GCC versions below 13 incorrectly detect the copy size as being static and too
small to fit in the "fds" array. Work around this by explicitly calculating the
size and returning EINVAL based on that, instead of based on the object count.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202502072019.LYoCR9bF-lkp@intel.com/
Suggested-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
--
Suggested-by as per Arnd's request, but the only thing I changed was preserving
array_size() [as noted by Geert in the linked thread]. I tested and found no
regressions.
v2: Add missing sign-off
Link: https://lore.kernel.org/r/20250220192334.549167-1-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Al Viro pointed out that dput() might sleep and must not be invoked
within an RCU section.
Keep only find_next_ancestor() winthin the RCU section.
Correct the wording in the comment.
Fixes: 6ef5b6fae3 ("kernfs: Drop kernfs_rwsem while invoking lookup_positive_unlocked().")
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://lore.kernel.org/r/20250221084232.xksA_IQ4@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The Intel Timed IO PPS generator driver outputs a PPS signal using
dedicated hardware that is more accurate than software actuated PPS.
The Timed IO hardware generates output events using the ART timer.
The ART timer period varies based on platform type, but is less than 100
nanoseconds for all current platforms. Timed IO output accuracy is
within 1 ART period.
PPS output is enabled by writing '1' the 'enable' sysfs attribute. The
driver uses hrtimers to schedule a wake-up 10 ms before each event
(edge) target time. At wakeup, the driver converts the target time in
terms of CLOCK_REALTIME to ART trigger time and writes this to the Timed
IO hardware. The Timed IO hardware generates an event precisely at the
requested system time without software involvement.
Co-developed-by: Christopher Hall <christopher.s.hall@intel.com>
Signed-off-by: Christopher Hall <christopher.s.hall@intel.com>
Co-developed-by: Pandith N <pandith.n@intel.com>
Signed-off-by: Pandith N <pandith.n@intel.com>
Co-developed-by: Thejesh Reddy T R <thejesh.reddy.t.r@intel.com>
Signed-off-by: Thejesh Reddy T R <thejesh.reddy.t.r@intel.com>
Signed-off-by: Lakshmi Sowjanya D <lakshmi.sowjanya.d@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Rodolfo Giometti <giometti@enneenne.com>
Signed-off-by: Subramanian Mohan <subramanian.mohan@intel.com>
Link: https://lore.kernel.org/r/20250219040618.70962-3-subramanian.mohan@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some PPS generator drivers may need to retrieve a pointer to their
internal data while executing the PPS generator enable() method.
During the driver registration the pps_gen_device pointer is returned
from the framework, and for that reason, there is difficulty in
getting generator driver data back in the enable function. We won't be
able to use container_of macro as it results in static assert, and we
might end up in using static pointer.
To solve the issue and to get back the generator driver data back, we
should not copy the struct pps_gen_source_info within the struct
pps_gen_device during the registration stage, but simply save the
pointer of the driver one. In this manner, driver may get a pointer
to their internal data as shown below:
struct pps_gen_foo_data_s {
...
struct pps_gen_source_info gen_info;
struct pps_gen_device *pps_gen;
...
};
static int __init pps_gen_foo_init(void)
{
struct pps_gen_foo_data_s *foo;
...
foo->pps_gen = pps_gen_register_source(&foo->gen_info);
...
}
Then, in the enable() method, we can retrieve the pointer to the main
struct by using the code below:
static int pps_gen_foo_enable(struct pps_gen_device *pps_gen, bool enable)
{
struct pps_gen_foo_data_s *foo = container_of(pps_gen->info,
struct pps_gen_foo_data_s, gen_info);
...
}
Signed-off-by: Rodolfo Giometti <giometti@enneenne.com>
Tested-by: Subramanian Mohan <subramanian.mohan@intel.com>
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Subramanian Mohan <subramanian.mohan@intel.com>
Link: https://lore.kernel.org/r/20250219040618.70962-2-subramanian.mohan@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add support for the 'eUSB2 Isochronous Endpoint Companion Descriptor'
introduced in the recent USB 2.0 specification 'USB 2.0 Double Isochronous
IN Bandwidth' ECN.
It allows embedded USB2 (eUSB2) devices to report and use higher bandwidths
for isochronous IN transfers in order to support higher camera resolutions
on the lid of laptops and tablets with minimal change to the USB2 protocol.
The motivation for expanding USB 2.0 is further clarified in an additional
Embedded USB2 version 2.0 (eUSB2v2) supplement to the USB 2.0
specification. It points out this is optimized for performance, power and
cost by using the USB 2.0 low-voltage, power efficient PHY and half-duplex
link for the asymmetric camera bandwidth needs, avoiding the costly and
complex full-duplex USB 3.x symmetric link and gigabit receivers.
eUSB2 devices that support the higher isochronous IN bandwidth and the new
descriptor can be identified by their device descriptor bcdUSB value of
0x0220
Co-developed-by: Amardeep Rai <amardeep.rai@intel.com>
Signed-off-by: Amardeep Rai <amardeep.rai@intel.com>
Signed-off-by: Kannappan R <r.kannappan@intel.com>
Co-developed-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Link: https://lore.kernel.org/r/20250220141339.1939448-1-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Document the compatible string "samsung,exynos7870-chipid". The
registers are entirely compatible with "samsung,exynos4210-chipid".
Signed-off-by: Kaustabh Chakraborty <kauschluss@disroot.org>
Link: https://lore.kernel.org/r/20250219-exynos7870-v3-1-e384fb610cad@disroot.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
While searching for a reference CPU within a given policy,
arch_freq_get_on_cpu relies on cpumask_next_wrap to iterate over
all available CPUs and to ensure each is verified only once.
Recent changes to cpumask_next_wrap will handle the latter no more,
so switching to for_each_cpu_wrap, which preserves expected behavior
while ensuring compatibility with the updates.
Not to mention that when iterating over each CPU, using a dedicated
iterator is preferable to an open-coded loop.
Fixes: 16d1e27475 ("arm64: Provide an AMU-based version of arch_freq_get_on_cpu")
Signed-off-by: Beata Michalska <beata.michalska@arm.com>
Link: https://lore.kernel.org/r/20250220091015.2319901-1-beata.michalska@arm.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
'struct tunertype', 'struct tuner_range' and 'struct tuner_params' are not
modified in this driver.
Constifying these structures moves some data to a read-only section, so
increase overall security.
On a x86_64, with allmodconfig:
Before:
======
2877 8554 0 11431 2ca7 drivers/media/tuners/tuner-types.o
After:
=====
text data bss dec hex filename
11421 48 0 11469 2ccd drivers/media/tuners/tuner-types.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
'struct i2c_algorithm' are not modified in these drivers.
Constifying this structure moves some data to a read-only section, so
increase overall security, especially when the structure holds some
function pointers.
More over, dvb_usb_device_properties->i2c_algo seems to only be copied in
i2c_adapter->algo, which is already a "const struct i2c_algorithm".
This is done in dvb_usbv2_i2c_init()
On a x86_64, with allmodconfig, as an example:
Before:
======
text data bss dec hex filename
35366 5832 36 41234 a112 drivers/media/usb/dvb-usb-v2/af9015.o
After:
=====
text data bss dec hex filename
35430 5768 36 41234 a112 drivers/media/usb/dvb-usb-v2/af9015.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
'struct i2c_algorithm' are not modified in these drivers.
Constifying this structure moves some data to a read-only section, so
increase overall security, especially when the structure holds some
function pointers.
More over, dvb_usb_device_properties->i2c_algo seems to only be copied in
i2c_adapter->algo, which is already a "const struct i2c_algorithm".
This is done in dvb_usb_i2c_init()
On a x86_64, with allmodconfig, as an example:
Before:
======
text data bss dec hex filename
30571 5916 36 36523 8eab drivers/media/usb/dvb-usb/af9005.o
After:
=====
text data bss dec hex filename
30667 5852 36 36555 8ecb drivers/media/usb/dvb-usb/af9005.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
'struct i2c_algorithm' is not modified in this driver.
Constifying this structure moves some data to a read-only section, so
increase overall security, especially when the structure holds some
function pointers.
To do so, the prototype of i2c_adapter_init() has to be updated as well.
On a x86_64, with allmodconfig:
Before:
======
text data bss dec hex filename
17213 932 20 18165 46f5 drivers/media/dvb-frontends/dibx000_common.o
After:
=====
text data bss dec hex filename
17490 660 20 18170 46fa drivers/media/dvb-frontends/dibx000_common.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
This function is unused and unlikely to be used in the near future.
Remove it.
Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The code is running arithmetics with the enum, which when not done with
care makes the compiler a bit nervous.
Because those enums are only used as defines (no argument or variable
from that enumeration type), convert it into a define and move on.
It is required to build with llvm 9 without these warnings:
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:673:17: warning: bitwise operation between different enumeration types ('enum TS_PORT' and 'enum POWE_TYPE') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:680:21: warning: bitwise operation between different enumeration types ('enum AVDEC_STATUS' and 'enum POWE_TYPE') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:687:21: warning: bitwise operation between different enumeration types ('enum AVDEC_STATUS' and 'enum POWE_TYPE') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:702:17: warning: bitwise operation between different enumeration types ('enum TS_PORT' and 'enum POWE_TYPE') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:709:21: warning: bitwise operation between different enumeration types ('enum TS_PORT' and 'enum POWE_TYPE') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:718:21: warning: bitwise operation between different enumeration types ('enum AVDEC_STATUS' and 'enum POWE_TYPE') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:727:21: warning: bitwise operation between different enumeration types ('enum AVDEC_STATUS' and 'enum TS_PORT') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:737:21: warning: bitwise operation between different enumeration types ('enum AVDEC_STATUS' and 'enum TS_PORT') [-Wenum-enum-conversion]
drivers/media/usb/cx231xx/cx231xx-pcb-cfg.c:749:21: warning: bitwise operation between different enumeration types ('enum AVDEC_STATUS' and 'enum TS_PORT') [-Wenum-enum-conversion]
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
[hverkuil: fix a tiny typo in the commit log]
After suspend-to-memory or suspend-to-disk, additional chips are no longer
reachable via i2c. Trying to tune to DVB-C on a cx23885 based
Hauppauge WinTV-HVR-4400-HD:
si2165 8-0064: could not set chip_mode
tda18271: performing RF tracking filter calibration
This patch implements the simplest possible suspend/resume that is
enough to tune to dvb-c channel after resume.
Afterwards dmesg looks like this:
si2165 8-0064: downloading firmware from file 'dvb-demod-si2165.fw' \
size=5768
si2165 8-0064: si2165_upload_firmware: extracted patch_version=0x9a, \
block_count=0x27, crc_expected=0xcc0a
si2165 8-0064: fw load finished
tda18271: performing RF tracking filter calibration
tda18271: RF tracking filter calibration complete
Signed-off-by: Matthias Schwarzott <zzam@gentoo.org>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
When the system goes into sleep mode, the dependent
power will be turned off, so when the system resumes,
coda_hw_init() needs to be called, otherwise it will
get 'CODA PIC_RUN timeout' after resuming.
Signed-off-by: Xiaolei Wang <xiaolei.wang@windriver.com>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
'struct usb_device_id' is not modified in these drivers.
Constifying this structure moves some data to a read-only section, so
increase overall security.
In order to do that, struct dvb_usb_device_description (in dvb-usb.h) also
needs to be updated.
On a x86_64, with allmodconfig, as an example:
Before:
======
text data bss dec hex filename
1983 4240 4 6227 1853 drivers/media/usb/dvb-usb/a800.o
After:
=====
text data bss dec hex filename
2079 4144 4 6227 1853 drivers/media/usb/dvb-usb/a800.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The ADV7180 family of chips have an INTRQ pin that can be connected
to a SoC GPIO.
Allow the 'interrupts' property to be described to fix the following
dt-schema warning:
'interrupt-parent', 'interrupts' do not match any of the regexes: 'pinctrl-[0-9]+'
Signed-off-by: Fabio Estevam <festevam@denx.de>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Use kmalloc_array() to replace kmalloc() with multiplication.
kmalloc_array() has multiply overflow check, which will be safer.
Signed-off-by: Zhang Heng <zhangheng@kylinos.cn>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
It was originally intended to introduce simple_strtol,
but since commit a081c3400f ("[media] pwc: Remove dev_hint
module parameter"), that simple_strtol has been removed,
so in order to prevent disputes, the header file should be removed.
Signed-off-by: Zhang Heng <zhangheng@kylinos.cn>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
[hverkuil: fixed small typos in commit log]
The mask to select the test-pattern in register ADV748X_SDP_FRP is
incorrect, it's the lower 3 bits which controls the pattern. The
GENMASK() macro is used incorrectly and the generated mask is 0x0e
instead of 0x07.
The result is that not all test patterns are selectable, and that in
some cases the wrong test pattern is activated. Fix this by correcting
the GENMASK().
Fixes: 3e89586a64 ("media: i2c: adv748x: add adv748x driver")
Cc: stable@vger.kernel.org
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
[hverkuil: fixed tiny typo in commit log: my -> by]
Added support for GMSL modules variants 3 and 4. Variant 3 is the same as
variant 2 from the driver's point of view. Variant 4 has "hardwired" daisy
chain loopback outputs and thus missing the v4l2 outputs.
Signed-off-by: Martin Tůma <martin.tuma@digiteqautomotive.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Document the (new) mgb4 GMSL modules variants.
Signed-off-by: Martin Tůma <martin.tuma@digiteqautomotive.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Correct the erroneous spelling of 'responsible' in
Documentation/admin-guide/media/cec.rst. This fixes all errors
pointed out by codespell in the file.
Signed-off-by: Chandra Pratap <chandrapratap3519@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
On Mediatek devices with a system companion processor (SCP) the mtk_scp
structure has to be removed explicitly to avoid a resource leak.
Free the structure in case the allocation of the firmware structure fails
during the firmware initialization.
Fixes: 53dbe08504 ("media: mtk-vcodec: potential null pointer deference in SCP")
Cc: stable@vger.kernel.org
Signed-off-by: Jiasheng Jiang <jiashengjiangcool@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The define used for the version in the example diagram does not match what
is defined in enum rksip1_ext_param_buffer_version, nor the description
above it. Correct the typo to make it clear which define to use.
Fixes: e9d05e9d5d ("media: uapi: rkisp1-config: Add extensible params format")
Cc: stable@vger.kernel.org
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Convert aspeed-video.txt to yaml format.
Update aspeed-video.txt to aspeed,video-engine.yaml in MAINTAINER file.
Signed-off-by: Jammy Huang <jammy_huang@aspeedtech.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Andrew Jeffery <andrew@codeconstruct.com.au>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The MIPID02 can use up to 2 data lanes which leads to having a maximum
item number of 3 for the lane-polarities since this also contains the
clock lane.
CC: stable@vger.kernel.org
Fixes: c2741cbe7f ("dt-bindings: media: st,stmipid02: Convert the text bindings to YAML")
Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Build-testing on x86 showed two modules with conflicting "iris.ko"
names after the addition of the qualcomm driver:
error: the following would cause module name conflict:
arch/x86/platform/iris/iris.ko
drivers/media/platform/qcom/iris/iris.ko
Since this a new module, nothing should rely on the name yet, so
rename this one to a more specific "qcom-iris.ko".
Fixes: 38506cb7e8 ("media: iris: add platform driver for iris video device")
Cc: Shérab <Sebastien.Hinderer@ens-lyon.org> # for x86 iris
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
This is one of three clang warnings about incompatible enum types
in a conditional expression:
drivers/media/platform/mediatek/vcodec/encoder/venc/venc_h264_if.c:597:29: error: conditional expression between different enumeration types ('enum scp_ipi_id' and 'enum ipi_id') [-Werror,-Wenum-compare-conditional]
597 | inst->vpu_inst.id = is_ext ? SCP_IPI_VENC_H264 : IPI_VENC_H264;
| ^ ~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~
The code is correct, so just rework it to avoid the warning.
Fixes: 0dc4b32861 ("media: mtk-vcodec: venc: support SCP firmware")
Cc: stable@vger.kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Alexandre Courbot <acourbot@google.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Since 2002 (change "Replace BKL for chain locking with sysvfs-private
rwlock") the sysv filesystem was doing IO under a rwlock in its
get_block() function (yes, a non-sleepable lock hold over a function
used to read inode metadata for all reads and writes). Nobody noticed
until syzbot in 2023 [1]. This shows nobody is using the filesystem.
Just drop it.
[1] https://lore.kernel.org/all/0000000000000ccf9a05ee84f5b0@google.com/
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20250220163940.10155-2-jack@suse.cz
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Adding an unlikely() hint on task comparisons on an unlikely error
return path improves run-time performance of the kcmp system call.
Benchmarking on an i9-12900 shows an improvement of ~5.5% on kcmp().
Results based on running 20 tests with turbo disabled (to reduce
clock freq turbo changes), with 10 second run per test and comparing
the number of kcmp calls per second. The % Standard deviation of 20
tests was ~0.25%, results are reliable.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Link: https://lore.kernel.org/r/20250213163916.709392-1-colin.i.king@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
The former is a no-op wrapper with the same argument.
I left it in place to not lose the information who needs it -- one day
"pseudo" inodes may start differing from what alloc_inode() returns.
In the meantime no point taking a detour.
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/r/20250212180459.1022983-1-mjguzik@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
Before 2011 there was no meaningful synchronization between
read/readdir/write/seek. Only in commit
ef3d0fd27e ("vfs: do (nearly) lockless generic_file_llseek")
synchronization was added for SEEK_CUR by taking f_lock around
vfs_setpos().
Then in 2014 full synchronization between read/readdir/write/seek was
added in commit 9c225f2655 ("vfs: atomic f_pos accesses as per POSIX")
by introducing f_pos_lock for regular files with FMODE_ATOMIC_POS and
for directories. At that point taking f_lock became unnecessary for such
files.
So only acquire f_lock for SEEK_CUR if this isn't a file that would have
acquired f_pos_lock if necessary.
Link: https://lore.kernel.org/r/20250207-daten-mahlzeit-99d2079864fb@brauner
Signed-off-by: Christian Brauner <brauner@kernel.org>
It is merely a trivial wrapper around getname_flags which adds a zeroed
argument, no point paying for an extra call.
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/r/20250206000105.432528-1-mjguzik@gmail.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Dumping processes with large allocated and mostly not-faulted areas is
very slow.
Borrowing a test case from Tavian Barnes:
int main(void) {
char *mem = mmap(NULL, 1ULL << 40, PROT_READ | PROT_WRITE,
MAP_ANONYMOUS | MAP_NORESERVE | MAP_PRIVATE, -1, 0);
printf("%p %m\n", mem);
if (mem != MAP_FAILED) {
mem[0] = 1;
}
abort();
}
That's 1TB of almost completely not-populated area.
On my test box it takes 13-14 seconds to dump.
The profile shows:
- 99.89% 0.00% a.out
entry_SYSCALL_64_after_hwframe
do_syscall_64
syscall_exit_to_user_mode
arch_do_signal_or_restart
- get_signal
- 99.89% do_coredump
- 99.88% elf_core_dump
- dump_user_range
- 98.12% get_dump_page
- 64.19% __get_user_pages
- 40.92% gup_vma_lookup
- find_vma
- mt_find
4.21% __rcu_read_lock
1.33% __rcu_read_unlock
- 3.14% check_vma_flags
0.68% vma_is_secretmem
0.61% __cond_resched
0.60% vma_pgtable_walk_end
0.59% vma_pgtable_walk_begin
0.58% no_page_table
- 15.13% down_read_killable
0.69% __cond_resched
13.84% up_read
0.58% __cond_resched
Almost 29% of the time is spent relocking the mmap semaphore between
calls to get_dump_page() which find nothing.
Whacking that results in times of 10 seconds (down from 13-14).
While here make the thing killable.
The real problem is the page-sized iteration and the real fix would
patch it up instead. It is left as an exercise for the mm-familiar
reader.
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/r/20250119103205.2172432-1-mjguzik@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
Fix the return type of several functions from long to int to match its actu
al behavior. These functions only return int values. This change improves
type consistency across the filesystem code and aligns the function signatu
re with its existing implementation and usage.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Yuichiro Tsuji <yuichtsu@amazon.com>
Link: https://lore.kernel.org/r/20250121070844.4413-3-yuichtsu@amazon.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
Fix the return type of several functions from long to int to match its actu
al behavior. These functions only return int values. This change improves
type consistency across the filesystem code and aligns the function signatu
re with its existing implementation and usage.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Yuichiro Tsuji <yuichtsu@amazon.com>
Link: https://lore.kernel.org/r/20250121070844.4413-2-yuichtsu@amazon.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
["fallen through the cracks" misc stuff]
A bunch of anon_inode_getfile() callers follow it with adjusting
->f_mode; we have a helper doing that now, so let's make use
of it.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Link: https://lore.kernel.org/r/20250118014434.GT1977892@ZenIV
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Mateusz Guzik <mjguzik@gmail.com> says:
This adds a super basic version just to get the mechanism going, along
with sample usage.
The macro set is incomplete (e.g., lack of locking macros) and
dump_inode routine fails to dump any state yet, to be implemented(tm).
I think despite the primitive state this is complete enough to start
sprinkling asserts as necessary.
* patches from https://lore.kernel.org/r/20250209185523.745956-1-mjguzik@gmail.com:
vfs: use the new debug macros in inode_set_cached_link()
vfs: catch invalid modes in may_open()
vfs: add initial support for CONFIG_DEBUG_VFS
Link: https://lore.kernel.org/r/20250209185523.745956-1-mjguzik@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
Replace the deprecated "simple-audio-card,hp-det-gpio" property by
"simple-audio-card,hp-det-gpios" in Simple Audio Card device nodes.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>