Split ->populate() implementation from ->init() code.
This decoupling will help for the further changes.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://patch.msgid.link/20251031080540.3970776-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In the future changes we would like to change the flow of the cache handling.
Add ->populate() callback in order to prepare for that.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://patch.msgid.link/20251031080540.3970776-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Commit e0762fd26a ("rtc: cpcap: Fix initial enable_irq/disable_irq
balance") set 'alarm_enabled' prior to calling the function
devm_request_threaded_irq() because this enables the IRQ. However, right
after calling devm_request_threaded_irq(), the driver calls
disable_irq() to disable the IRQ and so now 'alarm_enabled' will be true
but the IRQ is actually disabled. Revert this commit to fix the
'alarm_enabled' state.
Fixes: e0762fd26a ("rtc: cpcap: Fix initial enable_irq/disable_irq balance")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://patch.msgid.link/20251031103741.945460-2-jonathanh@nvidia.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Commit 1502fe0e97 ("rtc: tps6586x: Fix initial enable_irq/disable_irq
balance") breaks the wake-up alarm for the tps6586x. After this commit
was added RTC wake ups from suspend stopped working on the Tegra20
Ventana platform.
The problem is that this change set the 'irq_en' variable to true prior
to calling devm_request_threaded_irq() to indicate that the IRQ is
enabled, however, it was over looked that the flag IRQ_NOAUTOEN is
already set meaning that the IRQ is not enabled by default. This
prevents the IRQ from being enabled as expected. Revert this change to
fix this.
Fixes: 1502fe0e97 ("rtc: tps6586x: Fix initial enable_irq/disable_irq balance")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://patch.msgid.link/20251031103741.945460-1-jonathanh@nvidia.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
There is a hw restriction that we could enable the FBC for FP16
formats only if the pixel normalization block is enabled. Hence
enable the pixel normalizer block with normalzation factor as
1.0 for the supported FP16 formats to get the FBC enabled. Two
existing helper function definitions are moved up to avoid the
forward declarations as part of this patch as well.
v2: sw/hw state differentiation on handling pixel normalizer (Jani)
Bspec: 69863, 68881
Cc: Shekhar Chauhan <shekhar.chauhan@intel.com>
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Link: https://patch.msgid.link/20251027134001.325064-5-vinod.govindapillai@intel.com
As the pixel normalizer block is needed only for the FP16 formats,
separating out the FP16 formats handling will benefit in selectively
enabling the pixel normalizer block in the following patch.
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Link: https://patch.msgid.link/20251027134001.325064-4-vinod.govindapillai@intel.com
Add supported FP16 formats for FBC. FBC can be enabled with
FP16 formats only when plane pixel normalizer block is enabled.
The follow up patches will handle the pixel normalizer block
configuration.
v2: changes related to comments to use better tracking of pixel
normalizer
Bspec: 6881, 69863, 68904
Cc: Shekhar Chauhan <shekhar.chauhan@intel.com>
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Link: https://patch.msgid.link/20251027134001.325064-3-vinod.govindapillai@intel.com
Starting Xe3p_LPD, FBC is supported on UINT16 formats as well. Also
UINT16 being a 64bpp format, will use cpp of 8 for cfb stride and thus
size calculations.
v2: simplify getting the cpp per format (Ville)
simplify the pixel format is valid for xe3p_lp (Vinod)
Cc: Shekhar Chauhan <shekhar.chauhan@intel.com>
BSpec: 68881, 68904, 69560
Signed-off-by: Sai Teja Pottumuttu <sai.teja.pottumuttu@intel.com>
Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Link: https://patch.msgid.link/20251027134001.325064-2-vinod.govindapillai@intel.com
Reduce the compat i915_utils.h to only cover the requirements of
MISSING_CASE() in soc/ and the few i915_inject_probe_failure()
instances, instead of including the entire i915_utils.h from i915. This
prevents new users for the utilities from cropping up.
With this, we can remove the xe/display/ext/i915_utils.c altogether.
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://patch.msgid.link/c02f82e45f31caf95bd1339080b9099c3b7190be.1761146196.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
The remaining utils display needs from i915_utils.h are primarily
MISSING_CASE() and fetch_and_zero(), with a couple of
i915_inject_probe_failure() uses.
To avoid excessive churn, add duplicates of MISSING_CASE() and
fetch_and_zero() to intel_display_utils.h, and switch display to use the
display utils.
As long as there are display files that include i915_drv.h, which
includes i915_utils.h, we'll need #ifndef guards for MISSING_CASE() and
fetch_and_zero() in both utils headers. We can remove them once display
no longer depends on i915_drv.h.
A couple of files in display still need i915_utils.h for
i915_inject_probe_failure(). Annotate this. They will be handled
separately.
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://patch.msgid.link/79f9e31ca64c8c045834d48e20ceb0c515d1e9e1.1761146196.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Add intel_display_utils.c for display utilities that need more than a
header.
Start off with intel_display_run_as_guest(). The implementation is
intentional duplication of the i915_utils.h i915_run_as_guest(), with
the idea that it's small enough to not matter.
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://patch.msgid.link/469f9c41e0c3e3099314a3cf1a7671bf36ec8ffd.1761146196.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Add i915_jiffies.h and intel_display_jiffies.h for jiffies timeout and
wait helpers, and use them separately from i915 and display. This helps
reduce the display dependency on i915_utils.h.
Long term, both msecs_to_jiffies_timeout() and
wait_remaining_ms_from_jiffies() really belong in core kernel headers,
but for now unblock display refactoring.
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://patch.msgid.link/d8bc62b3a81afa05c849dde9b0f633572eaf5611.1761146196.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Besides blocks being invalidated, there is another case when the original
mapping could have changed between querying the rmap for GC and calling
xfs_zoned_map_extent. Document it there as it took us quite some time
to figure out what is going on while developing the multiple-GC
protection fix.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hans Holmberg <hans.holmberg@wdc.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
When we are picking a zone for gc it might already be in the pipeline
which can lead to us moving the same data twice resulting in in write
amplification and a very unfortunate case where we keep on garbage
collecting the zone we just filled with migrated data stopping all
forward progress.
Fix this by introducing a count of on-going GC operations on a zone, and
skip any zone with ongoing GC when picking a new victim.
Fixes: 080d01c41 ("xfs: implement zoned garbage collection")
Signed-off-by: Hans Holmberg <hans.holmberg@wdc.com>
Co-developed-by: Hans Holmberg <hans.holmberg@wdc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Tested-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
This might help debug issues better than just debug logging the function
name.
v2: Debug log type as hex (Ville)
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patch.msgid.link/20251029084603.2254982-2-jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
We might be getting send packet sequence errors and never know. Log them
as errors. Also upgrade the not supported read commands to errors.
v2: Also error log -EOPNOTSUPP (Ville)
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patch.msgid.link/20251029084603.2254982-1-jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
An ftrace warning was reported in ftrace_init_ool_stub():
WARNING: arch/powerpc/kernel/trace/ftrace.c:234 at ftrace_init_ool_stub+0x188/0x3f4, CPU#0: swapper/0
The problem is that the linker script is placing .text.startup in .text
rather than in .init.text, due to an inadvertent match of the TEXT_MAIN
'.text.[0-9a-zA-Z_]*' pattern.
This bug existed for some configurations before, but is only now coming
to light due to the TEXT_MAIN macro unification in commit 1ba9f89794
("vmlinux.lds: Unify TEXT_MAIN, DATA_MAIN, and related macros").
The .text.startup section consists of constructors which are used by
KASAN, KCSAN, and GCOV. The constructors are only called during boot,
so .text.startup is supposed to match the INIT_TEXT pattern so it can be
placed in .init.text and freed after init. But since INIT_TEXT comes
*after* TEXT_MAIN in the linker script, TEXT_MAIN needs to manually
exclude .text.startup.
Update TEXT_MAIN to exclude .text.startup (and its .text.startup.*
variant from -ffunction-sections), along with .text.exit and
.text.exit.* which should match EXIT_TEXT.
Specifically, use a series of more specific glob patterns to match
generic .text.* sections (for -ffunction-sections) while explicitly
excluding .text.startup[.*] and .text.exit[.*].
Also update INIT_TEXT and EXIT_TEXT to explicitly match their
-ffunction-sections variants (.text.startup.* and .text.exit.*).
Fixes: 1ba9f89794 ("vmlinux.lds: Unify TEXT_MAIN, DATA_MAIN, and related macros")
Closes: https://lore.kernel.org/72469502-ca37-4287-90b9-a751cecc498c@linux.ibm.com
Reported-by: Venkat Rao Bagalkote <venkat88@linux.ibm.com>
Debugged-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Venkat Rao Bagalkote <venkat88@linux.ibm.com>
Link: https://patch.msgid.link/07f74b4e5c43872572b7def30f2eac45f28675d9.1761872421.git.jpoimboe@kernel.org
The Ilitek IL79900A display controller is similar to the ILI9882T and can
be supported within the existing `panel-ilitek-ili9882t.c` driver.
This patch extends the ILI9882T driver to handle IL79900A-based panels,
such as the Tianma TL121BVMS07-00. The IL79900A uses a similar command
sequence and initialization flow, with minor differences in power supply
configuration and timing.
Signed-off-by: Langyan Ye <yelangyan@huaqin.corp-partner.google.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20251031100447.253164-3-yelangyan@huaqin.corp-partner.google.com
Add device tree bindings for the Tianma TL121BVMS07-00 12.1-inch
MIPI-DSI TFT LCD panel. The panel is based on the Ilitek IL79900A
controller.
Signed-off-by: Langyan Ye <yelangyan@huaqin.corp-partner.google.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20251031100447.253164-2-yelangyan@huaqin.corp-partner.google.com
Describe reset controllers for VI, MISC, AP, DSP and AO subsystems. The
one for AO subsystem is marked as reserved, since it may be used by AON
firmware.
Reviewed-by: Drew Fustini <fustini@kernel.org>
Signed-off-by: Yao Zi <ziyao@disroot.org>
Signed-off-by: Drew Fustini <fustini@kernel.org>
drm_sched_run_job_work() contains a comment which explains that an
entity being NULL means that there is no more work to do. It can,
however, also mean that there is work, but the scheduler doesn't have
enough credits to process the jobs right now.
Provide this detail in the comment.
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251028134602.94125-3-phasta@kernel.org
seqprop_sequence() is not intended for use outside seqlock.h.
Although it is accessible as a macro, it is meant to be used only
internally within the header.
Replace it with its proper wrapper, raw_read_seqcount().
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Co-developed-by: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Christoph Lameter (Ampere) <cl@gentwo.org>
Reviewed-by: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
Link: https://lore.kernel.org/r/20251023132802.654061-1-andi.shyti@linux.intel.com
Replace the manual string copying and parsing logic with a call to
simple_strtoull() to simplify and improve qat_uclo_parse_num().
Ensure that the parsed number does not exceed UINT_MAX, and add an
approximate upper-bound check (no more than 19 digits) to guard against
overflow.
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
Acked-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Before calling acc_get_sgl(), the mem_block has already been
created. acc_get_sgl() will not return NULL or any other error.
so the return value check can be removed.
Signed-off-by: nieweiqiang <nieweiqiang@huawei.com>
Signed-off-by: Chenghai Huang <huangchenghai2@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The isolate_strategy_store function is not protected
by a lock. If sysfs operations and functions that depend
on the err_threshold variable,such as qm_hw_err_isolate(),
execute concurrently, the outcome will be unpredictable.
Therefore, concurrency protection should be added for
the err_threshold variable.
Signed-off-by: nieweiqiang <nieweiqiang@huawei.com>
Signed-off-by: Chenghai Huang <huangchenghai2@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The eqe and aeqe are device updated values that include the
valid bit and queue number. In the current process, there is no
memory barrier added, so it cannot be guaranteed that the valid
bit is read before other processes are executed. Since eqe and aeqe
are only 4 bytes and the device writes them to memory in a single
operation, saving the values of eqe and aeqe ensures that the valid
bit and queue number read by the CPU were written by the device
simultaneously.
Signed-off-by: nieweiqiang <nieweiqiang@huawei.com>
Signed-off-by: Chenghai Huang <huangchenghai2@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
On multiple occasions the qce device have shown up in devices_deferred,
without the explanation that this came from the failure to acquire the
DMA channels from the associated BAM.
Use dev_err_probe() to associate this context with the failure to faster
pinpoint the culprit when this happens in the future.
Signed-off-by: Bjorn Andersson <bjorn.andersson@oss.qualcomm.com>
Reviewed-by: Abel Vesa <abel.vesa@linaro.org>
Reviewed-by: David Heidelberg <david@ixit.cz>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
VMA pinning to GGTT is now commited asynchronously in CHV / BXT+VDT
environments to avoid lock inversion among reservation_ww and cpu_hotplug
locks, the latter acquired from stop_machine(). Then,
vma->resource->page_sizes_gtt the test uses as shift count may still be
not populated (equal 0) after i915_vma_pin() returns. Wait for VMA bind
completion in those cases to avoid shift-out-of-bounds kernel warnings and
the test case failing with -EBADSLT.
v2: Explain why VMA pinning is commited asynchronously on CHV/BXT+VTD
(Krzysztof),
- use more precise wording in commit description.
Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Reviewed-by: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Acked-by: Andi Shyti <andi.shyti@linux.intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://lore.kernel.org/r/20251023082925.351307-8-janusz.krzysztofik@linux.intel.com
When running on a Cherryview, or on a Broxton with VTD enabled, pinning of
a VMA to GGTT is now committed asynchronously to avoid lock inversion
among reservation_ww and cpu_hotplug locks, the latter acquired from
stop_machine(). That may defer further processing of resources that
depend on that VMA. As a consequence, a 10ms delay in a multithreaded
migrate test case may occur too short and still incomplete threads may be
interrupted, and the test case may fail with -ERESTARTSYS or -EINTR error
code returned by any of those threads.
Extend the delay to empirically determined 100ms on affected platforms.
v3: Add an in-line comment that explains why 100ms (Andi).
v2: Fix spelling (Sebastian, Krzysztof),
- explain why VMA pinning is commited asynchronously on CHV/BXT+VTD
(Krzysztof).
Cc: Andi Shyti <andi.shyti@kernel.org>
Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Reviewed-by: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Acked-by: Andi Shyti <andi.shyti@linux.intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://lore.kernel.org/r/20251023082925.351307-7-janusz.krzysztofik@linux.intel.com
Add the __counted_by() compiler attribute to the flexible array member
'data' to improve access bounds-checking via CONFIG_UBSAN_BOUNDS and
CONFIG_FORTIFY_SOURCE.
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The AMD Seattle CCP is behind an IOMMU and has 4 entries, so add
the 'iommus' property.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add support for XTS mode of operation for AES algorithm in the AES
Engine of the DTHEv2 hardware cryptographic engine.
Signed-off-by: T Pratham <t-pratham@ti.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch introduces infrastructure for allocating req objects on the
stack for AEADs. The additions mirror the existing sync skcipher APIs.
This can be used in cases where simple sync AEAD operations are being
done. So allocating the request on stack avoides possible out-of-memory
errors.
The struct crypto_sync_aead is a wrapper around crypto_aead and should
be used in its place when sync only requests will be done on the stack.
Correspondingly, the request should be allocated with
SYNC_AEAD_REQUEST_ON_STACK().
Similar to sync_skcipher APIs, the new sync_aead APIs are wrappers
around the regular aead APIs to facilitate sync only operations. The
following crypto APIs are added:
- struct crypto_sync_aead
- crypto_alloc_sync_aead()
- crypto_free_sync_aead()
- crypto_aync_aead_tfm()
- crypto_sync_aead_setkey()
- crypto_sync_aead_setauthsize()
- crypto_sync_aead_authsize()
- crypto_sync_aead_maxauthsize()
- crypto_sync_aead_ivsize()
- crypto_sync_aead_blocksize()
- crypto_sync_aead_get_flags()
- crypto_sync_aead_set_flags()
- crypto_sync_aead_clear_flags()
- crypto_sync_aead_reqtfm()
- aead_request_set_sync_tfm()
- SYNC_AEAD_REQUEST_ON_STACK()
Signed-off-by: T Pratham <t-pratham@ti.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Some graphics APIs differentiate between out-of-graphics-memory and
out-of-host-memory (system memory). Add a device init flag to have -ENOSPC
propagated from the resource managers instead of being converted to
-ENOMEM, to aid driver stacks in determining what error code to return or
whether corrective action can be taken at the driver level.
Co-developed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-7-tvrtko.ursulin@igalia.com
While pidfs dentries are never hashed and thus retain_dentry() will never
consider them for placing them on the LRU it isn't great to always have
to go and remember that. Raise DCACHE_DONTCACHE explicitly as a visual
marker that dentries aren't kept but freed immediately instead.
Link: https://patch.msgid.link/20251029-work-namespace-nstree-listns-v4-4-2e6f823ebdc0@kernel.org
Tested-by: syzbot@syzkaller.appspotmail.com
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
While nsfs dentries are never hashed and thus retain_dentry() will never
consider them for placing them on the LRU it isn't great to always have
to go and remember that. Raise DCACHE_DONTCACHE explicitly as a visual
marker that dentries aren't kept but freed immediately instead.
Link: https://patch.msgid.link/20251029-work-namespace-nstree-listns-v4-3-2e6f823ebdc0@kernel.org
Tested-by: syzbot@syzkaller.appspotmail.com
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Currently nsfs uses the default inode_generic_drop() fallback which
drops the inode when it's unlinked or when it's unhashed. Since nsfs
never hashes inodes that always amounts to dropping the inode.
But that's just annoying to have to reason through every time we look at
this code. Switch to inode_just_drop() which always drops the inode
explicitly. This also aligns the behavior with pidfs which does the
same.
Link: https://patch.msgid.link/20251029-work-namespace-nstree-listns-v4-2-2e6f823ebdc0@kernel.org
Tested-by: syzbot@syzkaller.appspotmail.com
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Let the TTM pool allocator know that we can afford for it to expend less
effort for satisfying contiguous allocations larger than 2MiB. The latter
is the maximum relevant PTE entry size and the driver and hardware are
happy to get larger blocks only opportunistically.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-6-tvrtko.ursulin@igalia.com
GPUs typically benefit from contiguous memory via reduced TLB pressure and
improved caching performance, where the maximum size of contiguous block
which adds a performance benefit is related to hardware design.
TTM pool allocator by default tries (hard) to allocate up to the system
MAX_PAGE_ORDER blocks. This varies by the CPU platform and can also be
configured via Kconfig.
If that limit was set to be higher than the GPU can make an extra use of,
lets allow the individual drivers to let TTM know over which allocation
order can the pool allocator afford to make a little bit less effort with.
We implement this by disabling direct reclaim for those allocations, which
reduces the allocation latency and lowers the demands on the page
allocator, in cases where expending this effort is not critical for the
GPU in question.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-5-tvrtko.ursulin@igalia.com
Multiple consecutive boolean function arguments are usually not very
readable.
Replace the ones in ttm_device_init() with flags with the additional
benefit of soon being able to pass in more data with just a one off
code base churning cost.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Sui Jingfeng <suijingfeng@loongson.cn>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Zack Rusin <zack.rusin@broadcom.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Zack Rusin <zack.rusin@broadcom.com>
Acked-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> # For xe
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-4-tvrtko.ursulin@igalia.com
[tursulin: fixup checkpatch while applying]
