuration
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The code expect the PHY to be in power down which is only true after reset.
Allow changes of the SGMII parameters more than once.
Only power down when reconfiguring to avoid bouncing the link when there's
no reason to - based on code from Russell King.
There are cases when the SGMII_PHYA_PWD register contains 0x9 which
prevents SGMII from working. The SGMII still shows link but no traffic
can flow. Writing 0x0 to the PHYA_PWD register fix the issue. 0x0 was
taken from a good working state of the SGMII interface.
Fixes: 42c03844e93d ("net-next: mediatek: add support for MediaTek MT7622 SoC")
Suggested-by: Russell King (Oracle) <linux@armlinux.org.uk>
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
[ bmork: rebased and squashed into one patch ]
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Acked-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/mediatek/mtk_eth_soc.h | 2 ++
drivers/net/ethernet/mediatek/mtk_sgmii.c | 39 +++++++++++++++------
2 files changed, 30 insertions(+), 11 deletions(-)
--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
@@ -1073,11 +1073,13 @@ struct mtk_soc_data {
* @regmap: The register map pointing at the range used to setup
* SGMII modes
* @ana_rgc3: The offset refers to register ANA_RGC3 related to regmap
+ * @interface: Currently configured interface mode
* @pcs: Phylink PCS structure
*/
struct mtk_pcs {
struct regmap *regmap;
u32 ana_rgc3;
+ phy_interface_t interface;
struct phylink_pcs pcs;
};
--- a/drivers/net/ethernet/mediatek/mtk_sgmii.c
+++ b/drivers/net/ethernet/mediatek/mtk_sgmii.c
@@ -43,11 +43,6 @@ static int mtk_pcs_config(struct phylink
int advertise, link_timer;
bool changed, use_an;
- if (interface == PHY_INTERFACE_MODE_2500BASEX)
- rgc3 = RG_PHY_SPEED_3_125G;
- else
- rgc3 = 0;
-
advertise = phylink_mii_c22_pcs_encode_advertisement(interface,
advertising);
if (advertise < 0)
@@ -88,9 +83,22 @@ static int mtk_pcs_config(struct phylink
bmcr = 0;
}
- /* Configure the underlying interface speed */
- regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
- RG_PHY_SPEED_3_125G, rgc3);
+ if (mpcs->interface != interface) {
+ /* PHYA power down */
+ regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
+ SGMII_PHYA_PWD, SGMII_PHYA_PWD);
+
+ if (interface == PHY_INTERFACE_MODE_2500BASEX)
+ rgc3 = RG_PHY_SPEED_3_125G;
+ else
+ rgc3 = 0;
+
+ /* Configure the underlying interface speed */
+ regmap_update_bits(mpcs->regmap, mpcs->ana_rgc3,
+ RG_PHY_SPEED_3_125G, rgc3);
+
+ mpcs->interface = interface;
+ }
/* Update the advertisement, noting whether it has changed */
regmap_update_bits_check(mpcs->regmap, SGMSYS_PCS_ADVERTISE,
@@ -108,9 +116,17 @@ static int mtk_pcs_config(struct phylink
regmap_update_bits(mpcs->regmap, SGMSYS_PCS_CONTROL_1,
SGMII_AN_RESTART | SGMII_AN_ENABLE, bmcr);
- /* Release PHYA power down state */
- regmap_update_bits(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL,
- SGMII_PHYA_PWD, 0);
+ /* Release PHYA power down state
+ * Only removing bit SGMII_PHYA_PWD isn't enough.
+ * There are cases when the SGMII_PHYA_PWD register contains 0x9 which
+ * prevents SGMII from working. The SGMII still shows link but no traffic
+ * can flow. Writing 0x0 to the PHYA_PWD register fix the issue. 0x0 was
+ * taken from a good working state of the SGMII interface.
+ * Unknown how much the QPHY needs but it is racy without a sleep.
+ * Tested on mt7622 & mt7986.
+ */
+ usleep_range(50, 100);
+ regmap_write(mpcs->regmap, SGMSYS_QPHY_PWR_STATE_CTRL, 0);
return changed;
}
@@ -171,6 +187,7 @@ int mtk_sgmii_init(struct mtk_sgmii *ss,
return PTR_ERR(ss->pcs[i].regmap);
ss->pcs[i].pcs.ops = &mtk_pcs_ops;
+ ss->pcs[i].interface = PHY_INTERFACE_MODE_NA;
}
return 0;

Some files were not shown because too many files have changed in this diff Show More