Douglas Anderson [Fri, 3 May 2024 21:33:28 +0000 (14:33 -0700)]
drm/panel: sony-acx565akm: Don't call disable at remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
A grep through mainline for compatible strings used by this driver
indicates that it is used by TI OMAP boards. The TI OMAP driver
appears to be correctly calling drm_atomic_helper_shutdown() so we can
remove the calls.
Douglas Anderson [Fri, 3 May 2024 21:33:27 +0000 (14:33 -0700)]
drm/panel: sony-acx565akm: Don't double-check enabled state in disable
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
The acx565akm seems to do some unique stuff with the "enabled"
state. Specifically:
1. It seems to detect the enabled state based on how the bootloader
left the panel.
2. It uses the enabled state to prevent certain sysfs files from
accessing a disabled panel.
We'll leave the "enabled" state tracking for this. However, we can at
least get rid of the double-check when trying to disable.
Douglas Anderson [Fri, 3 May 2024 21:33:24 +0000 (14:33 -0700)]
drm/panel: sitronix-st7703: Don't call disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
The compatible strings used by this driver seem to show up across
boards using a variety of DRM drivers. It appears that the relevant
drivers have been converted, but at least one compatible string
doesn't seem to be found in any mainline dts files so we can't be 100%
sure. If it is found that the DRM modeset driver hasn't been fixed
then this patch could be temporarily reverted until it is.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
One thing to note for st7703 is that it has a special "allpixelson"
debugfs file. When this file is written the driver hacks a
disable/unprepare and then a prepare/enable to try to reset the
panel. Potentially that might have been relying on the old booleans we
removed. It'll still "work" because of the checks in the core but it
deserves a comment. This debugfs file didn't appear to be particularly
safe to use even before this patch since it would cause a
disabled/unprepared panel to become prepared/enabled.
Douglas Anderson [Fri, 3 May 2024 21:33:12 +0000 (14:33 -0700)]
drm/panel: xinpeng-xpp055c272: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
A grep through mainline for compatible strings used by this driver
indicates that it is used by Rockchip boards. The Rockchip driver
appears to be correctly calling drm_atomic_helper_shutdown() so we can
remove the calls.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:33:08 +0000 (14:33 -0700)]
drm/panel: simple: Add a comment about unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
Unfortunately, it's very difficult to know exactly which DRM modeset
drivers are using panel-simple due to the sheer number of panels it
handles. For now, we'll leave the calls and just add a comment to keep
people from copying this code.
Douglas Anderson [Fri, 3 May 2024 21:33:07 +0000 (14:33 -0700)]
drm/panel: simple: Stop tracking prepared/enabled
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:33:06 +0000 (14:33 -0700)]
drm/panel: samsung-atna33xc20: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
A grep through mainline for compatible strings used by this driver
indicates that it is used by Qualcomm boards. The Qualcomm driver
appears to be correctly calling drm_atomic_helper_shutdown() so we can
remove the calls.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:33:00 +0000 (14:33 -0700)]
drm/panel: novatek-nt36672a: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
A grep through mainline for compatible strings used by this driver
indicates that it is used by Qualcomm boards. The Qualcomm driver
appears to be correctly calling drm_atomic_helper_shutdown() so we can
remove the calls.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:32:58 +0000 (14:32 -0700)]
drm/panel: ltk500hd1829: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
Unfortunately, grepping mainline for this panel's compatible string
shows no hits, so we can't be 100% sure if the DRM modeset driver used
with this panel has been fixed. If it is found that the DRM modeset
driver hasn't been fixed then this patch could be temporarily reverted
until it is.
Douglas Anderson [Fri, 3 May 2024 21:32:57 +0000 (14:32 -0700)]
drm/panel: ltk500hd1829: Stop tracking prepared
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:32:56 +0000 (14:32 -0700)]
drm/panel: ltk050h3146w: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
Unfortunately, grepping mainline for this panel's compatible string
shows no hits, so we can't be 100% sure if the DRM modeset driver used
with this panel has been fixed. If it is found that the DRM modeset
driver hasn't been fixed then this patch could be temporarily reverted
until it is.
Douglas Anderson [Fri, 3 May 2024 21:32:55 +0000 (14:32 -0700)]
drm/panel: ltk050h3146w: Stop tracking prepared
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:32:54 +0000 (14:32 -0700)]
drm/panel: kingdisplay-kd097d04: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
A grep through mainline for compatible strings used by this driver
indicates that it is used by Rockchip boards. The Rockchip driver
appear to be correctly calling drm_atomic_helper_shutdown() so we can
remove the calls.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:32:50 +0000 (14:32 -0700)]
drm/panel: innolux-p079zca: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
A grep through mainline for compatible strings used by this driver
indicates that it is used by Rockchip boards. The Rockchip driver
appears to be correctly calling drm_atomic_helper_shutdown() so we can
remove the calls.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:32:48 +0000 (14:32 -0700)]
drm/panel: edp: Add a comment about unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
Unfortunately, it's very difficult to know exactly which DRM modeset
drivers are using panel-edp due to the sheer number of panels it
handles. For now, we'll leave the calls and just add a comment to keep
people from copying this code.
Douglas Anderson [Fri, 3 May 2024 21:32:47 +0000 (14:32 -0700)]
drm/panel: edp: Stop tracking prepared/enabled
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Douglas Anderson [Fri, 3 May 2024 21:32:46 +0000 (14:32 -0700)]
drm/panel: boe-tv101wum-nl6: Don't call unprepare+disable at shutdown/remove
It's the responsibility of a correctly written DRM modeset driver to
call drm_atomic_helper_shutdown() at shutdown time and that should be
disabling / unpreparing the panel if needed. Panel drivers shouldn't
be calling these functions themselves.
A recent effort was made to fix as many DRM modeset drivers as
possible [1] [2] [3] and most drivers are fixed now.
A grep through mainline for compatible strings used by this driver
indicates that it is used by Mediatek and Qualcomm boards. Both of
those drivers appear to be correctly calling
drm_atomic_helper_shutdown() so we can remove the calls.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
As talked about in commit d2aacaf07395 ("drm/panel: Check for already
prepared/enabled in drm_panel"), we want to remove needless code from
panel drivers that was storing and double-checking the
prepared/enabled state. Even if someone was relying on the
double-check before, that double-check is now in the core and not
needed in individual drivers.
Ville Syrjälä [Mon, 8 Apr 2024 17:04:11 +0000 (20:04 +0300)]
drm/sti: Include linux/io.h for devm_ioremap()
Include linux/io.h for devm_ioremap().
When built on x86_64 w/ COMPILE_TEST=y:
../drivers/gpu/drm/sti/sti_dvo.c:531:21: error: implicit declaration of function ‘devm_ioremap’ [-Werror=implicit-function-declaration]
531 | dvo->regs = devm_ioremap(dev, res->start,
| ^~~~~~~~~~~~
../drivers/gpu/drm/sti/sti_dvo.c:531:19: error: assignment to ‘void *’ from ‘int’ makes pointer from integer without a cast [-Werror=int-conversion]
531 | dvo->regs = devm_ioremap(dev, res->start,
| ^
Maxime Ripard [Mon, 27 May 2024 13:58:15 +0000 (15:58 +0200)]
drm/vc4: tests: Remove vc4_dummy_plane structure
The vc4_dummy_plane structure was introduced as a mean to add
mock-specific fields.
However, we never really used it and it's still strictly equivalent to
vc4_plane (which is in the same situation vs drm_plane), so we can
simply remove the vc4_dummy_plane structure and make the mock code
cleaner.
There has been some discussions recently about the infoframes sent by
drivers and if they were properly generated.
In parallel, there's been some interest in creating an infoframe-decode
tool similar to edid-decode.
Both would be much easier if we were to expose the infoframes programmed
in the hardware. It won't be perfect since we have no guarantee that
it's actually what goes through the wire, but it's the best we can do.
Maxime Ripard [Mon, 27 May 2024 13:58:11 +0000 (15:58 +0200)]
drm/connector: hdmi: Add Infoframes generation
Infoframes in KMS is usually handled by a bunch of low-level helpers
that require quite some boilerplate for drivers. This leads to
discrepancies with how drivers generate them, and which are actually
sent.
Now that we have everything needed to generate them in the HDMI
connector state, we can generate them in our common logic so that
drivers can simply reuse what we precomputed.
Maxime Ripard [Mon, 27 May 2024 13:58:10 +0000 (15:58 +0200)]
drm/tests: Add RGB Quantization tests
The previous commit added the infrastructure to the connector state to
track what RGB Quantization should be used in a given state for an HDMI
connector.
Let's add some kunit tests to make sure it works as expected.
Maxime Ripard [Mon, 27 May 2024 13:58:09 +0000 (15:58 +0200)]
drm/connector: hdmi: Add RGB Quantization Range to the connector state
HDMI controller drivers will need to figure out the RGB range they need
to configure based on a mode and property values. Let's expose that in
the HDMI connector state so drivers can just use that value.
Maxime Ripard [Mon, 27 May 2024 13:58:07 +0000 (15:58 +0200)]
drm/connector: hdmi: Add Broadcast RGB property
The i915 driver has a property to force the RGB range of an HDMI output.
The vc4 driver then implemented the same property with the same
semantics. KWin has support for it, and a PR for mutter is also there to
support it.
Both drivers implementing the same property with the same semantics,
plus the userspace having support for it, is proof enough that it's
pretty much a de-facto standard now and we can provide helpers for it.
Let's plumb it into the newly created HDMI connector.
Maxime Ripard [Mon, 27 May 2024 13:58:06 +0000 (15:58 +0200)]
drm/doc: Remove unused Broadcast RGB Property
The Broadcast RGB property has been documented as three separate entries
so far, each with a different set of values.
The first one is pretty much the generic one: it's used by i915 and vc4,
and is the one used by all the compositors.
The second one is used by the gma500 driver, and is a subset of the
first one: it can have the values "Full" or "Limited 16:235", but lack
the "Automatic" value.
The third one however isn't used by any driver and documents the values
"off", "auto" and "on".
It's unclear where the last one comes from. It was first documented in
commit 6c6a3996f2c5 ("Documentation: drm: describing drm properties
exposed by various drivers") which adds a number of properties used by
drivers, but without mentioning which driver was using what property.
Grepping at the 6c6a3996f2c5 commit however, it looks like no driver is
actually using it, and a quick look at the entire kernel history doesn't
show any match either.
At the time though, gma500 had an "audio" property used right next to
Broadcast RGB that did have the "off", "auto" and "on" values in the
same order.
As such, it was probably a copy/paste or scripting error back then, and
there's never been such property used in the kernel.
Either way, it certainly hasn't been used in a decade or two so we can
just get rid of it.
Maxime Ripard [Mon, 27 May 2024 13:58:04 +0000 (15:58 +0200)]
drm/connector: hdmi: Compute bpc and format automatically
Now that we have all the infrastructure needed, we can add some code
that will, for a given connector state and mode, compute the best output
format and bpc.
The algorithm is equivalent to the one already found in i915 and vc4.
The previous patch adds a new hook for HDMI connectors to filter out
configurations based on the TMDS character rate. Let's add some tests to
make sure it works as expected.
Maxime Ripard [Mon, 27 May 2024 13:58:02 +0000 (15:58 +0200)]
drm/connector: hdmi: Add custom hook to filter TMDS character rate
Most of the HDMI controllers have an upper TMDS character rate limit
they can't exceed. On "embedded"-grade display controllers, it will
typically be lower than what high-grade monitors can provide these days,
so drivers will filter the TMDS character rate based on the controller
capabilities.
To make that easier to handle for drivers, let's provide an optional
hook to be implemented by drivers so they can tell the HDMI controller
helpers if a given TMDS character rate is reachable for them or not.
This will then be useful to figure out the best format and bpc count for
a given mode.
Maxime Ripard [Mon, 27 May 2024 13:58:01 +0000 (15:58 +0200)]
drm/tests: Add TDMS character rate connector state tests
The previous patch stores in the connector state the expected TMDS
character rate matching the configuration of the HDMI connector. Let's
add a few tests to make sure it works as expected.
Maxime Ripard [Mon, 27 May 2024 13:58:00 +0000 (15:58 +0200)]
drm/connector: hdmi: Calculate TMDS character rate
Most HDMI drivers have some code to calculate the TMDS character rate,
usually to adjust an internal clock to match what the mode requires.
Since the TMDS character rates mostly depends on the resolution, whether
we need to repeat pixels or not, the bpc count and the format, we can
now derive it from the HDMI connector state that stores all those infos
and remove the duplication from drivers.
Maxime Ripard [Mon, 27 May 2024 13:57:58 +0000 (15:57 +0200)]
drm/display: hdmi: Add HDMI compute clock helper
A lot of HDMI drivers have some variation of the formula to calculate
the TMDS character rate from a mode, but few of them actually take all
parameters into account.
Let's create a helper to provide that rate taking all parameters into
account.
Maxime Ripard [Mon, 27 May 2024 13:57:50 +0000 (15:57 +0200)]
drm/connector: Introduce an HDMI connector initialization function
A lot of the various HDMI drivers duplicate some logic that depends on
the HDMI spec itself and not really a particular hardware
implementation.
Output BPC or format selection, infoframe generation are good examples
of such areas.
This creates a lot of boilerplate, with a lot of variations, which makes
it hard for userspace to rely on, and makes it difficult to get it right
for drivers.
In the next patches, we'll add a lot of infrastructure around the
drm_connector and drm_connector_state structures, which will allow to
abstract away the duplicated logic. This infrastructure comes with a few
requirements though, and thus we need a new initialization function.
Hopefully, this will make drivers simpler to handle, and their behaviour
more consistent.
Christophe JAILLET [Sat, 25 May 2024 16:00:31 +0000 (18:00 +0200)]
dma-buf/fence-array: Add flex array to struct dma_fence_array
This is an effort to get rid of all multiplications from allocation
functions in order to prevent integer overflows [1][2].
The "struct dma_fence_array" can be refactored to add a flex array in order
to have the "callback structures allocated behind the array" be more
explicit.
Do so:
- makes the code more readable and safer.
- allows using __counted_by() for additional checks
- avoids some pointer arithmetic in dma_fence_array_enable_signaling()
Kent Overstreet [Fri, 24 May 2024 15:42:09 +0000 (11:42 -0400)]
mm: percpu: Include smp.h in alloc_tag.h
percpu.h depends on smp.h, but doesn't include it directly because of
circular header dependency issues; percpu.h is needed in a bunch of low
level headers.
This fixes a randconfig build error on mips:
include/linux/alloc_tag.h: In function '__alloc_tag_ref_set':
include/asm-generic/percpu.h:31:40: error: implicit declaration of function 'raw_smp_processor_id' [-Werror=implicit-function-declaration]
Linus Torvalds [Sun, 26 May 2024 16:54:26 +0000 (09:54 -0700)]
Merge tag 'perf-tools-fixes-for-v6.10-1-2024-05-26' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools
Pull perf tool fix from Arnaldo Carvalho de Melo:
"Revert a patch causing a regression.
This made a simple 'perf record -e cycles:pp make -j199' stop working
on the Ampere ARM64 system Linus uses to test ARM64 kernels".
* tag 'perf-tools-fixes-for-v6.10-1-2024-05-26' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools:
Revert "perf parse-events: Prefer sysfs/JSON hardware events over legacy"
This made a simple 'perf record -e cycles:pp make -j199' stop working on
the Ampere ARM64 system Linus uses to test ARM64 kernels, as discussed
at length in the threads in the Link tags below.
The fix provided by Ian wasn't acceptable and work to fix this will take
time we don't have at this point, so lets revert this and work on it on
the next devel cycle.
Reported-by: Linus Torvalds <torvalds@linux-foundation.org> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Bhaskar Chowdhury <unixbhaskar@gmail.com> Cc: Ethan Adams <j.ethan.adams@gmail.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: James Clark <james.clark@arm.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kan Liang <kan.liang@linux.intel.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Thomas Richter <tmricht@linux.ibm.com> Cc: Tycho Andersen <tycho@tycho.pizza> Cc: Yang Jihong <yangjihong@bytedance.com> Link: https://lore.kernel.org/lkml/CAHk-=wi5Ri=yR2jBVk-4HzTzpoAWOgstr1LEvg_-OXtJvXXJOA@mail.gmail.com Link: https://lore.kernel.org/lkml/CAHk-=wiWvtFyedDNpoV7a8Fq_FpbB+F5KmWK2xPY3QoYseOf_A@mail.gmail.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Linus Torvalds [Sun, 26 May 2024 05:33:10 +0000 (22:33 -0700)]
Merge tag '6.10-rc-smb3-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client fixes from Steve French:
- two important netfs integration fixes - including for a data
corruption and also fixes for multiple xfstests
- reenable swap support over SMB3
* tag '6.10-rc-smb3-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
cifs: Fix missing set of remote_i_size
cifs: Fix smb3_insert_range() to move the zero_point
cifs: update internal version number
smb3: reenable swapfiles over SMB3 mounts
Linus Torvalds [Sat, 25 May 2024 22:10:33 +0000 (15:10 -0700)]
Merge tag 'mm-hotfixes-stable-2024-05-25-09-13' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Pull misc fixes from Andrew Morton:
"16 hotfixes, 11 of which are cc:stable.
A few nilfs2 fixes, the remainder are for MM: a couple of selftests
fixes, various singletons fixing various issues in various parts"
* tag 'mm-hotfixes-stable-2024-05-25-09-13' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
mm/ksm: fix possible UAF of stable_node
mm/memory-failure: fix handling of dissolved but not taken off from buddy pages
mm: /proc/pid/smaps_rollup: avoid skipping vma after getting mmap_lock again
nilfs2: fix potential hang in nilfs_detach_log_writer()
nilfs2: fix unexpected freezing of nilfs_segctor_sync()
nilfs2: fix use-after-free of timer for log writer thread
selftests/mm: fix build warnings on ppc64
arm64: patching: fix handling of execmem addresses
selftests/mm: compaction_test: fix bogus test success and reduce probability of OOM-killer invocation
selftests/mm: compaction_test: fix incorrect write of zero to nr_hugepages
selftests/mm: compaction_test: fix bogus test success on Aarch64
mailmap: update email address for Satya Priya
mm/huge_memory: don't unpoison huge_zero_folio
kasan, fortify: properly rename memintrinsics
lib: add version into /proc/allocinfo output
mm/vmalloc: fix vmalloc which may return null if called with __GFP_NOFAIL
Linus Torvalds [Sat, 25 May 2024 21:48:40 +0000 (14:48 -0700)]
Merge tag 'irq-urgent-2024-05-25' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq fixes from Ingo Molnar:
- Fix x86 IRQ vector leak caused by a CPU offlining race
- Fix build failure in the riscv-imsic irqchip driver
caused by an API-change semantic conflict
- Fix use-after-free in irq_find_at_or_after()
* tag 'irq-urgent-2024-05-25' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
genirq/irqdesc: Prevent use-after-free in irq_find_at_or_after()
genirq/cpuhotplug, x86/vector: Prevent vector leak during CPU offline
irqchip/riscv-imsic: Fixup riscv_ipi_set_virq_range() conflict
Linus Torvalds [Sat, 25 May 2024 21:40:09 +0000 (14:40 -0700)]
Merge tag 'x86-urgent-2024-05-25' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Ingo Molnar:
- Fix regressions of the new x86 CPU VFM (vendor/family/model)
enumeration/matching code
- Fix crash kernel detection on buggy firmware with
non-compliant ACPI MADT tables
- Address Kconfig warning
* tag 'x86-urgent-2024-05-25' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/cpu: Fix x86_match_cpu() to match just X86_VENDOR_INTEL
crypto: x86/aes-xts - switch to new Intel CPU model defines
x86/topology: Handle bogus ACPI tables correctly
x86/kconfig: Select ARCH_WANT_FRAME_POINTERS again when UNWINDER_FRAME_POINTER=y
Linus Torvalds [Sat, 25 May 2024 21:23:58 +0000 (14:23 -0700)]
Merge tag 'ceph-for-6.10-rc1' of https://github.com/ceph/ceph-client
Pull ceph updates from Ilya Dryomov:
"A series from Xiubo that adds support for additional access checks
based on MDS auth caps which were recently made available to clients.
This is needed to prevent scenarios where the MDS quietly discards
updates that a UID-restricted client previously (wrongfully) acked to
the user.
Other than that, just a documentation fixup"
* tag 'ceph-for-6.10-rc1' of https://github.com/ceph/ceph-client:
doc: ceph: update userspace command to get CephFS metadata
ceph: add CEPHFS_FEATURE_MDS_AUTH_CAPS_CHECK feature bit
ceph: check the cephx mds auth access for async dirop
ceph: check the cephx mds auth access for open
ceph: check the cephx mds auth access for setattr
ceph: add ceph_mds_check_access() helper
ceph: save cap_auths in MDS client when session is opened
Linus Torvalds [Sat, 25 May 2024 21:19:01 +0000 (14:19 -0700)]
Merge tag 'ntfs3_for_6.10' of https://github.com/Paragon-Software-Group/linux-ntfs3
Pull ntfs3 updates from Konstantin Komarov:
"Fixes:
- reusing of the file index (could cause the file to be trimmed)
- infinite dir enumeration
- taking DOS names into account during link counting
- le32_to_cpu conversion, 32 bit overflow, NULL check
- some code was refactored
Changes:
- removed max link count info display during driver init
Remove:
- atomic_open has been removed for lack of use"
* tag 'ntfs3_for_6.10' of https://github.com/Paragon-Software-Group/linux-ntfs3:
fs/ntfs3: Break dir enumeration if directory contents error
fs/ntfs3: Fix case when index is reused during tree transformation
fs/ntfs3: Mark volume as dirty if xattr is broken
fs/ntfs3: Always make file nonresident on fallocate call
fs/ntfs3: Redesign ntfs_create_inode to return error code instead of inode
fs/ntfs3: Use variable length array instead of fixed size
fs/ntfs3: Use 64 bit variable to avoid 32 bit overflow
fs/ntfs3: Check 'folio' pointer for NULL
fs/ntfs3: Missed le32_to_cpu conversion
fs/ntfs3: Remove max link count info display during driver init
fs/ntfs3: Taking DOS names into account during link counting
fs/ntfs3: remove atomic_open
fs/ntfs3: use kcalloc() instead of kzalloc()
Linus Torvalds [Sat, 25 May 2024 21:15:39 +0000 (14:15 -0700)]
Merge tag '6.10-rc-ksmbd-server-fixes' of git://git.samba.org/ksmbd
Pull smb server fixes from Steve French:
"Two ksmbd server fixes, both for stable"
* tag '6.10-rc-ksmbd-server-fixes' of git://git.samba.org/ksmbd:
ksmbd: ignore trailing slashes in share paths
ksmbd: avoid to send duplicate oplock break notifications
Linus Torvalds [Sat, 25 May 2024 20:33:53 +0000 (13:33 -0700)]
Merge tag 'rtc-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux
Pull RTC updates from Alexandre Belloni:
"There is one new driver and then most of the changes are the device
tree bindings conversions to yaml.
New driver:
- Epson RX8111
Drivers:
- Many Device Tree bindings conversions to dtschema
- pcf8563: wakeup-source support"
* tag 'rtc-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux:
pcf8563: add wakeup-source support
rtc: rx8111: handle VLOW flag
rtc: rx8111: demote warnings to debug level
rtc: rx6110: Constify struct regmap_config
dt-bindings: rtc: convert trivial devices into dtschema
dt-bindings: rtc: stmp3xxx-rtc: convert to dtschema
dt-bindings: rtc: pxa-rtc: convert to dtschema
rtc: Add driver for Epson RX8111
dt-bindings: rtc: Add Epson RX8111
rtc: mcp795: drop unneeded MODULE_ALIAS
rtc: nuvoton: Modify part number value
rtc: test: Split rtc unit test into slow and normal speed test
dt-bindings: rtc: nxp,lpc1788-rtc: convert to dtschema
dt-bindings: rtc: digicolor-rtc: move to trivial-rtc
dt-bindings: rtc: alphascale,asm9260-rtc: convert to dtschema
dt-bindings: rtc: armada-380-rtc: convert to dtschema
rtc: cros-ec: provide ID table for avoiding fallback match
Linus Torvalds [Sat, 25 May 2024 20:28:29 +0000 (13:28 -0700)]
Merge tag 'i3c/for-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/i3c/linux
Pull i3c updates from Alexandre Belloni:
"Runtime PM (power management) is improved and hot-join support has
been added to the dw controller driver.
Core:
- Allow device driver to trigger controller runtime PM
Drivers:
- dw: hot-join support
- svc: better IBI handling"
* tag 'i3c/for-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/i3c/linux:
i3c: dw: Add hot-join support.
i3c: master: Enable runtime PM for master controller
i3c: master: svc: fix invalidate IBI type and miss call client IBI handler
i3c: master: svc: change ENXIO to EAGAIN when IBI occurs during start frame
i3c: Add comment for -EAGAIN in i3c_device_do_priv_xfers()
Linus Torvalds [Sat, 25 May 2024 20:23:42 +0000 (13:23 -0700)]
Merge tag 'jffs2-for-linus-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs
Pull jffs2 updates from Richard Weinberger:
- Fix illegal memory access in jffs2_free_inode()
- Kernel-doc fixes
- print symbolic error names
* tag 'jffs2-for-linus-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
jffs2: Fix potential illegal address access in jffs2_free_inode
jffs2: Simplify the allocation of slab caches
jffs2: nodemgmt: fix kernel-doc comments
jffs2: print symbolic error name instead of error code
Linus Torvalds [Sat, 25 May 2024 20:17:48 +0000 (13:17 -0700)]
Merge tag 'uml-for-linus-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/uml/linux
Pull UML updates from Richard Weinberger:
- Fixes for -Wmissing-prototypes warnings and further cleanup
- Remove callback returning void from rtc and virtio drivers
- Fix bash location
* tag 'uml-for-linus-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/uml/linux: (26 commits)
um: virtio_uml: Convert to platform remove callback returning void
um: rtc: Convert to platform remove callback returning void
um: Remove unused do_get_thread_area function
um: Fix -Wmissing-prototypes warnings for __vdso_*
um: Add an internal header shared among the user code
um: Fix the declaration of kasan_map_memory
um: Fix the -Wmissing-prototypes warning for get_thread_reg
um: Fix the -Wmissing-prototypes warning for __switch_mm
um: Fix -Wmissing-prototypes warnings for (rt_)sigreturn
um: Stop tracking host PID in cpu_tasks
um: process: remove unused 'n' variable
um: vector: remove unused len variable/calculation
um: vector: fix bpfflash parameter evaluation
um: slirp: remove set but unused variable 'pid'
um: signal: move pid variable where needed
um: Makefile: use bash from the environment
um: Add winch to winch_handlers before registering winch IRQ
um: Fix -Wmissing-prototypes warnings for __warp_* and foo
um: Fix -Wmissing-prototypes warnings for text_poke*
um: Move declarations to proper headers
...
Linus Torvalds [Sat, 25 May 2024 00:28:02 +0000 (17:28 -0700)]
Merge tag 'drm-next-2024-05-25' of https://gitlab.freedesktop.org/drm/kernel
Pull drm fixes from Dave Airlie:
"Some fixes for the end of the merge window, mostly amdgpu and panthor,
with one nouveau uAPI change that fixes a bad decision we made a few
months back.
nouveau:
- fix bo metadata uAPI for vm bind
panthor:
- Fixes for panthor's heap logical block.
- Reset on unrecoverable fault
- Fix VM references.
- Reset fix.
xlnx:
- xlnx compile and doc fixes.
amdgpu:
- Handle vbios table integrated info v2.3
amdkfd:
- Handle duplicate BOs in reserve_bo_and_cond_vms
- Handle memory limitations on small APUs
dp/mst:
- MST null deref fix.
bridge:
- Don't let next bridge create connector in adv7511 to make probe
work"
* tag 'drm-next-2024-05-25' of https://gitlab.freedesktop.org/drm/kernel:
drm/amdgpu/atomfirmware: add intergrated info v2.3 table
drm/mst: Fix NULL pointer dereference at drm_dp_add_payload_part2
drm/amdkfd: Let VRAM allocations go to GTT domain on small APUs
drm/amdkfd: handle duplicate BOs in reserve_bo_and_cond_vms
drm/bridge: adv7511: Attach next bridge without creating connector
drm/buddy: Fix the warn on's during force merge
drm/nouveau: use tile_mode and pte_kind for VM_BIND bo allocations
drm/panthor: Call panthor_sched_post_reset() even if the reset failed
drm/panthor: Reset the FW VM to NULL on unplug
drm/panthor: Keep a ref to the VM at the panthor_kernel_bo level
drm/panthor: Force an immediate reset on unrecoverable faults
drm/panthor: Document drm_panthor_tiler_heap_destroy::handle validity constraints
drm/panthor: Fix an off-by-one in the heap context retrieval logic
drm/panthor: Relax the constraints on the tiler chunk size
drm/panthor: Make sure the tiler initial/max chunks are consistent
drm/panthor: Fix tiler OOM handling to allow incremental rendering
drm: xlnx: zynqmp_dpsub: Fix compilation error
drm: xlnx: zynqmp_dpsub: Fix few function comments
David Howells [Fri, 24 May 2024 14:23:36 +0000 (15:23 +0100)]
cifs: Fix missing set of remote_i_size
Occasionally, the generic/001 xfstest will fail indicating corruption in
one of the copy chains when run on cifs against a server that supports
FSCTL_DUPLICATE_EXTENTS_TO_FILE (eg. Samba with a share on btrfs). The
problem is that the remote_i_size value isn't updated by cifs_setsize()
when called by smb2_duplicate_extents(), but i_size *is*.
This may cause cifs_remap_file_range() to then skip the bit after calling
->duplicate_extents() that sets sizes.
Fix this by calling netfs_resize_file() in smb2_duplicate_extents() before
calling cifs_setsize() to set i_size.
This means we don't then need to call netfs_resize_file() upon return from
->duplicate_extents(), but we also fix the test to compare against the pre-dup
inode size.
[Note that this goes back before the addition of remote_i_size with the
netfs_inode struct. It should probably have been setting cifsi->server_eof
previously.]
Fixes: cfc63fc8126a ("smb3: fix cached file size problems in duplicate extents (reflink)") Signed-off-by: David Howells <dhowells@redhat.com>
cc: Steve French <sfrench@samba.org>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Shyam Prasad N <nspmangalore@gmail.com>
cc: Rohith Surabattula <rohiths.msft@gmail.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev Signed-off-by: Steve French <stfrench@microsoft.com>
David Howells [Wed, 22 May 2024 08:38:48 +0000 (09:38 +0100)]
cifs: Fix smb3_insert_range() to move the zero_point
Fix smb3_insert_range() to move the zero_point over to the new EOF.
Without this, generic/147 fails as reads of data beyond the old EOF point
return zeroes.
Fixes: 3ee1a1fc3981 ("cifs: Cut over to using netfslib") Signed-off-by: David Howells <dhowells@redhat.com>
cc: Shyam Prasad N <nspmangalore@gmail.com>
cc: Rohith Surabattula <rohiths.msft@gmail.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev Signed-off-by: Steve French <stfrench@microsoft.com>
Chengming Zhou [Mon, 13 May 2024 03:07:56 +0000 (11:07 +0800)]
mm/ksm: fix possible UAF of stable_node
The commit 2c653d0ee2ae ("ksm: introduce ksm_max_page_sharing per page
deduplication limit") introduced a possible failure case in the
stable_tree_insert(), where we may free the new allocated stable_node_dup
if we fail to prepare the missing chain node.
Then that kfolio return and unlock with a freed stable_node set... And
any MM activities can come in to access kfolio->mapping, so UAF.
Fix it by moving folio_set_stable_node() to the end after stable_node
is inserted successfully.
Link: https://lkml.kernel.org/r/20240513-b4-ksm-stable-node-uaf-v1-1-f687de76f452@linux.dev Fixes: 2c653d0ee2ae ("ksm: introduce ksm_max_page_sharing per page deduplication limit") Signed-off-by: Chengming Zhou <chengming.zhou@linux.dev> Acked-by: David Hildenbrand <david@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Stefan Roesch <shr@devkernel.io> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
memory_failure
try_memory_failure_hugetlb
me_huge_page
__page_handle_poison
dissolve_free_hugetlb_folio
drain_all_pages -- Buddy page can be isolated e.g. for compaction.
take_page_off_buddy -- Failed as page is not in the buddy list.
-- Page can be putback into buddy after compaction.
page_ref_inc -- Leads to buddy page with refcnt = 1.
Then unpoison_memory() can unpoison the page and send the buddy page back
into buddy list again leading to the above bad page state warning. And
bad_page() will call page_mapcount_reset() to remove PageBuddy from buddy
page leading to later VM_BUG_ON_PAGE(!PageBuddy(page)) when trying to
allocate this page.
Fix this issue by only treating __page_handle_poison() as successful when
it returns 1.
Link: https://lkml.kernel.org/r/20240523071217.1696196-1-linmiaohe@huawei.com Fixes: ceaf8fbea79a ("mm, hwpoison: skip raw hwpoison page in freeing 1GB hugepage") Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> Cc: Naoya Horiguchi <nao.horiguchi@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Yuanyuan Zhong [Thu, 23 May 2024 18:35:31 +0000 (12:35 -0600)]
mm: /proc/pid/smaps_rollup: avoid skipping vma after getting mmap_lock again
After switching smaps_rollup to use VMA iterator, searching for next entry
is part of the condition expression of the do-while loop. So the current
VMA needs to be addressed before the continue statement.
Otherwise, with some VMAs skipped, userspace observed memory
consumption from /proc/pid/smaps_rollup will be smaller than the sum of
the corresponding fields from /proc/pid/smaps.
Link: https://lkml.kernel.org/r/20240523183531.2535436-1-yzhong@purestorage.com Fixes: c4c84f06285e ("fs/proc/task_mmu: stop using linked list and highest_vm_end") Signed-off-by: Yuanyuan Zhong <yzhong@purestorage.com> Reviewed-by: Mohamed Khalfella <mkhalfella@purestorage.com> Cc: David Hildenbrand <david@redhat.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ryusuke Konishi [Mon, 20 May 2024 13:26:21 +0000 (22:26 +0900)]
nilfs2: fix potential hang in nilfs_detach_log_writer()
Syzbot has reported a potential hang in nilfs_detach_log_writer() called
during nilfs2 unmount.
Analysis revealed that this is because nilfs_segctor_sync(), which
synchronizes with the log writer thread, can be called after
nilfs_segctor_destroy() terminates that thread, as shown in the call trace
below:
Fix this issue by changing nilfs_segctor_sync() so that the log writer
thread returns normally without synchronizing after it terminates, and by
forcing tasks that are already waiting to complete once after the thread
terminates.
The skipped inode metadata flushout will then be processed together in the
subsequent cleanup work in nilfs_segctor_destroy().
Ryusuke Konishi [Mon, 20 May 2024 13:26:20 +0000 (22:26 +0900)]
nilfs2: fix unexpected freezing of nilfs_segctor_sync()
A potential and reproducible race issue has been identified where
nilfs_segctor_sync() would block even after the log writer thread writes a
checkpoint, unless there is an interrupt or other trigger to resume log
writing.
This turned out to be because, depending on the execution timing of the
log writer thread running in parallel, the log writer thread may skip
responding to nilfs_segctor_sync(), which causes a call to schedule()
waiting for completion within nilfs_segctor_sync() to lose the opportunity
to wake up.
The reason why waking up the task waiting in nilfs_segctor_sync() may be
skipped is that updating the request generation issued using a shared
sequence counter and adding an wait queue entry to the request wait queue
to the log writer, are not done atomically. There is a possibility that
log writing and request completion notification by nilfs_segctor_wakeup()
may occur between the two operations, and in that case, the wait queue
entry is not yet visible to nilfs_segctor_wakeup() and the wake-up of
nilfs_segctor_sync() will be carried over until the next request occurs.
Fix this issue by performing these two operations simultaneously within
the lock section of sc_state_lock. Also, following the memory barrier
guidelines for event waiting loops, move the call to set_current_state()
in the same location into the event waiting loop to ensure that a memory
barrier is inserted just before the event condition determination.
Ryusuke Konishi [Mon, 20 May 2024 13:26:19 +0000 (22:26 +0900)]
nilfs2: fix use-after-free of timer for log writer thread
Patch series "nilfs2: fix log writer related issues".
This bug fix series covers three nilfs2 log writer-related issues,
including a timer use-after-free issue and potential deadlock issue on
unmount, and a potential freeze issue in event synchronization found
during their analysis. Details are described in each commit log.
This patch (of 3):
A use-after-free issue has been reported regarding the timer sc_timer on
the nilfs_sc_info structure.
The problem is that even though it is used to wake up a sleeping log
writer thread, sc_timer is not shut down until the nilfs_sc_info structure
is about to be freed, and is used regardless of the thread's lifetime.
Fix this issue by limiting the use of sc_timer only while the log writer
thread is alive.
Link: https://lkml.kernel.org/r/20240520132621.4054-1-konishi.ryusuke@gmail.com Link: https://lkml.kernel.org/r/20240520132621.4054-2-konishi.ryusuke@gmail.com Fixes: fdce895ea5dd ("nilfs2: change sc_timer from a pointer to an embedded one in struct nilfs_sc_info") Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com> Reported-by: "Bai, Shuangpeng" <sjb7183@psu.edu> Closes: https://groups.google.com/g/syzkaller/c/MK_LYqtt8ko/m/8rgdWeseAwAJ Tested-by: Ryusuke Konishi <konishi.ryusuke@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Michael Ellerman [Tue, 21 May 2024 03:02:19 +0000 (13:02 +1000)]
selftests/mm: fix build warnings on ppc64
Fix warnings like:
In file included from uffd-unit-tests.c:8:
uffd-unit-tests.c: In function `uffd_poison_handle_fault':
uffd-common.h:45:33: warning: format `%llu' expects argument of type
`long long unsigned int', but argument 3 has type `__u64' {aka `long
unsigned int'} [-Wformat=]
By switching to unsigned long long for u64 for ppc64 builds.
The problem is because bpf_arch_text_copy() silently fails to write to the
read-only area as a result of patch_map() faulting and the resulting
-EFAULT being chucked away.
Update patch_map() to use CONFIG_EXECMEM instead of
CONFIG_STRICT_MODULE_RWX to check for vmalloc addresses.
Link: https://lkml.kernel.org/r/20240521213813.703309-1-rppt@kernel.org Fixes: 2c9e5d4a0082 ("bpf: remove CONFIG_BPF_JIT dependency on CONFIG_MODULES of") Signed-off-by: Will Deacon <will@kernel.org> Signed-off-by: Mike Rapoport (IBM) <rppt@kernel.org> Reported-by: Klara Modin <klarasmodin@gmail.com> Closes: https://lore.kernel.org/all/7983fbbf-0127-457c-9394-8d6e4299c685@gmail.com Tested-by: Klara Modin <klarasmodin@gmail.com> Cc: Björn Töpel <bjorn@kernel.org> Cc: Luis Chamberlain <mcgrof@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Dev Jain [Tue, 21 May 2024 07:43:58 +0000 (13:13 +0530)]
selftests/mm: compaction_test: fix bogus test success and reduce probability of OOM-killer invocation
Reset nr_hugepages to zero before the start of the test.
If a non-zero number of hugepages is already set before the start of the
test, the following problems arise:
- The probability of the test getting OOM-killed increases. Proof:
The test wants to run on 80% of available memory to prevent OOM-killing
(see original code comments). Let the value of mem_free at the start
of the test, when nr_hugepages = 0, be x. In the other case, when
nr_hugepages > 0, let the memory consumed by hugepages be y. In the
former case, the test operates on 0.8 * x of memory. In the latter,
the test operates on 0.8 * (x - y) of memory, with y already filled,
hence, memory consumed is y + 0.8 * (x - y) = 0.8 * x + 0.2 * y > 0.8 *
x. Q.E.D
- The probability of a bogus test success increases. Proof: Let the
memory consumed by hugepages be greater than 25% of x, with x and y
defined as above. The definition of compaction_index is c_index = (x -
y)/z where z is the memory consumed by hugepages after trying to
increase them again. In check_compaction(), we set the number of
hugepages to zero, and then increase them back; the probability that
they will be set back to consume at least y amount of memory again is
very high (since there is not much delay between the two attempts of
changing nr_hugepages). Hence, z >= y > (x/4) (by the 25% assumption).
Therefore, c_index = (x - y)/z <= (x - y)/y = x/y - 1 < 4 - 1 = 3
hence, c_index can always be forced to be less than 3, thereby the test
succeeding always. Q.E.D
Link: https://lkml.kernel.org/r/20240521074358.675031-4-dev.jain@arm.com Fixes: bd67d5c15cc1 ("Test compaction of mlocked memory") Signed-off-by: Dev Jain <dev.jain@arm.com> Cc: <stable@vger.kernel.org> Cc: Anshuman Khandual <anshuman.khandual@arm.com> Cc: Shuah Khan <shuah@kernel.org> Cc: Sri Jayaramappa <sjayaram@akamai.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Dev Jain [Tue, 21 May 2024 07:43:57 +0000 (13:13 +0530)]
selftests/mm: compaction_test: fix incorrect write of zero to nr_hugepages
Currently, the test tries to set nr_hugepages to zero, but that is not
actually done because the file offset is not reset after read(). Fix that
using lseek().
Link: https://lkml.kernel.org/r/20240521074358.675031-3-dev.jain@arm.com Fixes: bd67d5c15cc1 ("Test compaction of mlocked memory") Signed-off-by: Dev Jain <dev.jain@arm.com> Cc: <stable@vger.kernel.org> Cc: Anshuman Khandual <anshuman.khandual@arm.com> Cc: Shuah Khan <shuah@kernel.org> Cc: Sri Jayaramappa <sjayaram@akamai.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Dev Jain [Tue, 21 May 2024 07:43:56 +0000 (13:13 +0530)]
selftests/mm: compaction_test: fix bogus test success on Aarch64
Patch series "Fixes for compaction_test", v2.
The compaction_test memory selftest introduces fragmentation in memory
and then tries to allocate as many hugepages as possible. This series
addresses some problems.
On Aarch64, if nr_hugepages == 0, then the test trivially succeeds since
compaction_index becomes 0, which is less than 3, due to no division by
zero exception being raised. We fix that by checking for division by
zero.
Secondly, correctly set the number of hugepages to zero before trying
to set a large number of them.
Now, consider a situation in which, at the start of the test, a non-zero
number of hugepages have been already set (while running the entire
selftests/mm suite, or manually by the admin). The test operates on 80%
of memory to avoid OOM-killer invocation, and because some memory is
already blocked by hugepages, it would increase the chance of OOM-killing.
Also, since mem_free used in check_compaction() is the value before we
set nr_hugepages to zero, the chance that the compaction_index will
be small is very high if the preset nr_hugepages was high, leading to a
bogus test success.
This patch (of 3):
Currently, if at runtime we are not able to allocate a huge page, the test
will trivially pass on Aarch64 due to no exception being raised on
division by zero while computing compaction_index. Fix that by checking
for nr_hugepages == 0. Anyways, in general, avoid a division by zero by
exiting the program beforehand. While at it, fix a typo, and handle the
case where the number of hugepages may overflow an integer.
The root cause is that HWPoison flag will be set for huge_zero_folio
without increasing the folio refcnt. But then unpoison_memory() will
decrease the folio refcnt unexpectedly as it appears like a successfully
hwpoisoned folio leading to VM_BUG_ON_PAGE(page_ref_count(page) == 0) when
releasing huge_zero_folio.
Skip unpoisoning huge_zero_folio in unpoison_memory() to fix this issue.
We're not prepared to unpoison huge_zero_folio yet.
Link: https://lkml.kernel.org/r/20240516122608.22610-1-linmiaohe@huawei.com Fixes: 478d134e9506 ("mm/huge_memory: do not overkill when splitting huge_zero_page") Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> Acked-by: David Hildenbrand <david@redhat.com> Reviewed-by: Yang Shi <shy828301@gmail.com> Reviewed-by: Oscar Salvador <osalvador@suse.de> Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com> Cc: Naoya Horiguchi <nao.horiguchi@gmail.com> Cc: Xu Yu <xuyu@linux.alibaba.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
projects / users / jedix / linux-maple.git / log