If the vop2_plane_atomic_disable function is called with NULL as a
state, accessing the old_pstate runs into a null pointer exception.
However, the drm_atomic_helper_disable_planes_on_crtc function calls the
atomic_disable callback with state NULL.
Allow to disable a plane without passing a plane state by checking the
old_pstate only if a state is passed.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221028095206.2136601-2-m.tretter@pengutronix.de
lcdsel_grf_reg is defined as u32, so "< 0" comaprison is always false,
which breaks VOP selection on eg. RK3399. Compare against 0.
Fixes: f3aaa6125b ("drm/rockchip: dsi: add rk3568 support")
Signed-off-by: Ondrej Jirman <megi@xff.cz>
Tested-by: Chris Morgan <macromorgan@hotmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221023160747.607943-1-megi@xff.cz
When switching to the generic fbdev infrastructure, it was missed that
framebuffers were created with the alloc_kmap parameter to
rockchip_gem_create_object() set to true. The generic infrastructure
calls this via the .dumb_create() driver operation and thus creates a
buffer without an associated kmap.
alloc_kmap only makes a difference on devices without an IOMMU, but when
it is missing rockchip_gem_prime_vmap() fails and the framebuffer cannot
be used.
Detect the case where a buffer is being allocated for the framebuffer
and ensure a kernel mapping is created in this case.
Fixes: 24af7c34b2 ("drm/rockchip: use generic fbdev setup")
Reported-by: Johan Jonker <jbx6244@gmail.com>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020181248.2497065-1-john@metanate.com
We can't safely probe a dual-DSI display asynchronously
(driver_async_probe='*' or driver_async_probe='dw-mipi-dsi-rockchip'
cmdline), because dw_mipi_dsi_rockchip_find_second() pokes one DSI
device's drvdata from the other device without any locking.
Request synchronous probe, at least until this driver learns some
appropriate locking for dual-DSI initialization.
Cc: <stable@vger.kernel.org>
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221019170255.2.I6b985b0ca372b7e35c6d9ea970b24bcb262d4fc1@changeid
If we fail to attach the first time (especially: EPROBE_DEFER), we fail
to clean up 'usage_mode', and thus will fail to attach on any subsequent
attempts, with "dsi controller already in use".
Re-set to DW_DSI_USAGE_IDLE on attach failure.
This is especially common to hit when enabling asynchronous probe on a
duel-DSI system (such as RK3399 Gru/Scarlet), such that we're more
likely to fail dw_mipi_dsi_rockchip_find_second() the first time.
Fixes: 71f68fe7f1 ("drm/rockchip: dsi: add ability to work as a phy instead of full dsi")
Cc: <stable@vger.kernel.org>
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221019170255.1.Ia68dfb27b835d31d22bfe23812baf366ee1c6eac@changeid
When the avdd-0v9 or avdd-1v8 supply are not yet available, EPROBE_DEFER
is returned by rockchip_hdmi_parse_dt(). This causes the following error
message to be printed multiple times:
dwhdmi-rockchip fe0a0000.hdmi: [drm:dw_hdmi_rockchip_bind [rockchipdrm]] *ERROR* Unable to parse OF data
Fix that by not printing the message when rockchip_hdmi_parse_dt()
returns -EPROBE_DEFER.
Fixes: ca80c4eb4b ("drm/rockchip: dw_hdmi: add regulator support")
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reviewed-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220926203752.5430-1-aurelien@aurel32.net
Up until now, the external MDIO controller frequency values relied
either on the default ones out of reset or on those setup by u-boot.
Let's just properly specify the MDC frequency in the DTS so that even
without u-boot's intervention Linux can drive the MDIO bus.
Fixes: 0420dde30a ("arm64: dts: ls208xa: add the external MDIO nodes")
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Up until now, the external MDIO controller frequency values relied
either on the default ones out of reset or on those setup by u-boot.
Let's just properly specify the MDC frequency in the DTS so that even
without u-boot's intervention Linux can drive the MDIO bus.
Fixes: bbe75af7b0 ("arm64: dts: ls1088a: add external MDIO device nodes")
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Up until now, the external MDIO controller frequency values relied
either on the default ones out of reset or on those setup by u-boot.
Let's just properly specify the MDC frequency in the DTS so that even
without u-boot's intervention Linux can drive the MDIO bus.
Fixes: 6e1b8fae89 ("arm64: dts: lx2160a: add emdio1 node")
Fixes: 5705b9dcda ("arm64: dts: lx2160a: add emdio2 node")
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Input voltage channel enum is compared against iio address instead
of the channel.
Fixes: 7abd9fb646 ("iio: adc: mp2629: Add support for mp2629 ADC driver")
Signed-off-by: Saravanan Sekar <sravanhome@gmail.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Link: https://lore.kernel.org/r/20221029093000.45451-2-sravanhome@gmail.com
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Don't hardcode the ms5611 SPI speed, limit it instead.
Signed-off-by: Mitja Spes <mitja@lxnav.com>
Fixes: c0644160a8 ("iio: pressure: add support for MS5611 pressure and temperature sensor")
Link: https://lore.kernel.org/r/20221021135827.1444793-3-mitja@lxnav.com
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
When using multiple instances of this driver the compensation PROM was
overwritten by the last initialized sensor. Now each sensor has own PROM
storage.
Signed-off-by: Mitja Spes <mitja@lxnav.com>
Fixes: 9690d81a02 ("iio: pressure: ms5611: add support for MS5607 temperature and pressure sensor")
Link: https://lore.kernel.org/r/20221021135827.1444793-2-mitja@lxnav.com
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The regulator enables were after the check on the chip variant, which was
very unlikely to return a correct value when not powered.
Presumably all the device anyone is testing on have a regulator that
is already powered up when this code runs for reasons beyond the scope
of this driver. Move the read call down a few lines.
Fixes: 3cf7ded15e ("iio: accel: bma400: basic regulator support")
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dan Robertson <dan@dlrobertson.com>
Cc: <Stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20221002144133.3771029-1-jic23@kernel.org
platform_get_irq() returns negative error number on failure, fix the
return value check in pxa2xx_ac97_hw_probe() and assign the error code
to 'ret'.
Fixes: 2548e6c76e ("ARM: pxa: pxa2xx-ac97-lib: use IRQ resource")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221029082001.3207380-1-yangyingliang@huawei.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently the return value of 'sub_driver->init' is not checked. If
sparse_keymap_setup() called in the init function fails, 'generic_
inputdev' is freed, then it will lead a UAF when using it in generic_
acpi_laptop_init(). Fix it by checking the return value and setting
generic_inputdev to NULL after free, so as to avoid double free it.
The error code in generic_subdriver_init() is always negative, so the
return of generic_subdriver_init() can be simplified.
Fixes: 6246ed0911 ("LoongArch: Add ACPI-based generic laptop driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Some laptops don't support SW_LID, but still have backlight control,
move backlight resuming before SW_LID event handling so as to avoid
backlight mistake due to early return.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Not all compilers support declare variables in switch-case, so move
declarations to the beginning of a function. Otherwise we may get such
build errors:
arch/loongarch/net/bpf_jit.c: In function ‘emit_atomic’:
arch/loongarch/net/bpf_jit.c:362:3: error: a label can only be part of a statement and a declaration is not a statement
u8 r0 = regmap[BPF_REG_0];
^~
arch/loongarch/net/bpf_jit.c: In function ‘build_insn’:
arch/loongarch/net/bpf_jit.c:727:3: error: a label can only be part of a statement and a declaration is not a statement
u8 t7 = -1;
^~
arch/loongarch/net/bpf_jit.c:778:3: error: a label can only be part of a statement and a declaration is not a statement
int ret;
^~~
arch/loongarch/net/bpf_jit.c:779:3: error: expected expression before ‘u64’
u64 func_addr;
^~~
arch/loongarch/net/bpf_jit.c:780:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement]
bool func_addr_fixed;
^~~~
arch/loongarch/net/bpf_jit.c:784:11: error: ‘func_addr’ undeclared (first use in this function); did you mean ‘in_addr’?
&func_addr, &func_addr_fixed);
^~~~~~~~~
in_addr
arch/loongarch/net/bpf_jit.c:784:11: note: each undeclared identifier is reported only once for each function it appears in
arch/loongarch/net/bpf_jit.c:814:3: error: a label can only be part of a statement and a declaration is not a statement
u64 imm64 = (u64)(insn + 1)->imm << 32 | (u32)insn->imm;
^~~
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Eliminate the following coccicheck warning:
./arch/loongarch/include/asm/ptrace.h:32:15-21: WARNING use flexible-array member instead
Reviewed-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Yushan Zhou <katrinzhou@tencent.com>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
The current LoongArch kernel stack is padded as if obeying the MIPS o32
calling convention (32 bytes), signifying the port's MIPS lineage but no
longer making sense. Remove the padding for clarity.
Reviewed-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Jinyang He <hejinyang@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
In imx93_pd_probe(); if an error occurs, some resources need to be released
as done in the remove function.
Fixes: 0a0f7cc25d ("soc: imx: add i.MX93 SRC power domain driver")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Per binding doc, interrupt names should be tx and rx.
Fixes: 0dfb380d24 ("arm64: dts: imx93: add s4 mu node")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Some pgc power-domain requires a parent power domain, so
add an optional power-domains property, otherwise there will be
dt check warning:
gpc@303a0000: pgc:power-domain@1: 'power-domains' does not match
any of the regexes: 'pinctrl-[0-9]+'
Fixes: 30af8513bd ("dt-bindings: power: add defines for i.MX8MM power domains")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Per bindings/mmc/fsl-imx-esdhc.yaml, the clock order is ipg, ahb, per,
otherwise warning: "
mmc@5b020000: clock-names:1: 'ahb' was expected
mmc@5b020000: clock-names:2: 'per' was expected "
Fixes: 16c4ea7501 ("arm64: dts: imx8: switch to new lpcg clock binding")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Fix our design flaw in supply voltage distribution on the Quad and QuadPlus
based boards.
The problem is that we supply the SoC cache (VDD_CACHE_CAP) from VDD_PU
instead of VDD_SOC. The VDD_PU internal regulator can be disabled by PM
if VPU or GPU is not used. If that happens the system freezes. To prevent
that configure the reg_pu regulator to be always on.
Fixes: 0de4ab81ab ("ARM: dts: imx6dl-yapp4: Add Y Soft IOTA Crux/Crux+ board")
Cc: petrben@gmail.com
Cc: stable@vger.kernel.org
Signed-off-by: Petr Benes <petr.benes@ysoft.com>
Signed-off-by: Michal Vokáč <michal.vokac@ysoft.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
The GW5910 and GW5913 have a user pushbutton that is tied to the
Gateworks System Controller GPIO offset 2. Fix the invalid offset of 0.
Fixes: 64bf0a0af1 ("ARM: dts: imx6qdl-gw: add Gateworks System Controller support")
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
pgc_otg1 is actual the power domain of usb PHY, usb controller
is in hsio power domain, and pgc_otg1 is required to be powered
up to detect usb remote wakeup, so move the pgc_otg1 power domain
to the usb phy node.
Fixes: ea2b5af58a ("arm64: dts: imx8mn: put USB controller into power-domains")
Signed-off-by: Li Jun <jun.li@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
pgc_otg1 is an independent power domain of hsio, it's for usb phy,
so remove hsio power domain from its node.
Fixes: 8b8ebec673 ("arm64: dts: imx8mn: add GPC node")
Signed-off-by: Li Jun <jun.li@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
pgc_otg1/2 is actual the power domain of usb PHY, usb controller
is in hsio power domain, and pgc_otg1/2 is required to be powered
up to detect usb remote wakeup, so move the pgc_otg1/2 power domain
to the usb phy node.
Fixes: 01df28d808 ("arm64: dts: imx8mm: put USB controllers into power-domains")
Signed-off-by: Li Jun <jun.li@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
pgc_otg1/2 are independent power domain of hsio, they for usb phy,
so remove hsio power domain dependency from its node.
Fixes: d39d4bb153 ("arm64: dts: imx8mm: add GPC node")
Signed-off-by: Li Jun <jun.li@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
The GPIO signaling ctrl_sleep_moci is currently handled as a gpio hog.
But the gpio-hog node is made a child of the wrong gpio controller.
Move it to the node representing gpio4 so that it actually works.
Without this carrier board components jumpered to use the signal are
unconditionally switched off.
Fixes: a39ed23bdf ("arm64: dts: freescale: add initial support for verdin imx8m plus")
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
To keep the power domain on if usb remote wakeup is enabled, add the
GENPD_FLAG_ACTIVE_WAKEUP for otg1/2 of imx8mm/n.
Signed-off-by: Li Jun <jun.li@nxp.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
The rk3399 has a crypto IP handled by the rk3288 crypto driver so adds a
node for it.
Tested-by Diederik de Haas <didi.debian@cknow.org>
Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Link: https://lore.kernel.org/r/20220927075511.3147847-29-clabbe@baylibre.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
i.MX8MP ECSPI is derived from i.MX6UL, so update compatible
Add assigned-clocks settings
Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
i.MX8MP ECSPI is derived from i.MX6UL, so update the binding.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
gpio driver need these gpio alises to config the gpio base.
Reviewed-by: Jun Li <jun.li@nxp.com>
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Add MU clock for mailbox usage
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
There is no internal '/3' divider in i.MX93 sysctr, so need this
property.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
SW steering updates from Yevgeny Kliteynik:
1) 1st Four patches: small fixes / optimizations for SW steering:
- Patch 1: Don't abort destroy flow if failed to destroy table - continue
and free everything else.
- Patches 2 and 3 deal with fast teardown:
+ Skip sync during fast teardown, as PCI device is not there any more.
+ Check device state when polling CQ - otherwise SW steering keeps polling
the CQ forever, because nobody is there to flush it.
- Patch 4: Removing unneeded function argument.
2) Deal with the hiccups that we get during rules insertion/deletion,
which sometimes reach 1/4 of a second. While insertion/deletion rate
improvement was not the focus here, it still is a by-product of removing these
hiccups.
Another by-product is the reduced standard deviation in measuring the duration
of rules insertion/deletion bursts.
In the testing we add K rules (warm-up phase), and then continuously do
insertion/deletion bursts of N rules.
During the test execution, the driver measures hiccups (amount and duration)
and total time for insertion/deletion of a batch of rules.
Here are some numbers, before and after these patches:
+--------------------------------------------+-----------------+----------------+
| | Create rules | Delete rules |
| +--------+--------+--------+-------+
| | Before | After | Before | After |
+--------------------------------------------+--------+--------+--------+-------+
| Max hiccup [msec] | 253 | 42 | 254 | 68 |
+--------------------------------------------+--------+--------+--------+-------+
| Avg duration of 10K rules add/remove [msec]| 140.07 | 124.32 | 106.99 | 99.51 |
+--------------------------------------------+--------+--------+--------+-------+
| Num of hiccups per 100K rules add/remove | 7.77 | 7.97 | 12.60 | 11.57 |
+--------------------------------------------+--------+--------+--------+-------+
| Avg hiccup duration [msec] | 36.92 | 33.25 | 36.15 | 33.74 |
+--------------------------------------------+--------+--------+--------+-------+
- Patch 5: Allocate a short array on stack instead of dynamically- it is
destroyed at the end of the function.
- Patch 6: Rather than cleaning the corresponding chunk's section of
ste_arrays on chunk deletion, initialize these areas upon chunk creation.
Chunk destruction tend to come in large batches (during pool syncing),
so instead of doing huge memory initialization during pool sync,
we amortize this by doing small initsializations on chunk creation.
- Patch 7: In order to simplifies error flow and allows cleaner addition
of new pools, handle creation/destruction of all the domain's memory pools
and other memory-related fields in a separate init/uninit functions.
- Patch 8: During rehash, write each table row immediately instead of waiting
for the whole table to be ready and writing it all - saves allocations
of ste_send_info structures and improves performance.
- Patch 9: Instead of allocating/freeing send info objects dynamically,
manage them in pool. The number of send info objects doesn't depend on
number of rules, so after pre-populating the pool with an initial batch of
send info objects, the pool is not expected to grow.
This way we save alloc/free during writing STEs to ICM, which by itself can
sometimes take up to 40msec.
- Patch 10: Allocate icm_chunks from their own slab allocator, which lowered
the alloc/free "hiccups" frequency.
- Patch 11: Similar to patch 9, allocate htbl from its own slab allocator.
- Patch 12: Lower sync threshold for ICM hot memory - set the threshold for
sync to 1/4 of the pool instead of 1/2 of the pool. Although we will have
more syncs, each sync will be shorter and will help with insertion rate
stability. Also, notice that the overall number of hiccups wasn't increased
due to all the other patches.
- Patch 13: Keep track of hot ICM chunks in an array instead of list.
After steering sync, we traverse the hot list and finally free all the
chunks. It appears that traversing a long list takes unusually long time
due to cache misses on many entries, which causes a big "hiccup" during
rule insertion. This patch replaces the list with pre-allocated array that
stores only the bookkeeping information that is needed to later free the
chunks in its buddy allocator.
- Patch 14: Remove the unneeded buddy used_list - we don't need to have the
list of used chunks, we only need the total amount of used memory.
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEGhZs6bAKwk/OTgTpSD+KveBX+j4FAmNamsEACgkQSD+KveBX
+j6hWgf/ec6O0Ige4Az9quVtN1YiLRpcA4RJs5prV/2qyqcjzpkumTLpWgFzD6SM
T7uz/lQY4/JTLAkFNQBE6aynjtFfUP7bJ2LulqE+8QXBmHoHndHA+S3ZBGAmjLgK
9tY73Bb5qxsHCzEvaab+UxEIWXiBPtaNaw5mkzKKO5ULCplVl1loKxVEmLO1ri7j
fa7G7I1VHgSg6/7GWPzMN9tsR8b927H9gdRw3atTC91T8jgwf+9YYXmhd4Bj2Dk0
uB1n4AyVCcLxGZiFtNHUSyBNIvFwaO87DzBDDftIJPSMcJvRJxeBx0692Z7sZdE0
cfta+4bPDpjHVNN2slcYgzJu/jDh8w==
=5XcL
-----END PGP SIGNATURE-----
Merge tag 'mlx5-updates-2022-10-24' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-updates-2022-10-24
SW steering updates from Yevgeny Kliteynik:
1) 1st Four patches: small fixes / optimizations for SW steering:
- Patch 1: Don't abort destroy flow if failed to destroy table - continue
and free everything else.
- Patches 2 and 3 deal with fast teardown:
+ Skip sync during fast teardown, as PCI device is not there any more.
+ Check device state when polling CQ - otherwise SW steering keeps polling
the CQ forever, because nobody is there to flush it.
- Patch 4: Removing unneeded function argument.
2) Deal with the hiccups that we get during rules insertion/deletion,
which sometimes reach 1/4 of a second. While insertion/deletion rate
improvement was not the focus here, it still is a by-product of removing these
hiccups.
Another by-product is the reduced standard deviation in measuring the duration
of rules insertion/deletion bursts.
In the testing we add K rules (warm-up phase), and then continuously do
insertion/deletion bursts of N rules.
During the test execution, the driver measures hiccups (amount and duration)
and total time for insertion/deletion of a batch of rules.
Here are some numbers, before and after these patches:
+--------------------------------------------+-----------------+----------------+
| | Create rules | Delete rules |
| +--------+--------+--------+-------+
| | Before | After | Before | After |
+--------------------------------------------+--------+--------+--------+-------+
| Max hiccup [msec] | 253 | 42 | 254 | 68 |
+--------------------------------------------+--------+--------+--------+-------+
| Avg duration of 10K rules add/remove [msec]| 140.07 | 124.32 | 106.99 | 99.51 |
+--------------------------------------------+--------+--------+--------+-------+
| Num of hiccups per 100K rules add/remove | 7.77 | 7.97 | 12.60 | 11.57 |
+--------------------------------------------+--------+--------+--------+-------+
| Avg hiccup duration [msec] | 36.92 | 33.25 | 36.15 | 33.74 |
+--------------------------------------------+--------+--------+--------+-------+
- Patch 5: Allocate a short array on stack instead of dynamically- it is
destroyed at the end of the function.
- Patch 6: Rather than cleaning the corresponding chunk's section of
ste_arrays on chunk deletion, initialize these areas upon chunk creation.
Chunk destruction tend to come in large batches (during pool syncing),
so instead of doing huge memory initialization during pool sync,
we amortize this by doing small initsializations on chunk creation.
- Patch 7: In order to simplifies error flow and allows cleaner addition
of new pools, handle creation/destruction of all the domain's memory pools
and other memory-related fields in a separate init/uninit functions.
- Patch 8: During rehash, write each table row immediately instead of waiting
for the whole table to be ready and writing it all - saves allocations
of ste_send_info structures and improves performance.
- Patch 9: Instead of allocating/freeing send info objects dynamically,
manage them in pool. The number of send info objects doesn't depend on
number of rules, so after pre-populating the pool with an initial batch of
send info objects, the pool is not expected to grow.
This way we save alloc/free during writing STEs to ICM, which by itself can
sometimes take up to 40msec.
- Patch 10: Allocate icm_chunks from their own slab allocator, which lowered
the alloc/free "hiccups" frequency.
- Patch 11: Similar to patch 9, allocate htbl from its own slab allocator.
- Patch 12: Lower sync threshold for ICM hot memory - set the threshold for
sync to 1/4 of the pool instead of 1/2 of the pool. Although we will have
more syncs, each sync will be shorter and will help with insertion rate
stability. Also, notice that the overall number of hiccups wasn't increased
due to all the other patches.
- Patch 13: Keep track of hot ICM chunks in an array instead of list.
After steering sync, we traverse the hot list and finally free all the
chunks. It appears that traversing a long list takes unusually long time
due to cache misses on many entries, which causes a big "hiccup" during
rule insertion. This patch replaces the list with pre-allocated array that
stores only the bookkeeping information that is needed to later free the
chunks in its buddy allocator.
- Patch 14: Remove the unneeded buddy used_list - we don't need to have the
list of used chunks, we only need the total amount of used memory.
* tag 'mlx5-updates-2022-10-24' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
net/mlx5: DR, Remove the buddy used_list
net/mlx5: DR, Keep track of hot ICM chunks in an array instead of list
net/mlx5: DR, Lower sync threshold for ICM hot memory
net/mlx5: DR, Allocate htbl from its own slab allocator
net/mlx5: DR, Allocate icm_chunks from their own slab allocator
net/mlx5: DR, Manage STE send info objects in pool
net/mlx5: DR, In rehash write the line in the entry immediately
net/mlx5: DR, Handle domain memory resources init/uninit separately
net/mlx5: DR, Initialize chunk's ste_arrays at chunk creation
net/mlx5: DR, For short chains of STEs, avoid allocating ste_arr dynamically
net/mlx5: DR, Remove unneeded argument from dr_icm_chunk_destroy
net/mlx5: DR, Check device state when polling CQ
net/mlx5: DR, Fix the SMFS sync_steering for fast teardown
net/mlx5: DR, In destroy flow, free resources even if FW command failed
====================
Link: https://lore.kernel.org/r/20221027145643.6618-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Alex Elder says:
====================
net: ipa: start adding IPA v5.0 functionality
The biggest change for IPA v5.0 is that it supports more than 32
endpoints. However there are two other unrelated changes:
- The STATS_TETHERING memory region is not required
- Filter tables no longer support a "global" filter
Beyond this, refactoring some code makes supporting more than 32
endpoints (in an upcoming series) easier. So this series includes
a few other changes (not in this order):
- The maximum endpoint ID in use is determined during config
- Loops over all endpoints only involve those in use
- Endpoints IDs and their directions are checked for validity
differently to simplify comparison against the maximum
====================
Link: https://lore.kernel.org/r/20221027122632.488694-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Define a new field in the IPA structure that records the maximum
number of entries that will be used in the IPA endpoint array. Use
that value rather than IPA_ENDPOINT_MAX to determine the end
condition for two loops that iterate over all endpoints.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Each endpoint ID has an entry in the IPA endpoint array. But the
size of that array is defined at compile time. Instead, rename
ipa_endpoint_data_valid() to be ipa_endpoint_max() and have it
return the maximum endpoint ID defined in configuration data.
That function will still validate configuration data.
Zero is returned on error; it's a valid endpoint ID, but we need
more than one, so it can't be the maximum. The next patch makes use
of the returned maximum value.
Finally, rename the "initialized" mask of endpoints defined by
configuration data to be "defined".
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Change two functions that iterate over all endpoints to use while
loops, using "endpoint_id" as the index variables in both spots.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Ensure all defined TX endpoints are in the range [0, CONS_PIPES) and
defined RX endpoints are within [PROD_LOWEST, PROD_LOWEST+PROD_PIPES).
Modify the way local variables are used to make the checks easier
to understand. Check for each endpoint being in valid range in the
loop, and drop the logical-AND check of initialized against
unavailable IDs.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
IPA v5.0 eliminates the global filter table entry. As a result,
there is no need to shift the filtered endpoint bitmap when it is
written to IPA local memory.
Update comments to explain this. Also delete a redundant block of
comments above the function.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
