As the potential failure of the clk_enable(),
it should be better to check it, like mxs_saif_trigger().
Fixes: d0ba4c014934 ("ASoC: mxs-saif: set a base clock rate for EXTMASTER mode work") Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn> Link: https://lore.kernel.org/r/20220301081717.3727190-1-jiasheng@iscas.ac.cn Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If an invalid option value is used with "printk.devkmsg=<value>",
it is silently ignored.
If a valid option value is used, it is honored but the wrong return
value (0) is used, indicating that the command line option had an
error and was not handled. This string is not added to init's
environment strings due to init/main.c::unknown_bootoption()
checking for a '.' in the boot option string and then considering
that string to be an "Unused module parameter".
Print a warning message if a bad option string is used.
Always return 1 from the __setup handler to indicate that the command
line option has been handled.
Fixes: 750afe7babd1 ("printk: add kernel parameter to control writes to /dev/kmsg") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Reported-by: Igor Zhbanov <i.zhbanov@omprussia.ru>
Link: lore.kernel.org/r/64644a2f-4a20-bab3-1e15-3b2cdd0defe3@omprussia.ru Cc: Borislav Petkov <bp@suse.de> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Petr Mladek <pmladek@suse.com> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: John Ogness <john.ogness@linutronix.de> Reviewed-by: John Ogness <john.ogness@linutronix.de> Reviewed-by: Sergey Senozhatsky <senozhatsky@chromium.org> Reviewed-by: Petr Mladek <pmladek@suse.com> Signed-off-by: Petr Mladek <pmladek@suse.com> Link: https://lore.kernel.org/r/20220228220556.23484-1-rdunlap@infradead.org Signed-off-by: Sasha Levin <sashal@kernel.org>
Broadcom ns2 platform has spi-cpol and spi-cpho properties set
incorrectly. As per spi-slave-peripheral-prop.yaml, these properties are
of flag or boolean type and not integer type. Fix the values.
Fixes: d69dbd9f41a7c (arm64: dts: Add ARM PL022 SPI DT nodes for NS2) Signed-off-by: Kuldeep Singh <singh.kuldeep87k@gmail.com> CC: Ray Jui <rjui@broadcom.com> CC: Scott Branden <sbranden@broadcom.com> CC: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
As the potential failure of the clk_enable(),
it should be better to check it and return error
if fails.
Fixes: 5f9a50c3e55e ("ASoC: Davinci: McBSP: add device tree support for McBSP") Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn> Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com> Link: https://lore.kernel.org/r/20220228031540.3571959-1-jiasheng@iscas.ac.cn Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The MMCSD IPs used in AM64 are the same as the ones used in J721E.
Therefore, fix this by using the driver data from J721E for AM64 too, for
both 8 and 4 bit instances.
The function devm_kzalloc() in rt5663_parse_dp() can fail, so its return
value should be checked.
Fixes: 457c25efc592 ("ASoC: rt5663: Add the function of impedance sensing") Reported-by: TOTE Robot <oslab@tsinghua.edu.cn> Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com> Link: https://lore.kernel.org/r/20220225131030.27248-1-baijiaju1990@gmail.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Although kptr_restrict is set to 0 and the kernel is booted with
no_hash_pointers parameter, the content of /proc/vmallocinfo is
lacking the real addresses.
UBSAN_BOUNDS and UBSAN_TRAP depend on UBSAN config option.
merge_config.sh script generates following warnings if parent config
doesn't have UBSAN config already enabled and UBSAN_BOUNDS/UBSAN_TRAP
config options don't get added to the parent config.
Value requested for CONFIG_UBSAN_BOUNDS not in final .config
Requested value: CONFIG_UBSAN_BOUNDS=y
Actual value:
Value requested for CONFIG_UBSAN_TRAP not in final .config
Requested value: CONFIG_UBSAN_TRAP=y
Actual value:
In ov5640_set_fmt, pending_fmt_change will always be false, because the
sensor format is saved before comparing it with the previous format:
fmt = &sensor->fmt;...
*fmt = *mbus_fmt;...
if (mbus_fmt->code != sensor->fmt.code)
sensor->pending_fmt_change = true;
This causes the sensor to capture with the previous pixelcode.
Also, changes might happen even for V4L2_SUBDEV_FORMAT_TRY, so fix that.
Basically, revert back to the state before
commit 071154499193 ("media: ov5640: Fix set format regression")
as it was more clear, and then update format even when pixelcode does
not change, as resolution might change.
Fixes: 071154499193 ("media: ov5640: Fix set format regression") Fixes: 6949d864776e ("media: ov5640: do not change mode if format or frame interval is unchanged") Fixes: fb98e29ff1ea5 ("media: ov5640: fix mode change regression") Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com> Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org> Acked-by: Hugues Fruchet <hugues.fruchet@st.com> Tested-by: Hugues Fruchet <hugues.fruchet@st.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the final H264 API, it is not required to set scaling matrix if
they are not present in the bitstream. A flag was added in order to let
the driver know. The downside is that it leaves the default control
value to 0, which isn't valid. As per the spec (see formulas 7-8/7-9),
when the scaling matrix are absent from the bitstream, flat values
of 16 should be used. This improves this control semantic in a way
that the control value are always valid. Drivers can then use
the scaling_matrix control values without having to check its presence.
Same method was employed for MPEG2_QUANTISATION.
This fixes issues with MTK VCODEC H264 decoder when using GStreamer.
GStreamer does not set this control if its not present in the bitstream.
As MTK VDCODEC was using the initialized to 0 values, the frames ended
up completely gray.
Fixes: 54889c51b833d236 ("media: uapi: h264: Rename and clarify PPS_FLAG_SCALING_MATRIX_PRESENT") Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> Tested-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Sasha Levin <sashal@kernel.org>
set "HPH Type" Kcontrol max value of WCD_MBHC_HPH_STEREO instead of UINT_MAX.
set "HPHL/R Impedance" Kcontrols max value to INT_MAX instead of UINT_MAX as
max field is integer type.
Without this patch amixer for these controls will show -1 as max value to userspace.
set "HPH Type" Kcontrol max value of WCD_MBHC_HPH_STEREO instead of UINT_MAX.
set "HPHL/R Impedance" Kcontrols max value to INT_MAX instead of UINT_MAX as
max field is integer type.
Without this patch amixer for these controls will show -1 as max value to userspace.
Fixes: bcee7ed09b8e ("ASoC: codecs: wcd938x: add Multi Button Headset Control support") Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> Link: https://lore.kernel.org/r/20220222183212.11580-8-srinivas.kandagatla@linaro.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Accessing enums using integer would result in array out of bounds access
on platforms like aarch64 where sizeof(long) is 8 compared to enum size
which is 4 bytes.
Fix this by using enumerated items instead of integers.
Accessing enums using integer would result in array out of bounds access
on platforms like aarch64 where sizeof(long) is 8 compared to enum size
which is 4 bytes.
Fixes: 908e6b1df26e ("ASoC: codecs: lpass-va-macro: Add support to VA Macro") Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> Link: https://lore.kernel.org/r/20220222183212.11580-5-srinivas.kandagatla@linaro.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Accessing enums using integer would result in array out of bounds access
on platforms like aarch64 where sizeof(long) is 8 compared to enum size
which is 4 bytes.
Fixes: 4f692926f562 ("ASoC: codecs: lpass-rx-macro: add dapm widgets and route") Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> Link: https://lore.kernel.org/r/20220222183212.11580-3-srinivas.kandagatla@linaro.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
According to BSP library source, H264 neighbour info buffer size needs
to be 32 kiB for H6. This is similar to H265 decoding, which also needs
double buffer size in comparison to older Cedrus core generations.
Increase buffer size to cover H6 needs. Since increase is not that big
in absolute numbers, it doesn't make sense to complicate logic for older
generations.
Issue was discovered using iommu and cross checked with BSP library
source.
Neighbour info buffer size needs to be 794 kiB in H6. This is actually
already indirectly mentioned in the comment, but smaller size is used
nevertheless.
Increase buffer size to cover H6 needs. Since increase is not that big
in absolute numbers, it doesn't make sense to complicate logic for older
generations.
Bug was discovered using iommu, which reported access error when trying
to play H265 video.
Call i2c_unregister_device(audio) on this error path.
Fixes: d3b2ccd9e307 ("[media] s2250: convert to the control framework") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Sasha Levin <sashal@kernel.org>
The commit 47677e51e2a4("[media] em28xx: Only deallocate struct
em28xx after finishing all extensions") adds kref_get to many init
functions (e.g., em28xx_audio_init). However, kref_init is called too
late in em28xx_usb_probe, since em28xx_init_dev before will invoke
those init functions and call kref_get function. Then refcount bug
occurs in my local syzkaller instance.
Fix it by moving kref_init before em28xx_init_dev. This issue occurs
not only in dev but also dev->dev_next.
Fixes: 47677e51e2a4 ("[media] em28xx: Only deallocate struct em28xx after finishing all extensions") Reported-by: syzkaller <syzkaller@googlegroups.com> Signed-off-by: Dongliang Mu <mudongliangabcd@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Sasha Levin <sashal@kernel.org>
In cal_ctx_v4l2_init_formats(), devm_kzalloc() is assigned to
ctx->active_fmt and there is a dereference of it after that, which could
lead to NULL pointer dereference on failure of devm_kzalloc().
Fix this bug by adding a NULL check of ctx->active_fmt.
This bug was found by a static analyzer.
Builds with 'make allyesconfig' show no new warnings, and our static
analyzer no longer warns about this code.
Calling hdmi_infoframe_unpack() with static sizeof(buffer) skips all
the size checking done later in hdmi_infoframe_unpack(). A better
value is the amount of data read into buffer.
Fixes: 480b8b3e42c3 ("video/hdmi: Pass buffer size to infoframe unpack functions") Signed-off-by: Tom Rix <trix@redhat.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The M53Menlo display unit uses an LVDS-to-DPI bridge, TI DS90CF364A.
Describe this bridge in DT, otherwise the DT incorrectly describes
DPI panel attached directly to LVDS source.
The %p4cc specifier in some cases might get an unaligned pointer.
Due to this we need to make copy to local variable once to avoid
potential crashes on some architectures due to improper access.
Fixes: af612e43de6d ("lib/vsprintf: Add support for printing V4L2 and DRM fourccs") Cc: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Petr Mladek <pmladek@suse.com> Signed-off-by: Petr Mladek <pmladek@suse.com> Link: https://lore.kernel.org/r/20220127181233.72910-1-andriy.shevchenko@linux.intel.com Signed-off-by: Sasha Levin <sashal@kernel.org>
At the moment, the "qcom-spm-cpuidle" platform device is always created,
even if none of the CPUs is actually managed by the SPM. On non-qcom
platforms this will result in infinite probe-deferral due to the
failing qcom_scm_is_available() call.
To avoid this, look through the CPU DT nodes and check if there is
actually any CPU managed by a SPM (as indicated by the qcom,saw property).
It should also be available because e.g. MSM8916 has qcom,saw defined
but it's typically not enabled with ARM64/PSCI firmwares.
This is needed in preparation of a follow-up change that calls
qcom_scm_set_warm_boot_addr() a single time before registering any
cpuidle drivers. Otherwise this call might be made even on devices
that have this driver enabled but actually make use of PSCI.
This is needed due changes in commit 0519d1d0bf33 ("clk: qcom:
gcc-msm8994: Modernize the driver"), which removed struct
clk_fixed_factor. Preparation for next commit for enabling SD/eMMC.
Inspired by 2c2f64ae36d9.
This is required for both msm8994-huawei-angler (sdhc1 will be enabled
in next commit) and msm8992-lg-bullhead (where actually fixes sdhc1
- tested on bullhead rev 1.01).
Block at <ff800400 0x4c> is a TWD that contains timers, watchdog and
reset. Actual timers happen to be at block beginning but they only span
across the first 0x28 registers. It means the old block description was
incorrect (size 0x3c).
Drop timers binding for now and use documented TWD binding. Timers
should be properly documented and defined as TWD subnode.
Fixes: 2961f69f151c ("arm64: dts: broadcom: add BCM4908 and Asus GT-AC5300 early DTS files") Signed-off-by: Rafał Miłecki <rafal@milecki.pl> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
replace millivolt with correct microvolt and adjust value to
the minimal value allowed by documentation.
Found with `make qcom/sdm845-oneplus-fajita.dtb`.
Fixes:
arch/arm64/boot/dts/qcom/sdm845-oneplus-fajita.dt.yaml: codec@1: 'qcom,micbias1-microvolt' is a required property
From schema: Documentation/devicetree/bindings/sound/qcom,wcd934x.yaml
arch/arm64/boot/dts/qcom/sdm845-oneplus-fajita.dt.yaml: codec@1: 'qcom,micbias2-microvolt' is a required property
From schema: Documentation/devicetree/bindings/sound/qcom,wcd934x.yaml
arch/arm64/boot/dts/qcom/sdm845-oneplus-fajita.dt.yaml: codec@1: 'qcom,micbias3-microvolt' is a required property
From schema: Documentation/devicetree/bindings/sound/qcom,wcd934x.yaml
arch/arm64/boot/dts/qcom/sdm845-oneplus-fajita.dt.yaml: codec@1: 'qcom,micbias4-microvolt' is a required property
From schema: Documentation/devicetree/bindings/sound/qcom,wcd934x.yaml
arch/arm64/boot/dts/qcom/sdm845-oneplus-fajita.dt.yaml: codec@1: 'qcom,micbias1-millivolt', 'qcom,micbias2-millivolt', 'qcom,micbias3-millivolt', 'qcom,micbias4-millivolt' do not match any of the regexes: '^.*@[0-9a-f]+$', 'pinctrl-[0-9]+'
Quoting the header comments, IRQF_ONESHOT is "Used by threaded interrupts
which need to keep the irq line disabled until the threaded handler has
been run.". When applied to an interrupt that doesn't request a threaded
irq then IRQF_ONESHOT has a lesser known (undocumented?) side effect,
which it to disable the forced threading of the irq. For "normal" kernels
(without forced threading) then, if there is no thread_fn, then
IRQF_ONESHOT is a nop.
In this case disabling forced threading is not appropriate for this driver
because it calls wake_up_all() and this API cannot be called from
no-thread interrupt handlers on PREEMPT_RT systems (deadlock risk, triggers
sleeping-while-atomic warnings).
The reference taken by 'of_find_device_by_node()' must be released when
not needed anymore.
Add the corresponding 'put_device()' in the error handling paths.
Fixes: 8c75d585b931 ("soc: qcom: aoss: Expose send for generic usecase") Signed-off-by: Miaoqian Lin <linmq006@gmail.com> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org> Link: https://lore.kernel.org/r/20220108095931.21527-1-linmq006@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org>
The reference taken by 'of_find_device_by_node()' must be released when
not needed anymore.
Add the corresponding 'put_device()' in the error handling path.
Because of the possible failure of the allocation, data->domains might
be NULL pointer and will cause the dereference of the NULL pointer
later.
Therefore, it might be better to check it and directly return -ENOMEM
without releasing data manually if fails, because the comment of the
devm_kmalloc() says "Memory allocated with this function is
automatically freed on driver detach.".
Fixes: bbe3a66c3f5a ("soc: qcom: rpmpd: Add a Power domain driver to model corners") Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org> Link: https://lore.kernel.org/r/20211231094419.1941054-1-jiasheng@iscas.ac.cn Signed-off-by: Sasha Levin <sashal@kernel.org>
Member assignments to qcom_scm_desc were moved into struct initializers
in 57d3b816718c ("firmware: qcom_scm: Remove thin wrappers") including
the case in qcom_scm_iommu_secure_ptbl_init, except that the - now
duplicate - assignment to desc was left in place. While not harmful,
remove this unnecessary extra reassignment.
Fixes: 57d3b816718c ("firmware: qcom_scm: Remove thin wrappers") Signed-off-by: Marijn Suijten <marijn.suijten@somainline.org> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@somainline.org> Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org> Link: https://lore.kernel.org/r/20211208083423.22037-2-marijn.suijten@somainline.org Signed-off-by: Sasha Levin <sashal@kernel.org>
Add a header file that provides the missing function prototypes
and macro to fix these build errors (seen on arch/alpha/):
../sound/soc/codecs/max98927.c: In function 'max98927_i2c_probe':
../sound/soc/codecs/max98927.c:902:19: error: implicit declaration of function 'devm_gpiod_get_optional'; did you mean 'devm_regulator_get_optional'? [-Werror=implicit-function-declaration]
902 | = devm_gpiod_get_optional(&i2c->dev, "reset", GPIOD_OUT_HIGH);
| ^~~~~~~~~~~~~~~~~~~~~~~
../sound/soc/codecs/max98927.c:902:63: error: 'GPIOD_OUT_HIGH' undeclared (first use in this function); did you mean 'GPIOF_INIT_HIGH'?
902 | = devm_gpiod_get_optional(&i2c->dev, "reset", GPIOD_OUT_HIGH);
| ^~~~~~~~~~~~~~
../sound/soc/codecs/max98927.c:909:17: error: implicit declaration of function 'gpiod_set_value_cansleep'; did you mean 'gpio_set_value_cansleep'? [-Werror=implicit-function-declaration]
909 | gpiod_set_value_cansleep(max98927->reset_gpio, 0);
| ^~~~~~~~~~~~~~~~~~~~~~~~
Fixes: 4d67dc1998f1 ("ASoC: max98927: Handle reset gpio when probing i2c") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Reported-by: kernel test robot <lkp@intel.com> Cc: Alejandro Tafalla <atafalla@dnyon.com> Cc: Mark Brown <broonie@kernel.org> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Takashi Iwai <tiwai@suse.com> Cc: alsa-devel@alsa-project.org Link: https://lore.kernel.org/r/20220129080259.19964-1-rdunlap@infradead.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
We've ended up with some inconsistencies between the addresses in the
DT node labels and the actual offsets of the partitions; this brings
them back in sync.
This code has two bugs:
1) "cnt" is 255 but the size of the buffer is 256 so the last byte is
not used.
2) If we try to print more than 255 characters then "cnt" will be
negative and that will trigger a WARN() in snprintf(). The fix for
this is to use scnprintf() instead of snprintf().
We can re-write this code to be cleaner:
1) Rename "offset" to "off" because that's shorter.
2) Get rid of the "cnt" variable and just use "size - off" directly.
3) Get rid of the "read" variable and just increment "off" directly.
Fixes: 96fe6a2109db ("fbdev: Add VESA Coordinated Video Timings (CVT) support") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Helge Deller <deller@gmx.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
If fb_alloc_cmap() fails in ufx_usb_probe(), fb_destroy_modelist() will
be called to destroy modelist in the error handling path. But modelist
has not been initialized yet, so it will result in null-ptr-deref.
Initialize modelist before calling fb_alloc_cmap() to fix this bug.
Fixes: 3c8a63e22a08 ("Add support for SMSC UFX6000/7000 USB display adapters") Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Wang Hai <wanghai38@huawei.com> Acked-by: Thomas Zimmermann <tzimmermann@suse.de> Signed-off-by: Helge Deller <deller@gmx.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
In file included from ./arch/powerpc/include/asm/pgtable.h:20:0,
from ./include/linux/pgtable.h:6,
from ./include/linux/mm.h:33,
from drivers/video/fbdev/controlfb.c:37:
./arch/powerpc/include/asm/nohash/pgtable.h:243:0: note: this is the location of the previous definition
#define pgprot_cached_wthru(prot) (__pgprot((pgprot_val(prot) & ~_PAGE_CACHE_CTL) | \
Start from commit 11be60bd66d54 "matroxfb: add Matrox MGA-G200eW board
support", when maxvram is 0x800000, monitor become black w/ error message
said: "The current input timing is not supported by the monitor display.
Please change your input timing to 1920x1080@60Hz ...".
Fixes: 11be60bd66d5 ("matroxfb: add Matrox MGA-G200eW board support") Signed-off-by: Z. Liu <liuzx@knownsec.com> Signed-off-by: Helge Deller <deller@gmx.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
Pointers V4L2 pixelformat and dataformat fields in a few packed structs
are directly passed to printk family of functions. This could result in an
unaligned access albeit no such possibility appears to exist at the
moment i.e. this clang warning appears to be a false positive.
Address the warning by copying the pixelformat or dataformat value to a
local variable first.
Reported-by: kernel test robot <lkp@intel.com> Fixes: e927e1e0f0dd ("v4l: ioctl: Use %p4cc printk modifier to print FourCC codes") Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
According to subdevice interface specification found in V4L2 API
documentation, set format pad operations should not affect image
geometry set in preceding image processing steps. Unfortunately, that
requirement is not respected by the driver implementation of set format
as it was not the case when that code was still implementing a pair of
now obsolete .s_mbus_fmt() / .try_mbus_fmt() video operations before
they have been merged and reused as an implementation of .set_fmt() pad
operation by commit 717fd5b4907a ("[media] v4l2: replace try_mbus_fmt
by set_fmt").
In case of set format active processing path the issue can be fixed
easily by excluding a call to set active selection from that path. That
will effectively limit frame size processing to optimal frame scaling
against active crop rectangle without touching it. Users can just call
set active selection themselves to obtain desired frame size. However,
set format try processing path needs more work.
First of all, the driver should be extended with set try selection
support. Lack of it constraints video device drivers to not use
subdevice cropping at all while processing user requested active frame
size, otherwise their set try format results might differ from active.
Next, set format try processing path should use pad config crop
rectangle as a reference, not the active one as it does now. That
issue can be resolved easily as soon as set try selection support is
added to the driver so pad config crop rectangle can be maintained by
users via selection API.
Last, set format try processing path should give the same results as
active in respect to active vs. pad config crop rectangle geometry.
Both rectangles should be either not touched by set format (that's what
we are going to achieve) or modified the same way, otherwise users
won't be able to obtain equal results from both paths while iterating
through set format and set selection operations in order to obtain
desired frame size.
We can't begin with modifying set format pad operation as not to touch
crop rectangle since that depends on availability of set try selection
for symmetry. Neither can we begin with adding set try selection since
that in turn depends on equal handling of active and pad config crop
rectangles by set format. We can either implement all required
modifications in a single patch, or begin with fixing current set
format try processing path to appropriately handle pad config crop
rectangle. This patch implements the latter approach as believed to
be more readable.
Move crop rectangle adjustments code from a helper (the former
implementation of .s_fmt(), now called from set format active
processing path) to the body of set format pad operation function
where it can be also used for processing try requests for symmetry with
active ones. As the helper no longer processes frame geometry, only
frame format and half scaling, simplify its API accordingly and update
its users.
Moreover, extract code that applies crop rectangle hardware limits
(now a part of .set_selection() operation which is called from set
format active processing path) to a new helper and call that helper
from set format try processing path as well for symmetry with active.
[Sakari Ailus: Rebase on subdev state patches]
Fixes: 717fd5b4907a ("[media] v4l2: replace try_mbus_fmt by set_fmt") Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The Hantro H1 hardware can crop off pixels from the right and bottom of
the source frame. These are controlled with the H1_REG_IN_IMG_CTRL_OVRFLB
and H1_REG_IN_IMG_CTRL_OVRFLR in the H1_REG_IN_IMG_CTRL register.
The ChromeOS kernel driver that this was based on incorrectly added the
_D4 suffix H1_REG_IN_IMG_CTRL_OVRFLB. This field crops the bottom of the
input frame, and the number is _not_ divided by 4. [1]
Correct the name to avoid confusion when crop support with the selection
API is added.
As the possible failure of the kzalloc(), the 'new_ts' could be NULL
pointer.
Therefore, it should be better to check it in order to avoid the
dereference of the NULL pointer.
Also, the caller esparser_queue() needs to deal with the return value of
the amvdec_add_ts().
Fixes: 876f123b8956 ("media: meson: vdec: bring up to compliance") Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn> Suggested-by: Neil Armstrong <narmstrong@baylibre.com> Reviewed-by: Neil Armstrong <narmstrong@baylibre.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The reference taken by 'of_find_device_by_node()' must be released when
not needed anymore.
Add the corresponding 'put_device()' in the error handling path.
Fixes: e7f3c5481035 ("[media] coda: use VDOA for un-tiling custom macroblock format") Signed-off-by: Miaoqian Lin <linmq006@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Memory chip select is swapped when we read mode register, correct it.
We didn't have devices that use a single LPDDR chip and both chips are
always identical, hence this change is just a minor improvement.
Fixes: 131dd9a436d8 ("memory: tegra20-emc: Support matching timings by LPDDR2 configuration") Signed-off-by: Dmitry Osipenko <digetx@gmail.com> Link: https://lore.kernel.org/r/20211222043215.28237-2-digetx@gmail.com Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
If an mclk-fs value was provided in the device tree configuration, the
calculated MCLK was fed into the downstream codec DAI and CPU DAI,
however set_sysclk was not being called on the platform device. Some
platform devices such as the Xilinx Audio Formatter need to know the MCLK
as well.
Call snd_soc_component_set_sysclk on each component in the stream to set
the proper sysclk value in addition to the existing call of
snd_soc_dai_set_sysclk on the codec DAI and CPU DAI. This may end up
resulting in redundant calls if one of the snd_soc_dai_set_sysclk calls
ends up calling snd_soc_component_set_sysclk itself, but that isn't
expected to cause any significant harm.
Fixes: f48dcbb6d47d ("ASoC: simple-card-utils: share asoc_simple_hw_param()") Signed-off-by: Robert Hancock <robert.hancock@calian.com> Reviewed-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Link: https://lore.kernel.org/r/20220120195832.1742271-5-robert.hancock@calian.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This driver did not set the MM2S Fs Multiplier Register to the proper
value for playback streams. This needs to be set to the sample rate to
MCLK multiplier, or random stream underflows can occur on the downstream
I2S transmitter.
Store the sysclk value provided via the set_sysclk callback and use that
in conjunction with the sample rate in the hw_params callback to calculate
the proper value to set for this register.
Fixes: 6f6c3c36f091 ("ASoC: xlnx: add pcm formatter platform driver") Signed-off-by: Robert Hancock <robert.hancock@calian.com> Link: https://lore.kernel.org/r/20220120195832.1742271-2-robert.hancock@calian.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since the potential failure of the devm_regmap_init_mmio(), it will
return error pointer and be assigned to the regmap.
Then the error pointer will be dereferenced.
For example rx->regmap will be used in rx_macro_mclk_enable().
Therefore, it should be better to check it.
Fixes: af3d54b99764 ("ASoC: codecs: lpass-rx-macro: add support for lpass rx macro") Fixes: c39667ddcfc5 ("ASoC: codecs: lpass-tx-macro: add support for lpass tx macro") Fixes: 809bcbcecebf ("ASoC: codecs: lpass-wsa-macro: Add support to WSA Macro") Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn> Link: https://lore.kernel.org/r/20220121171031.2826198-1-jiasheng@iscas.ac.cn Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 2161536516ed ("media: media/pci: set device_caps in struct video_device")
introduced a regression: V4L2_CAP_TUNER is always present in device_caps,
even when the device has no tuner.
This causes a warning:
WARNING: CPU: 0 PID: 249 at drivers/media/v4l2-core/v4l2-ioctl.c:1102 v4l_querycap+0xa0/0xb0 [videodev]
Fixes: 2161536516ed ("media: media/pci: set device_caps in struct video_device") Signed-off-by: Ondrej Zary <linux@zary.sk> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
during system resume, interconnect bandwidth would currently be requested
even though the device is runtime suspended. This leaves the system in an
unbalanced state.
Fix that by only doing that in runtimem pm and splitting up runtime and
system suspend to be a more readable:
imx8mq_mipi_csi_pm_*() does the generic things called from system- and
runtime functions that each do specific things on top.
Fixes: f33fd8d77dd0 ("media: imx: add a driver for i.MX8MQ mipi csi rx phy and controller") Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The place where this register writel() that masks one interrupt is placed
does not guarantee that the device is powered so that's not allowed.
Moreover imx8mq_mipi_csi_start_stream() masks the interrupt anyway so the
write is not even needed. Remove it as this is a mistake that slipped in
with the driver.
Fixes: f33fd8d77dd0 ("media: imx: add a driver for i.MX8MQ mipi csi rx phy and controller") Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
DST_QUEUE_OFF_BASE is applied to offset/mem_offset on MMAP capture buffers
only for the VIDIOC_QUERYBUF ioctl, while the userspace fields (including
offset/mem_offset) are filled in for VIDIOC_{QUERY,PREPARE,Q,DQ}BUF
ioctls. This leads to differences in the values presented to userspace.
If userspace attempts to mmap the capture buffer directly using values
from DQBUF, it will fail.
Move the code that applies the magic offset into a helper, and call
that helper from all four ioctl entry points.
[hverkuil: drop unnecessary '= 0' in v4l2_m2m_querybuf() for ret]
Fixes: 7f98639def42 ("V4L/DVB: add memory-to-memory device helper framework for videobuf") Fixes: 908a0d7c588e ("[media] v4l: mem2mem: port to videobuf2") Signed-off-by: Chen-Yu Tsai <wenst@chromium.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When multiple CSIS instances are present in a single graph, they are
currently all named "imx7-mipi-csis.0", which breaks the entity name
uniqueness requirement. Fix it by using the device name to create the
subdev name.
This function waits for halt_complete but doesn't do anything to cause
it to complete, and always hits the "VFE halt timeout" error. Just delete
this code for now.
Fixes: 7319cdf189bb ("media: camss: Add support for VFE hardware version Titan 170") Signed-off-by: Jonathan Marek <jonathan@marek.ca> Reviewed-by: Robert Foss <robert.foss@linaro.org> Tested-by: Julian Grahsl <jgrahsl@snap.com> Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Use the "HALT_CMD_RESUME_AT_FRAME_BOUNDARY" define instead of a "1" which
is otherwise confusing, and add a "HALT_CMD_HALT_AT_FRAME_BOUNDARY" which
is set when disabling.
Fixes: eebe6d00e9bf ("media: camss: Add support for CSID hardware version Titan 170") Signed-off-by: Jonathan Marek <jonathan@marek.ca> Reviewed-by: Robert Foss <robert.foss@linaro.org> Tested-by: Julian Grahsl <jgrahsl@snap.com> Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
csid_isr() only checks for the reset irq, so enabling any other irqs
doesn't make sense. The "RDI irq" comment is also wrong, the register
should be CSID_CSI2_RDIN_IRQ_MASK. Without this fix there may be an
excessive amount of irqs.
Fixes: eebe6d00e9bf ("media: camss: Add support for CSID hardware version Titan 170") Signed-off-by: Jonathan Marek <jonathan@marek.ca> Reviewed-by: Robert Foss <robert.foss@linaro.org> Tested-by: Julian Grahsl <jgrahsl@snap.com> Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
vb2_dma_contig_set_max_seg_size need to have a size in parameter and not
a DMA_BIT_MASK().
While fixing this issue, also fix error handling of all DMA size
setting.
Reported-by: kernel test robot <lkp@intel.com> Fixes: d4ae3689226e5 ("media: zoran: device support only 32bit DMA address") Signed-off-by: Corentin Labbe <clabbe@baylibre.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
GPIO_ALIVE and GPIO_CMGP blocks in Exynos850 SoC don't have EINT
capabilities (like EINT_SVC register), and there are no corresponding
interrupts wired to GIC. Instead those blocks have wake-up interrupts
for each pin. The ".eint_gpio_init" callbacks were specified by mistake
for these blocks, when porting pinctrl code from downstream kernel. That
leads to error messages like this:
samsung-pinctrl 11850000.pinctrl: irq number not available
Remove ".eint_gpio_init" for pinctrl_alive and pinctrl_gpmc to fix this
error. This change doesn't affect proper interrupt handling for related
pins, as all those pins are handled in ".eint_wkup_init".
Fixes: cdd3d945dcec ("pinctrl: samsung: Add Exynos850 SoC specific data") Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org> Link: https://lore.kernel.org/r/20220114203757.4860-1-semen.protsenko@linaro.org Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In function kunit_test_timeout, it is declared "300 * MSEC_PER_SEC"
represent 5min. However, it is wrong when dealing with arm64 whose
default HZ = 250, or some other situations. Use msecs_to_jiffies to fix
this, and kunit_test_timeout will work as desired.
Link: https://lkml.kernel.org/r/20220309083753.1561921-3-liupeng256@huawei.com Fixes: 5f3e06208920 ("kunit: test: add support for test abort") Signed-off-by: Peng Liu <liupeng256@huawei.com> Reviewed-by: Marco Elver <elver@google.com> Reviewed-by: Daniel Latypov <dlatypov@google.com> Reviewed-by: Brendan Higgins <brendanhiggins@google.com> Tested-by: Brendan Higgins <brendanhiggins@google.com> Cc: Alexander Potapenko <glider@google.com> Cc: Dmitry Vyukov <dvyukov@google.com> Cc: Wang Kefeng <wangkefeng.wang@huawei.com> Cc: David Gow <davidgow@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If register_memory() fails, we freed the memory block but already added
the memory block to the group list, not good. Let's defer adding the
block to the memory group to after registering the memory block device.
We do handle it properly during unregister_memory(), but that's not
called when the registration fails.
Link: https://lkml.kernel.org/r/20220128144540.153902-1-david@redhat.com Fixes: 028fc57a1c36 ("drivers/base/memory: introduce "memory groups" to logically group memory blocks") Signed-off-by: David Hildenbrand <david@redhat.com> Reviewed-by: Oscar Salvador <osalvador@suse.de> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The $(CC) variable used in Makefiles could contain several arguments
such as "ccache gcc". These need to be passed as a single string to
check_cc.sh, otherwise only the first argument will be used as the
compiler command. Without quotes, the $(CC) variable is passed as
distinct arguments which causes the script to fail to build trivial
programs.
Fix this by adding quotes around $(CC) when calling check_cc.sh to pass
the whole string as a single argument to the script even if it has
several words such as "ccache gcc".
When compressed file has blocks, f2fs_ioc_start_atomic_write will succeed,
but compressed flag will be remained in inode. If write partial compreseed
cluster and commit atomic write will cause data corruption.
This is the reproduction process:
Step 1:
create a compressed file ,write 64K data , call fsync(), then the blocks
are write as compressed cluster.
Step2:
iotcl(F2FS_IOC_START_ATOMIC_WRITE) --- this should be fail, but not.
write page 0 and page 3.
iotcl(F2FS_IOC_COMMIT_ATOMIC_WRITE) -- page 0 and 3 write as normal file,
Step3:
drop cache.
read page 0-4 -- Since page 0 has a valid block address, read as
non-compressed cluster, page 1 and 2 will be filled with compressed data
or zero.
The root cause is, after commit 7eab7a696827 ("f2fs: compress: remove
unneeded read when rewrite whole cluster"), in step 2, f2fs_write_begin()
only set target page dirty, and in f2fs_commit_inmem_pages(), we will write
partial raw pages into compressed cluster, result in corrupting compressed
cluster layout.
Fixes: 4c8ff7095bef ("f2fs: support data compression") Fixes: 7eab7a696827 ("f2fs: compress: remove unneeded read when rewrite whole cluster") Reported-by: kernel test robot <lkp@intel.com> Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Fengnan Chang <changfengnan@vivo.com> Reviewed-by: Chao Yu <chao@kernel.org> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When reflinking an inline extent, we assert that its file offset is 0 and
that its uncompressed length is not greater than the sector size. We then
return an error if one of those conditions is not satisfied. However we
use a return statement, which results in returning from btrfs_clone()
without freeing the path and buffer that were allocated before, as well as
not clearing the flag BTRFS_INODE_NO_DELALLOC_FLUSH for the destination
inode.
Fix that by jumping to the 'out' label instead, and also add a WARN_ON()
for each condition so that in case assertions are disabled, we get to
known which of the unexpected conditions triggered the error.
Fixes: a61e1e0df9f321 ("Btrfs: simplify inline extent handling when doing reflinks") Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Quoted from Jing Xia's report, there is a potential deadlock may happen
between kworker and checkpoint as below:
[T:writeback] [T:checkpoint]
- wb_writeback
- blk_start_plug
bio contains NodeA was plugged in writeback threads
- do_writepages -- sync write inodeB, inc wb_sync_req[DATA]
- f2fs_write_data_pages
- f2fs_write_single_data_page -- write last dirty page
- f2fs_do_write_data_page
- set_page_writeback -- clear page dirty flag and
PAGECACHE_TAG_DIRTY tag in radix tree
- f2fs_outplace_write_data
- f2fs_update_data_blkaddr
- f2fs_wait_on_page_writeback -- wait NodeA to writeback here
- inode_dec_dirty_pages
- writeback_sb_inodes
- writeback_single_inode
- do_writepages
- f2fs_write_data_pages -- skip writepages due to wb_sync_req[DATA]
- wbc->pages_skipped += get_dirty_pages() -- PAGECACHE_TAG_DIRTY is not set but get_dirty_pages() returns one
- requeue_inode -- requeue inode to wb->b_dirty queue due to non-zero.pages_skipped
- blk_finish_plug
Let's try to avoid deadlock condition by forcing unplugging previous bio via
blk_finish_plug(current->plug) once we'v skipped writeback in writepages()
due to valid sbi->wb_sync_req[DATA/NODE].
Fixes: 687de7f1010c ("f2fs: avoid IO split due to mixed WB_SYNC_ALL and WB_SYNC_NONE") Signed-off-by: Zhiguo Niu <zhiguo.niu@unisoc.com> Signed-off-by: Jing Xia <jing.xia@unisoc.com> Signed-off-by: Chao Yu <chao@kernel.org> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
John reported that push_rt_task() can end up invoking
find_lowest_rq(rq->curr) when curr is not an RT task (in this case a CFS
one), which causes mayhem down convert_prio().
This can happen when current gets demoted to e.g. CFS when releasing an
rt_mutex, and the local CPU gets hit with an rto_push_work irqwork before
getting the chance to reschedule. Exactly who triggers this work isn't
entirely clear to me - switched_from_rt() only invokes rt_queue_pull_task()
if there are no RT tasks on the local RQ, which means the local CPU can't
be in the rto_mask.
My current suspected sequence is something along the lines of the below,
with the demoted task being current.
mark_wakeup_next_waiter()
rt_mutex_adjust_prio()
rt_mutex_setprio() // deboost originally-CFS task
check_class_changed()
switched_from_rt() // Only rt_queue_pull_task() if !rq->rt.rt_nr_running
switched_to_fair() // Sets need_resched
__balance_callbacks() // if pull_rt_task(), tell_cpu_to_push() can't select local CPU per the above
raw_spin_rq_unlock(rq)
// need_resched is set, so task_woken_rt() can't
// invoke push_rt_tasks(). Best I can come up with is
// local CPU has rt_nr_migratory >= 2 after the demotion, so stays
// in the rto_mask, and then:
<some other CPU running rto_push_irq_work_func() queues rto_push_work on this CPU>
push_rt_task()
// breakage follows here as rq->curr is CFS
Move an existing check to check rq->curr vs the next pushable task's
priority before getting anywhere near find_lowest_rq(). While at it, add an
explicit sched_class of rq->curr check prior to invoking
find_lowest_rq(rq->curr). Align the DL logic to also reschedule regardless
of next_task's migratability.
Fixes: a7c81556ec4d ("sched: Fix migrate_disable() vs rt/dl balancing") Reported-by: John Keeping <john@metanate.com> Signed-off-by: Valentin Schneider <valentin.schneider@arm.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com> Tested-by: John Keeping <john@metanate.com> Link: https://lore.kernel.org/r/20220127154059.974729-1-valentin.schneider@arm.com Signed-off-by: Sasha Levin <sashal@kernel.org>
The cpuacct_account_field() is always called by the current task
itself, so it's ok to use __this_cpu_add() to charge the tick time.
But cpuacct_charge() maybe called by update_curr() in load_balance()
on a random CPU, different from the CPU on which the task is running.
So __this_cpu_add() will charge that cputime to a random incorrect CPU.
The nfsd file cache table can be pretty large and its allocation
may require as many as 80 contigious pages.
Employ the same fix that was employed for similar issue that was
reported for the reply cache hash table allocation several years ago
by commit 8f97514b423a ("nfsd: more robust allocation failure handling
in nfsd_reply_cache_init").
Fixes: 65294c1f2c5e ("nfsd: add a new struct file caching facility to nfsd") Link: https://lore.kernel.org/linux-nfs/e3cdaeec85a6cfec980e87fc294327c0381c1778.camel@kernel.org/ Suggested-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Tested-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
There are inconsistencies when determining if a NUMA imbalance is allowed
that should be corrected.
o allow_numa_imbalance changes types and is not always examining
the destination group so both the type should be corrected as
well as the naming.
o find_idlest_group uses the sched_domain's weight instead of the
group weight which is different to find_busiest_group
o find_busiest_group uses the source group instead of the destination
which is different to task_numa_find_cpu
o Both find_idlest_group and find_busiest_group should account
for the number of running tasks if a move was allowed to be
consistent with task_numa_find_cpu
Fixes: 7d2b5dd0bcc4 ("sched/numa: Allow a floating imbalance between NUMA nodes") Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com> Link: https://lore.kernel.org/r/20220208094334.16379-2-mgorman@techsingularity.net Signed-off-by: Sasha Levin <sashal@kernel.org>
Change from shifting 'unsigned long' to 'u64' to prevent the config bits
being lost on a 32-bit kernel.
Fixes: eadf48cab4b6b0 ("perf/x86/intel/pt: Add support for address range filtering in PT") Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Link: https://lore.kernel.org/r/20220131072453.2839535-5-adrian.hunter@intel.com Signed-off-by: Sasha Levin <sashal@kernel.org>
The rseq rseq_cs.ptr.{ptr32,padding} uapi endianness handling is
entirely wrong on 32-bit little endian: a preprocessor logic mistake
wrongly uses the big endian field layout on 32-bit little endian
architectures.
Fortunately, those ptr32 accessors were never used within the kernel,
and only meant as a convenience for user-space.
Remove those and replace the whole rseq_cs union by a __u64 type, as
this is the only thing really needed to express the ABI. Document how
32-bit architectures are meant to interact with this field.
Fixes: ec9c82e03a74 ("rseq: uapi: Declare rseq_cs field as union, update includes") Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Link: https://lkml.kernel.org/r/20220127152720.25898-1-mathieu.desnoyers@efficios.com Signed-off-by: Sasha Levin <sashal@kernel.org>
iowait_boost signal is applied independently of util and doesn't take
into account uclamp settings of the rq. An io heavy task that is capped
by uclamp_max could still request higher frequency because
sugov_iowait_apply() doesn't clamp the boost via uclamp_rq_util_with()
like effective_cpu_util() does.
Make sure that iowait_boost honours uclamp requests by calling
uclamp_rq_util_with() when applying the boost.
Fixes: 982d9cdc22c9 ("sched/cpufreq, sched/uclamp: Add clamps for FAIR and RT tasks") Signed-off-by: Qais Yousef <qais.yousef@arm.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Link: https://lore.kernel.org/r/20211216225320.2957053-3-qais.yousef@arm.com Signed-off-by: Sasha Levin <sashal@kernel.org>
The older format of /proc/pid/sched printed home node info which
required the mempolicy and task lock around mpol_get(). However
the format has changed since then and there is no need for
sched_show_numa() any more to have mempolicy argument,
asssociated mpol_get/put and task_lock/unlock. Remove them.
Fixes: 397f2378f1361 ("sched/numa: Fix numa balancing stats in /proc/pid/sched") Signed-off-by: Bharata B Rao <bharata@amd.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Acked-by: Mel Gorman <mgorman@suse.de> Link: https://lore.kernel.org/r/20220118050515.2973-1-bharata@amd.com Signed-off-by: Sasha Levin <sashal@kernel.org>
projects / thirdparty / kernel / stable.git / log
