Fix debug log when looking for a DisplayPort adapter pair of DP IN and
DP OUT. In case of no DP adapter available, log the type of the DP
adapter that is not available.
Signed-off-by: Gil Fine <gil.fine@linux.intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
There is no point disabling CL states if the router is unplugged so in
that case return early.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
It turns out there is no need to use the actual link rate when
reclaiming bandwidth for USB 3.x. The reason is that we use consumed
bandwidth which is coming from xHCI when releasing bandwidth (for
example for DisplayPort tunneling) and this can be anything between
1000 Mb/s to maximum, so when reclaiming we can just bump it up back to
maximum instead of actual link rate (which is always <= maximum).
This allows us to get rid of couple of unnecessary lines of code.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
atomic-helper:
- Relax checks for unregistered connectors
dma-buf:
- Work around race condition when retrieving fence timestamp
gem:
- Avoid OOB access in BO memory range
panel:
- boe-tv101wun-ml6: Fix flickering
simpledrm:
- Fix error output
vwmgfx:
- Fix size calculation in texture-state code
- Ref GEM BOs in surfaces
msm:
- PHY/link training reset fix
- msm8998 - correct highest bank bit
- skip video mode if timing engine disabled
- check irq_of_parse_and_map return code
- add new lines to some prints
- fail atomic check for max mdp clk test
amdgpu:
- Seemless boot fix
- Fix TTM BO resource check
- SI fix for doorbell handling
-----BEGIN PGP SIGNATURE-----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=+7xQ
-----END PGP SIGNATURE-----
Merge tag 'drm-fixes-2023-10-13' of git://anongit.freedesktop.org/drm/drm
Pull drm fixes from Dave Airlie:
"Weekly fixes, the core is msm and amdgpu with some scattered fixes
across vmwgfx, panel and the core stuff.
atomic-helper:
- Relax checks for unregistered connectors
dma-buf:
- Work around race condition when retrieving fence timestamp
gem:
- Avoid OOB access in BO memory range
panel:
- boe-tv101wun-ml6: Fix flickering
simpledrm:
- Fix error output
vwmgfx:
- Fix size calculation in texture-state code
- Ref GEM BOs in surfaces
msm:
- PHY/link training reset fix
- msm8998 - correct highest bank bit
- skip video mode if timing engine disabled
- check irq_of_parse_and_map return code
- add new lines to some prints
- fail atomic check for max mdp clk test
amdgpu:
- Seamless boot fix
- Fix TTM BO resource check
- SI fix for doorbell handling"
* tag 'drm-fixes-2023-10-13' of git://anongit.freedesktop.org/drm/drm:
drm/tiny: correctly print `struct resource *` on error
drm: Do not overrun array in drm_gem_get_pages()
drm/atomic-helper: relax unregistered connector check
drm/panel: boe-tv101wum-nl6: Completely pull GPW to VGL before TP term
drm/amdgpu: fix SI failure due to doorbells allocation
drm/amdgpu: add missing NULL check
drm/amd/display: Don't set dpms_off for seamless boot
drm/vmwgfx: Keep a gem reference to user bos in surfaces
drm/vmwgfx: fix typo of sizeof argument
drm/msm/dpu: fail dpu_plane_atomic_check() based on mdp clk limits
dma-buf: add dma_fence_timestamp helper
drm/msm/dp: Add newlines to debug printks
drm/msm/dpu: change _dpu_plane_calc_bw() to use u64 to avoid overflow
drm/msm/dsi: fix irq_of_parse_and_map() error checking
drm/msm/dsi: skip the wait for video mode done if not applicable
drm/msm/mdss: fix highest-bank-bit for msm8998
drm/msm/dp: do not reinitialize phy unless retry during link training
Adds the I2C alias ports to each NVMe drive via the
backplane card.
Besides that, it also adds the eeprom and temperature sensor
on the backplane card.
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20231005035525.19036-8-chanh@os.amperecomputing.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
Mt.Mitchell DVT and later hardware do not use adc1. It only uses
adc0 with channels 0, 1 and 2. This commit removes redundant ADC
configurations.
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20231005035525.19036-7-chanh@os.amperecomputing.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
Add the GPIOR5 as a gpio-hog with output high so that can
power the OCP card once the BMC booting.
Add the GPIOAC5 as a gpio-hog with output high to notice
the BMC state.
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20231005035525.19036-4-chanh@os.amperecomputing.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
Update GPIO line-name to follow naming convention specified at
github.com/openbmc/docs/blob/master/designs/device-tree-gpio-naming.md
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20231005035525.19036-2-chanh@os.amperecomputing.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
Add linux device tree entry related to the Minerva Chassis Management
Controller (CMC) specific devices connected to the Aspeed SoC (AST2600).
Signed-off-by: Yang Chen <yangchen.openbmc@gmail.com>
Link: https://lore.kernel.org/r/20230914125648.3966519-3-yangchen.openbmc@gmail.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
Document the new compatibles used on the Facebook Minerva Chassis
Management Controller (CMC).
Signed-off-by: Yang Chen <yangchen.openbmc@gmail.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20230914125648.3966519-2-yangchen.openbmc@gmail.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
Move the call to ata_dev_power_set_active() to transition a disk in
standby power mode to the active power mode from
ata_eh_revalidate_and_attach() before doing revalidation to the end of
ata_eh_recover(), after the link speed for the device is reconfigured
(if that was necessary). This is safer as this ensure that the VERIFY
command executed to spinup the disk is executed with the drive properly
reconfigured first.
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Niklas Cassel <niklas.cassel@wdc.com>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
Fix new smatch warnings:
fs/smb/server/smb2pdu.c:6131 smb2_read_pipe() error: double free of 'rpc_resp'
Fixes: e2b76ab8b5 ("ksmbd: add support for read compound")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Coverity Scan report the following one. This report is a false alarm.
Because fp is never NULL when rc is zero. This patch add null check for fp
in ksmbd_update_fstate to make alarm silence.
*** CID 1568583: Null pointer dereferences (FORWARD_NULL)
/fs/smb/server/smb2pdu.c: 3408 in smb2_open()
3402 path_put(&path);
3403 path_put(&parent_path);
3404 }
3405 ksmbd_revert_fsids(work);
3406 err_out1:
3407 if (!rc) {
>>> CID 1568583: Null pointer dereferences (FORWARD_NULL)
>>> Passing null pointer "fp" to "ksmbd_update_fstate", which dereferences it.
3408 ksmbd_update_fstate(&work->sess->file_table, fp, FP_INITED);
3409 rc = ksmbd_iov_pin_rsp(work, (void *)rsp, iov_len);
3410 }
3411 if (rc) {
3412 if (rc == -EINVAL)
3413 rsp->hdr.Status = STATUS_INVALID_PARAMETER;
Fixes: e2b76ab8b5 ("ksmbd: add support for read compound")
Reported-by: Coverity Scan <scan-admin@coverity.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
set_smb2_rsp_status() after __process_request() sets the wrong error
status. This patch resets all iov vectors and sets the error status
on clean one.
Fixes: e2b76ab8b5 ("ksmbd: add support for read compound")
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Cthon test fail with the following error.
check for proper open/unlink operation
nfsjunk files before unlink:
-rwxr-xr-x 1 root root 0 9월 25 11:03 ./nfs2y8Jm9
./nfs2y8Jm9 open; unlink ret = 0
nfsjunk files after unlink:
-rwxr-xr-x 1 root root 0 9월 25 11:03 ./nfs2y8Jm9
data compare ok
nfsjunk files after close:
ls: cannot access './nfs2y8Jm9': No such file or directory
special tests failed
Cthon expect to second unlink failure when file is already unlinked.
ksmbd can not allow to open file if flags of ksmbd inode is set with
S_DEL_ON_CLS flags.
Cc: stable@vger.kernel.org
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Fixes for v6.6-rc5
- fix to not reset the PHY everytime we start link training but only
do it if link training fails. Without this, the PLL unlocked
interrupt fires causing "Unexpected DP AUX IRQ 0x01000000 when
not busy" spam in the logs since last 2-3 cycles
- correct the highest bank bit to match downstream device tree for
msm8998
- skip the video mode wait if the timing engine is not enabled. This
was introduced after pre_enable flag for DSI video mode panels
where we would end up waiting for the video mode done interrupt
even before enabling timing engine causing error spam and long
bootup times.
- check the correct return code of irq_of_parse_and_map() in DSI code
- avoid overflow issues in the dpu bandwidth calculation . This was
exposed for high resolution displays and a critical fix to avoid
atomic_check failure
- minor fix to add new lines in DP print messages.
- Fix to fail atomic_check() if the resolution exceeds max mdp clk.
This leads to underflow otherwise if we try to allow that frame.
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rob Clark <robdclark@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/CAF6AEGv-HNxQ=VBtZ8geGzYJum9jtManEdbvhcjo_WWF_J9Ziw@mail.gmail.com
-----BEGIN PGP SIGNATURE-----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=sbyr
-----END PGP SIGNATURE-----
Merge tag 'block-6.6-2023-10-12' of git://git.kernel.dk/linux
Pull block fix from Jens Axboe:
"Just a single fix for a longstanding regression with using fallocate
on a block device"
* tag 'block-6.6-2023-10-12' of git://git.kernel.dk/linux:
block: Don't invalidate pagecache for invalid falloc modes
Only DT based probing is used for the Motorola CPCAP charger driver, so
drop the !CONFIG_OF parts and redundant of_match_device() call.
Signed-off-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20231009172923.2457844-20-robh@kernel.org
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Depend on the explicit SoC defines rather than generic
architectures like most of the rest of the HW drivers do.
This makes the drivers only available for the HW and for
compile testing.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
Cc: Sebastian Reichel <sre@kernel.org>
Cc: linux-pm@vger.kernel.org
Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://lore.kernel.org/r/20231009135833.17880-3-pbrobinson@gmail.com
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
The divider_ro_round_rate() function could potentially return -EINVAL on
error but the error handling doesn't work because "rate" is unsigned.
It should be a type long.
Fixes: 06ed0fc0fb ("clk: stm32: composite: Switch to determine_rate")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: https://lore.kernel.org/r/d9a78453-9b40-48c1-830e-00751ba3ecb8@kili.mountain
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
In the possible_parent_show function, ensure proper handling of the return
value from of_clk_get_parent_name to prevent potential issues arising from
a NULL return.
The current implementation invokes seq_puts directly on the result of
of_clk_get_parent_name without verifying the return value, which can lead
to kernel panic if the function returns NULL.
This patch addresses the concern by introducing a check on the return
value of of_clk_get_parent_name. If the return value is not NULL, the
function proceeds to call seq_puts, providing the returned value as
argument.
However, if of_clk_get_parent_name returns NULL, the function provides a
static string as argument, avoiding the panic.
Fixes: 1ccc0ddf04 ("clk: Use seq_puts() in possible_parent_show()")
Reported-by: Philip Daly <pdaly@redhat.com>
Signed-off-by: Alessandro Carminati (Red Hat) <alessandro.carminati@gmail.com>
Link: https://lore.kernel.org/r/20230921073217.572151-1-alessandro.carminati@gmail.com
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Commit 9607beb917 ("clk: socfpga: gate: Add a determine_rate hook")
added a determine_rate implementation set to the
clk_hw_determine_rate_no_reparent, but failed to account for the
internal divider that wasn't used before anywhere but in recalc_rate.
This led to inconsistencies between the clock rate stored in
clk_core->rate and the one returned by clk_round_rate() that leverages
determine_rate().
Since that driver seems to be widely used (and thus regression-prone)
and not supporting rate changes (since it's missing a .set_rate
implementation), we can just report the current divider programmed in
the clock but not try to change it in any way.
This should be good enough to fix the issues reported, and if someone
ever wants to allow the divider to change then it should be easy enough
using the clk-divider helpers.
Link: https://lore.kernel.org/linux-clk/20231005095927.12398-2-b.spranger@linutronix.de/
Fixes: 9607beb917 ("clk: socfpga: gate: Add a determine_rate hook")
Reported-by: Benedikt Spranger <b.spranger@linutronix.de>
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20231012083729.2148044-1-mripard@kernel.org
[sboyd@kernel.org: Fix hw -> hwclk]
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
- In cgroup1, the `tasks` file could have duplicate pids which can trigger a
warning in seq_file. Fix it by removing duplicate items after sorting.
- Comment update.
-----BEGIN PGP SIGNATURE-----
iIQEABYIACwWIQTfIjM1kS57o3GsC/uxYfJx3gVYGQUCZSh+2A4cdGpAa2VybmVs
Lm9yZwAKCRCxYfJx3gVYGfASAP9dgEe1Ay6jkJoCCGROnjPRDj2j7Cm9WWcHV79X
0Pr3zQEA/vFIpUzRZGbisrvnyXwNNLX12Hq/nwRX6DzN4UIkDgE=
=tPdI
-----END PGP SIGNATURE-----
Merge tag 'cgroup-for-6.6-rc5-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup
Pull cgroup fixes from Tejun Heo:
- In cgroup1, the `tasks` file could have duplicate pids which can
trigger a warning in seq_file. Fix it by removing duplicate items
after sorting
- Comment update
* tag 'cgroup-for-6.6-rc5-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: Fix incorrect css_set_rwsem reference in comment
cgroup: Remove duplicates in cgroup v1 tasks file
Add a priority property equal to gpio-restart to allow increasing the
priority of the gpio-poweroff handler.
Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Link: https://lore.kernel.org/r/20231006130428.11259-5-francesco@dolcini.it
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Add the priority property to the gpio-poweroff bindings description.
Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20231006130428.11259-4-francesco@dolcini.it
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Use the new sys-off handler API for gpio-poweroff. This allows us to
have more than one poweroff handler and prioritise them.
Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Link: https://lore.kernel.org/r/20231006130428.11259-3-francesco@dolcini.it
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Use a struct to store the module variables. This is required to later
move to notifier_blocks where we can have several instances.
Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Link: https://lore.kernel.org/r/20231006130428.11259-2-francesco@dolcini.it
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
* Fix access-after-free in pwq allocation error path.
* Implicitly ordered unbound workqueues should lose the implicit ordering if
an attribute change which isn't compatible with ordered operation is
requested. However, attribute changes requested through the sysfs
interface weren't doing that leaving no way to override the implicit
ordering through the sysfs interface. Fix it.
* Other doc and misc updates.
-----BEGIN PGP SIGNATURE-----
iIQEABYIACwWIQTfIjM1kS57o3GsC/uxYfJx3gVYGQUCZSh9vQ4cdGpAa2VybmVs
Lm9yZwAKCRCxYfJx3gVYGTG4AQCklH7aGqSbzGBPuV19gN6q+BPjkNNLTkEtOzW7
3t1gewEAuwGiGr5FwuxCuGhzTUm5dkELFFsKdzYk+Pt7B2M5Pg0=
=YQtE
-----END PGP SIGNATURE-----
Merge tag 'wq-for-6.6-rc5-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq
Pull workqueue fixes from Tejun Heo:
- Fix access-after-free in pwq allocation error path
- Implicitly ordered unbound workqueues should lose the implicit
ordering if an attribute change which isn't compatible with ordered
operation is requested. However, attribute changes requested through
the sysfs interface weren't doing that leaving no way to override the
implicit ordering through the sysfs interface. Fix it.
- Other doc and misc updates
* tag 'wq-for-6.6-rc5-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
workqueue: fix -Wformat-truncation in create_worker
workqueue: Override implicit ordered attribute in workqueue_apply_unbound_cpumask()
workqueue: Use the kmem_cache_free() instead of kfree() to release pwq
workqueue: doc: Fix function and sysfs path errors
workqueue: Fix UAF report by KASAN in pwq_release_workfn()
Cross-merge networking fixes after downstream PR.
No conflicts.
Adjacent changes:
kernel/bpf/verifier.c
829955981c ("bpf: Fix verifier log for async callback return values")
a923819fb2 ("bpf: Treat first argument as return value for bpf_throw")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This reverts commit 92e24e0e57.
While the patch itself is correct, it uncovered an issue with fallback
to PS/2 mode, where we were leaving psmouse->fast_reconnect handler set
to psmouse_smbus_reconnect(), which caused crashes.
While discussing various approaches to fix the issue it was noted that
this patch ass undesired delay in the "fast" resume path of PS/2 device,
and it would be better to actually use "reset_delay" option defined in
struct rmi_device_platform_data and have RMI code handle it for SMBus
transport as well. So this patch is being reverted to deal with crashes
and a better solution will be merged shortly.
Reported-by: Thorsten Leemhuis <linux@leemhuis.info>
Closes: https://lore.kernel.org/all/ca0109fa-c64b-43c1-a651-75b294d750a1@leemhuis.info/
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
In light of the recent discovery that the fractional divisor
approximation does not utilize the full available range for clocks that
are flagged CLK_FRAC_DIVIDER_ZERO_BASED [1], implement tests for the
edge cases of this clock type.
Signed-off-by: Frank Oltmanns <frank@oltmanns.dev>
Link: https://lore.kernel.org/lkml/20230529133433.56215-1-frank@oltmanns.dev [1]
Link: https://lore.kernel.org/r/20230617131041.18313-3-frank@oltmanns.dev
[sboyd@kernel.org: Rename suite and tests slightly, drop unused
includes, store parent rate to compare instead of repeating equation]
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Consider the CLK_FRAC_DIVIDER_ZERO_BASED flag when finding the best
approximation for m and n. By doing so, increase the range of valid
values for the numerator and denominator by 1.
Furthermore, export the approximation function so that users of this
function can be compiled as modules.
Cc: A.s. Dong <aisheng.dong@nxp.com>
Signed-off-by: Frank Oltmanns <frank@oltmanns.dev>
Link: https://lore.kernel.org/r/20230617131041.18313-2-frank@oltmanns.dev
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Commit df8fc4e934 ("kbuild: Enable -fstrict-flex-arrays=3") enabled
'-fstrict-flex-arrays=3' globally, but bindgen does not recognized this
compiler option, triggering the following build error:
error: unknown argument: '-fstrict-flex-arrays=3', err: true
[ Miguel: Commit df8fc4e934 ("kbuild: Enable -fstrict-flex-arrays=3")
did it so only conditionally (i.e. only if the C compiler supports
it). This explains what Andrea was seeing: he was compiling with a
modern enough GCC, which enables the option, but with an old enough
Clang. Andrea confirmed this was the case: he was using Clang 14 with
GCC 13; and that Clang 15 worked for him.
While it is possible to construct code (see mailing list for an
example I came up with) where this could break, it is fairly
contrived, and anyway GCC-built kernels with Rust enabled should
only be used for experimentation until we get support for
`rustc_codegen_gcc` and/or GCC Rust. So let's add this for the
time being in case it helps somebody. ]
Add '-fstrict-flex-arrays' to the list of cflags that should be ignored
by bindgen.
Fixes: df8fc4e934 ("kbuild: Enable -fstrict-flex-arrays=3")
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Tested-by: Gary Guo <gary@garyguo.net>
Link: https://lore.kernel.org/r/20230815065346.131387-1-andrea.righi@canonical.com
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
A mistake was made and the description of `ECHILD` is wrong (it reuses
the description of `ENOEXEC`). This fixes it to reflect what's in
`errno-base.h`.
Signed-off-by: Wedson Almeida Filho <walmeida@microsoft.com>
Reviewed-by: Martin Rodriguez Reboredo <yakoyoku@gmail.com>
Reviewed-by: Trevor Gross <tmgross@umich.edu>
Reviewed-by: Finn Behrens <me@kloenk.dev>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Fixes: 266def2a0f ("rust: error: add codes from `errno-base.h`")
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20230930144958.46051-1-wedsonaf@gmail.com
[ Use the plural, as noticed by Benno. ]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
If there is a burst of message the receive worker will filling up the
processing queue but where are too slow to process dlm messages. This
patch will slow down the receiver worker to keep the buffer on the
socket layer to tell the sender to backoff. This is done by a threshold
to get the next buffers from the socket after all messages were
processed done by a flush_workqueue(). This however only occurs when we
have a message burst when we e.g. create 1 million locks. If we put more
and more new messages to process in the processqueue we will soon run out
of memory.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
