can_split_folio() is just a refcount comparison, making sure only the
split caller holds an extra pin. Open code it with
folio_expected_ref_count() != folio_ref_count() - 1. For the extra_pins
used by folio_ref_freeze(), add folio_cache_ref_count() to calculate it.
Also replace folio_expected_ref_count() with folio_cache_ref_count() used
by folio_ref_unfreeze(), since they are returning the same values when a
folio is frozen and folio_cache_ref_count() does not have unnecessary
folio_mapcount() in its implementation.
Link: https://lkml.kernel.org/r/20251126210618.1971206-3-ziy@nvidia.com
Signed-off-by: Zi Yan <ziy@nvidia.com>
Suggested-by: David Hildenbrand (Red Hat) <david@kernel.org>
Reviewed-by: Wei Yang <richard.weiyang@gmail.com>
Acked-by: David Hildenbrand (Red Hat) <david@kernel.org>
Cc: Balbir Singh <balbirs@nvidia.com>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Barry Song <baohua@kernel.org>
Cc: Dev Jain <dev.jain@arm.com>
Cc: Lance Yang <lance.yang@linux.dev>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Naoya Horiguchi <nao.horiguchi@gmail.com>
Cc: Nico Pache <npache@redhat.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "Improve folio split related functions", v4.
This patchset improves several folio split related functions to avoid
future misuse. The changes are:
1. Consolidated folio splittable checks by moving truncated folio check,
huge zero folio check, and writeback folio check into
folio_split_supported(). Changed the function return type. Renamed it
to folio_check_splittable() for clarification.
2. Replaced can_split_folio() with open coded folio_expected_ref_count()
and folio_ref_count() and introduced folio_cache_ref_count().
3. Changed min_order_for_split() to always return an order.
4. Fixed folio split stats counting.
Motivation
==========
This is based on Wei's observation[1] and solves several potential
issues:
1. Dereferencing NULL folio->mapping in try_folio_split_to_order() if it
is called on truncated folios.
2. Not handling of negative return value of min_order_for_split() in
mm/memory-failure.c
There is no bug in the current code.
This patch (of 4):
folio_split_supported() used in try_folio_split_to_order() requires
folio->mapping to be non NULL, but current try_folio_split_to_order() does
not check it. There is no issue in the current code, since
try_folio_split_to_order() is only used in truncate_inode_partial_folio(),
where folio->mapping is not NULL.
To prevent future misuse, move folio->mapping NULL check (i.e., folio is
truncated) into folio_split_supported(). Since folio->mapping NULL check
returns -EBUSY and folio_split_supported() == false means -EINVAL, change
folio_split_supported() return type from bool to int and return error
numbers accordingly. Rename folio_split_supported() to
folio_check_splittable() to match the return type change.
While at it, move is_huge_zero_folio() check and folio_test_writeback()
check into folio_check_splittable() and add kernel-doc.
Remove all warnings inside folio_check_splittable() and give warnings
in __folio_split() instead, so that bool warns parameter can be removed.
Link: https://lkml.kernel.org/r/20251126210618.1971206-1-ziy@nvidia.com
Link: https://lkml.kernel.org/r/20251126210618.1971206-2-ziy@nvidia.com
Signed-off-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Wei Yang <richard.weiyang@gmail.com>
Acked-by: Balbir Singh <balbirs@nvidia.com>
Acked-by: David Hildenbrand (Red Hat) <david@kernel.org>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Barry Song <baohua@kernel.org>
Cc: Dev Jain <dev.jain@arm.com>
Cc: Lance Yang <lance.yang@linux.dev>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Naoya Horiguchi <nao.horiguchi@gmail.com>
Cc: Nico Pache <npache@redhat.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
When CONFIG_SPARSEMEM is disabled, the macro
sparse_vmemmap_init_nid_early(_nid, _use) passes two arguments, while the
actual function accepts only nid. Drop the extra argument _use.
Link: https://lkml.kernel.org/r/20251127092512.278-1-guojinhui.liam@bytedance.com
Fixes: d65917c423 ("mm/sparse: allow for alternate vmemmap section init at boot")
Signed-off-by: Jinhui Guo <guojinhui.liam@bytedance.com>
Cc: Frank van der Linden <fvdl@google.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: "David Hildenbrand (Red Hat)" <david@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Let's properly adjust BALLOON_MIGRATE like the other drivers.
Note that the INFLATE/DEFLATE events are triggered from the core when
enqueueing/dequeueing pages.
This was found by code inspection.
Link: https://lkml.kernel.org/r/20251021100606.148294-3-david@redhat.com
Fixes: fe030c9b85 ("powerpc/pseries/cmm: Implement balloon compaction")
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "powerpc/pseries/cmm: two smaller fixes".
Two smaller fixes identified while doing a bigger rework.
This patch (of 2):
We always have to initialize the balloon_dev_info, even when compaction is
not configured in: otherwise the containing list and the lock are left
uninitialized.
Likely not many such configs exist in practice, but let's CC stable to
be sure.
This was found by code inspection.
Link: https://lkml.kernel.org/r/20251021100606.148294-1-david@redhat.com
Link: https://lkml.kernel.org/r/20251021100606.148294-2-david@redhat.com
Fixes: fe030c9b85 ("powerpc/pseries/cmm: Implement balloon compaction")
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The xe3p_xpc platform supports Indirect Ring State and
it is required for the upcoming multi-queue feature.
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patch.msgid.link/20251204063451.1180387-2-niranjana.vishwanathapura@intel.com
Consider a multi-link AP configuration:
MLD vif (MAC addr: aa:bb)
|-- 2.4 GHz link (BSSID: aa:bb)
|-- 5 GHz link (BSSID: cc:dd)
For AP vdevs, ath12k creates a DP peer using the arvif's BSSID and stores
it in dp_hw->dp_peers_list. During scan operations, the driver assigns an
arvif to the scan vdev and uses the vif's MAC address as its BSSID. In
the above scenario, the scan vdev MAC address (aa:bb) matches the BSSID
of the 2.4 GHz AP link, causing a duplicate entry in dp_hw->dp_peers_list
and leading to scan vdev creation failure.
Failure in vif bringup sequence:
1. Create AP vdev for 2.4 GHz link:
- Assign arvif with BSSID = aa:bb and link_id = 0.
- Create DP peer with address aa:bb and add to dp_hw->dp_peers_list.
2. Create scan vdev for 5 GHz link:
- Assign arvif with BSSID = aa:bb (same as vif MAC address) and
link_id = 15.
- Attempt to create another DP peer with address aa:bb.
- Operation fails because aa:bb already exists in dp_hw->dp_peers_list,
resulting in duplicate entry conflict.
3. Delete scan vdev for 5 GHz link.
4. Create AP vdev for 5 GHz link.
Since DP peer is not needed for scan operations, identify scan vdev using
arvif->link_id >= IEEE80211_MLD_MAX_NUM_LINKS and skip DP peer creation
and deletion.
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.5-01651-QCAHKSWPL_SILICONZ-1
Signed-off-by: Ripan Deuri <quic_rdeuri@quicinc.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Link: https://patch.msgid.link/20251207072717.95542-1-quic_rdeuri@quicinc.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
blk_mq_{add,del}_queue_tag_set() functions add and remove queues from
tagset, the functions make sure that tagset and queues are marked as
shared when two or more queues are attached to the same tagset.
Initially a tagset starts as unshared and when the number of added
queues reaches two, blk_mq_add_queue_tag_set() marks it as shared along
with all the queues attached to it. When the number of attached queues
drops to 1 blk_mq_del_queue_tag_set() need to mark both the tagset and
the remaining queues as unshared.
Both functions need to freeze current queues in tagset before setting on
unsetting BLK_MQ_F_TAG_QUEUE_SHARED flag. While doing so, both functions
hold set->tag_list_lock mutex, which makes sense as we do not want
queues to be added or deleted in the process. This used to work fine
until commit 98d81f0df7 ("nvme: use blk_mq_[un]quiesce_tagset")
made the nvme driver quiesce tagset instead of quiscing individual
queues. blk_mq_quiesce_tagset() does the job and quiesce the queues in
set->tag_list while holding set->tag_list_lock also.
This results in deadlock between two threads with these stacktraces:
__schedule+0x47c/0xbb0
? timerqueue_add+0x66/0xb0
schedule+0x1c/0xa0
schedule_preempt_disabled+0xa/0x10
__mutex_lock.constprop.0+0x271/0x600
blk_mq_quiesce_tagset+0x25/0xc0
nvme_dev_disable+0x9c/0x250
nvme_timeout+0x1fc/0x520
blk_mq_handle_expired+0x5c/0x90
bt_iter+0x7e/0x90
blk_mq_queue_tag_busy_iter+0x27e/0x550
? __blk_mq_complete_request_remote+0x10/0x10
? __blk_mq_complete_request_remote+0x10/0x10
? __call_rcu_common.constprop.0+0x1c0/0x210
blk_mq_timeout_work+0x12d/0x170
process_one_work+0x12e/0x2d0
worker_thread+0x288/0x3a0
? rescuer_thread+0x480/0x480
kthread+0xb8/0xe0
? kthread_park+0x80/0x80
ret_from_fork+0x2d/0x50
? kthread_park+0x80/0x80
ret_from_fork_asm+0x11/0x20
__schedule+0x47c/0xbb0
? xas_find+0x161/0x1a0
schedule+0x1c/0xa0
blk_mq_freeze_queue_wait+0x3d/0x70
? destroy_sched_domains_rcu+0x30/0x30
blk_mq_update_tag_set_shared+0x44/0x80
blk_mq_exit_queue+0x141/0x150
del_gendisk+0x25a/0x2d0
nvme_ns_remove+0xc9/0x170
nvme_remove_namespaces+0xc7/0x100
nvme_remove+0x62/0x150
pci_device_remove+0x23/0x60
device_release_driver_internal+0x159/0x200
unbind_store+0x99/0xa0
kernfs_fop_write_iter+0x112/0x1e0
vfs_write+0x2b1/0x3d0
ksys_write+0x4e/0xb0
do_syscall_64+0x5b/0x160
entry_SYSCALL_64_after_hwframe+0x4b/0x53
The top stacktrace is showing nvme_timeout() called to handle nvme
command timeout. timeout handler is trying to disable the controller and
as a first step, it needs to blk_mq_quiesce_tagset() to tell blk-mq not
to call queue callback handlers. The thread is stuck waiting for
set->tag_list_lock as it tries to walk the queues in set->tag_list.
The lock is held by the second thread in the bottom stack which is
waiting for one of queues to be frozen. The queue usage counter will
drop to zero after nvme_timeout() finishes, and this will not happen
because the thread will wait for this mutex forever.
Given that [un]quiescing queue is an operation that does not need to
sleep, update blk_mq_[un]quiesce_tagset() to use RCU instead of taking
set->tag_list_lock, update blk_mq_{add,del}_queue_tag_set() to use RCU
safe list operations. Also, delete INIT_LIST_HEAD(&q->tag_set_list)
in blk_mq_del_queue_tag_set() because we can not re-initialize it while
the list is being traversed under RCU. The deleted queue will not be
added/deleted to/from a tagset and it will be freed in blk_free_queue()
after the end of RCU grace period.
Signed-off-by: Mohamed Khalfella <mkhalfella@purestorage.com>
Fixes: 98d81f0df7 ("nvme: use blk_mq_[un]quiesce_tagset")
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
__bio_for_each_segment() uses the returned struct bio_vec's bv_len field
to advance the struct bvec_iter at the end of each loop iteration. So
it's incorrect to modify it during the loop. Don't assign to bv_len (or
bv_offset, for that matter) in ublk_copy_user_pages().
Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Fixes: e87d66ab27 ("ublk: use rq_for_each_segment() for user copy")
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that all potential callers of bio_chain_endio have been
eliminated, completely prohibit any future calls to this function.
Suggested-by: Ming Lei <ming.lei@redhat.com>
Suggested-by: Andreas Gruenbacher <agruenba@redhat.com>
Suggested-by: Christoph Hellwig <hch@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Shida Zhang <zhangshida@kylinos.cn>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Don't call bio->bi_end_io() directly. Use the bio_endio() helper
function instead, which handles completion more safely and uniformly.
Suggested-by: Christoph Hellwig <hch@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Shida Zhang <zhangshida@kylinos.cn>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
GCC notices that the 16-byte uabi_name field could theoretically be too
small for the formatted string if the instance number exceeds 100.
So grow the field to 20 bytes.
drivers/gpu/drm/i915/intel_memory_region.c: In function ‘intel_memory_region_create’:
drivers/gpu/drm/i915/intel_memory_region.c:273:61: error: ‘%u’ directive output may be truncated writing between 1 and 5 bytes into a region of size between 3 and 11 [-Werror=format-truncation=]
273 | snprintf(mem->uabi_name, sizeof(mem->uabi_name), "%s%u",
| ^~
drivers/gpu/drm/i915/intel_memory_region.c:273:58: note: directive argument in the range [0, 65535]
273 | snprintf(mem->uabi_name, sizeof(mem->uabi_name), "%s%u",
| ^~~~~~
drivers/gpu/drm/i915/intel_memory_region.c:273:9: note: ‘snprintf’ output between 7 and 19 bytes into a destination of size 16
273 | snprintf(mem->uabi_name, sizeof(mem->uabi_name), "%s%u",
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
274 | intel_memory_type_str(type), instance);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fixes: 3b38d35157 ("drm/i915: Add stable memory region names")
Cc: <stable@vger.kernel.org> # v6.8+
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251205113500.684286-2-ardb@kernel.org
(cherry picked from commit 18476087f1)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
During fbdev probe, the xe driver allocates and pins a framebuffer
BO (via xe_bo_create_pin_map_novm() → xe_ggtt_insert_bo()).
Without a runtime PM reference, xe_pm_runtime_get_noresume() warns about
missing outer PM protection as below:
xe 0000:03:00.0: [drm] Missing outer runtime PM protection
Acquire a runtime PM reference before framebuffer allocation to ensure
xe_ggtt_insert_bo() executes under active runtime PM context.
Changes in v2:
- Update commit message to add Fixes tag (Jani Nikula)
Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/6350
Fixes: 44e694958b ("drm/xe/display: Implement display support")
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Dibin Moolakadan Subrahmanian <dibin.moolakadan.subrahmanian@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Link: https://patch.msgid.link/20251111135403.3415947-1-dibin.moolakadan.subrahmanian@intel.com
(cherry picked from commit 37fc7b7b3a)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
intel_display_driver_probe_nogem() is too high of an abstraction level
to be looping and initializing individual CRTCs. Move this to
intel_crtc.c and repurpose intel_crtc_init() to initialize all
CRTCs. Make the original a static __intel_crtc_init() for initializing a
single CRTC.
Reviewed-by: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
Link: https://patch.msgid.link/20251204143235.3138973-1-jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Bump the driver version to reflect the new cached-CPU mapping
capability.
v2:
- Quickly describe what the new version exposes in the commit message
v3:
- Add Steve's R-b
v4:
- No changes
v5:
- No changes
v6:
- No changes
v7:
- No changes
v8:
- No changes
Signed-off-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-14-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Will be used by the UMD to optimize CPU accesses to buffers
that are frequently read by the CPU, or on which the access
pattern makes non-cacheable mappings inefficient.
Mapping buffers CPU-cached implies taking care of the CPU
cache maintenance in the UMD, unless the GPU is IO coherent.
v2:
- Add more to the commit message
v3:
- No changes
v4:
- Fix the map_wc test in panfrost_ioctl_query_bo_info()
v5:
- Drop Steve's R-b (enough has changed to justify a new review)
v6:
- Collect R-b
v7:
- No changes
v8:
- Fix double drm_gem_object_funcs::export assignment
Signed-off-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-13-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
This is useful when importing BOs, so we can know about cacheability
and flush the caches when needed.
v2:
- New commit
v3:
- Add Steve's R-b
v4:
- No changes
v5:
- No changes
v6:
- No changes
v7:
- No changes
v8:
- No changes
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-12-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
This will be used by the UMD to synchronize CPU-cached mappings when
the UMD can't do it directly (no usermode cache maintenance instruction
on Arm32).
v2:
- Add more to the commit message
- Change the flags to better match the drm_gem_shmem_sync semantics
v3:
- Add Steve's R-b
v4:
- No changes
v5:
- Drop Steve's R-b (semantics changes requiring a new review)
v6:
- Bail out early in panfrost_ioctl_sync_bo() if op_count is zero
v7:
- Hand-roll our own bo_sync() helper
v8:
- Collect R-b
Signed-off-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-11-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Will be needed if we want to skip CPU cache maintenance operations when
the GPU can snoop CPU caches.
v2:
- New commit
v3:
- Fix the coherency values (enum instead of bitmask)
v4:
- Fix init/test on coherency_features
v5:
- No changes
v6:
- Collect R-b
v7:
- No changes
v8:
- No changes
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-10-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Before we introduce cached CPU mappings, we want a dma_buf
implementation satisfying synchronization requests around CPU
accesses coming from a dma_buf exported by our driver. Let's
provide our own implementation relying on the default
gem_shmem_prime helpers designed for that purpose.
v5:
- New patch
v6:
- Collect R-b
v7:
- Hand-roll our own dma_buf boilerplate
v8:
- No changes
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-9-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Bump the driver version to reflect the new cached-CPU mapping
capability.
v2:
- Quickly describe what the new version exposes in the commit message
v3:
- Add Steve's R-b
v4:
- No changes
v5:
- No changes
v6:
- No changes
v7:
- No changes
v8:
- Bump to 1.7 (1.6 was picked in the meantime)
Signed-off-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-8-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Will be used by the UMD to optimize CPU accesses to buffers
that are frequently read by the CPU, or on which the access
pattern makes non-cacheable mappings inefficient.
Mapping buffers CPU-cached implies taking care of the CPU
cache maintenance in the UMD, unless the GPU is IO coherent.
v2:
- Add more to the commit message
- Tweak the doc
- Make sure we sync the section of the BO pointing to the CS
syncobj before we read its seqno
v3:
- Fix formatting/spelling issues
v4:
- Add Steve's R-b
v5:
- Drop Steve's R-b (changes in the ioctl semantics requiring
new review)
v6:
- Fix the uAPI doc
- Fix inverted logic in some comment
v7:
- No changes
v8:
- Collect R-b
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-7-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
This is useful when importing BOs, so we can know about cacheability
and flush the caches when needed.
We can also know when the buffer comes from a different subsystem and
take proper actions (avoid CPU mappings, or do kernel-based syncs
instead of userland cache flushes).
v2:
- New commit
v3:
- Add Steve's R-b
v4:
- No changes
v5:
- No changes
v6:
- No changes
v7:
- No changes
v8:
- No changes
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-6-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
This will be used by the UMD to synchronize CPU-cached mappings when
the UMD can't do it directly (no usermode cache maintenance instruction
on Arm32).
v2:
- Change the flags so they better match the drm_gem_shmem_sync()
semantics
v3:
- Add Steve's R-b
v4:
- No changes
v5:
- Drop Steve's R-b (the semantics changes call for a new review)
v6:
- Drop ret initialization in panthor_ioctl_bo_sync()
- Bail out early in panthor_ioctl_bo_sync() if ops.count is zero
- Drop unused PANTHOR_BO_SYNC_OP_FLAGS definition
v7:
- Hand-roll the sync logic (was previously provided by gem_shmem)
v8:
- Collect R-b
Signed-off-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-5-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
If we want to be able to skip CPU cache maintenance operations on
CPU-cached mappings, the UMD needs to know the kind of coherency
in place. Add a field to drm_panthor_gpu_info to do that. We can re-use
a padding field for that since this object is write-only from the
KMD perspective, and the UMD should just ignore it.
v2:
- New commit
v3:
- Make coherency protocol a real enum, not a bitmask
- Add BUILD_BUG_ON()s to make sure the values in panthor_regs.h and
those exposed through the uAPI match
v4:
- Add Steve's R-b
v5:
- No changes
v6:
- No changes
v7:
- Fix kernel doc
v8:
- No changes
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Karunika Choo <karunika.choo@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-4-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
GPU_COHERENCY_PROTOCOL takes one of GPU_COHERENCY_xx
not BIT(GPU_COHERENCY_xx).
v3:
- New commit
v4:
- Add Steve's R-b
v5:
- No changes
v6:
- No changes
v7:
- No changes
v8:
- No changes
Cc: Akash Goel <akash.goel@arm.com>
Fixes: dd7db8d911 ("drm/panthor: Explicitly set the coherency mode")
Reported-by: Steven Price <steven.price@arm.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-3-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Before we introduce cached CPU mappings, we want a dma_buf
implementation satisfying synchronization requests around CPU
accesses coming from a dma_buf exported by our driver. Let's
provide our own implementation relying on the default
gem_shmem_prime helpers designed for that purpose.
v5:
- New patch
v6:
- Collect R-b
v7:
- Hand-roll the dma_buf sync/import logic (was previously done by
generic prime/shmem helpers)
v8:
- No changes
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251208100841.730527-2-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
For events with inherit_stat enabled, a "read" event will be generated
to collect per task event counts on task exit.
The call chain is as follows:
do_exit
-> perf_event_exit_task
-> perf_event_exit_task_context
-> perf_event_exit_event
-> perf_remove_from_context
-> perf_child_detach
-> sync_child_event
-> perf_event_read_event
However, the child event context detaches the task too early in
perf_event_exit_task_context, which causes sync_child_event to never
generate the read event in this case, since child_event->ctx->task is
always set to TASK_TOMBSTONE. Fix that by moving context lock section
backward to ensure ctx->task is not set to TASK_TOMBSTONE before
generating the read event.
Because perf_event_free_task calls perf_event_exit_task_context with
exit = false to tear down all child events from the context, and the
task never lived, accessing the task PID can lead to a use-after-free.
To fix that, let sync_child_event read task from argument and move the
call to the only place it should be triggered to avoid the effect of
setting ctx->task to TASK_TOMESTONE, and add a task parameter to
perf_event_exit_event to trigger the sync_child_event properly when
needed.
This bug can be reproduced by running "perf record -s" and attaching to
any program that generates perf events in its child tasks. If we check
the result with "perf report -T", the last line of the report will leave
an empty table like "# PID TID", which is expected to contain the
per-task event counts by design.
Fixes: ef54c1a476 ("perf: Rework perf_event_exit_event()")
Signed-off-by: Thaumy Cheng <thaumy.love@gmail.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Ian Rogers <irogers@google.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: linux-perf-users@vger.kernel.org
Link: https://patch.msgid.link/20251209041600.963586-1-thaumy.love@gmail.com
Group is_permission_fault() with is_translation_fault(), which is
needed to use is_permission_fault() in __do_kernel_fault(). As
this is static inline, there is no need for this to be under
CONFIG_MMU.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
In the inline assembly inside load_unaligned_zeropad(), the "addr" is
constrained as input-only operand. The compiler assumes that on exit
from the asm statement these operands contain the same values as they
had before executing the statement, but when kernel page fault happened, the assembly fixup code "bic %2 %2, #0x3" modify the value of "addr", which may lead to an unexpected behavior.
Use a temporary variable "tmp" to handle it, instead of modifying the
input-only operand, just like what arm64's load_unaligned_zeropad()
does.
Fixes: b9a50f7490 ("ARM: 7450/1: dcache: select DCACHE_WORD_ACCESS for little-endian ARMv6+ CPUs")
Co-developed-by: Xie Yuanbin <xieyuanbin1@huawei.com>
Signed-off-by: Xie Yuanbin <xieyuanbin1@huawei.com>
Signed-off-by: Liyuan Pang <pangliyuan1@huawei.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Matthieu Baerts says:
====================
mptcp: misc fixes for v6.19-rc1
Here are various unrelated fixes:
- Patches 1-2: ignore unknown in-kernel PM endpoint flags instead of
pretending they are supported. A fix for v5.7.
- Patch 3: avoid potential unnecessary rtx timer expiration. A fix for
v5.15.
- Patch 4: avoid a deadlock on fallback in case of SKB creation failure
while re-injecting.
====================
Link: https://patch.msgid.link/20251205-net-mptcp-misc-fixes-6-19-rc1-v1-0-9e4781a6c1b8@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The MPTCP protocol usually schedule the retransmission timer only
when there is some chances for such retransmissions to happen.
With a notable exception: __mptcp_push_pending() currently schedule
such timer unconditionally, potentially leading to unnecessary rtx
timer expiration.
The issue is present since the blamed commit below but become easily
reproducible after commit 27b0e701d3 ("mptcp: drop bogus optimization
in __mptcp_check_push()")
Fixes: 33d41c9cd7 ("mptcp: more accurate timeout")
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251205-net-mptcp-misc-fixes-6-19-rc1-v1-3-9e4781a6c1b8@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This validates the previous commit: the userspace can set unknown flags
-- the 7th bit is currently unused -- without errors, but only the
supported ones are printed in the endpoints dumps.
The 'Fixes' tag here below is the same as the one from the previous
commit: this patch here is not fixing anything wrong in the selftests,
but it validates the previous fix for an issue introduced by this commit
ID.
Fixes: 01cacb00b3 ("mptcp: add netlink-based PM")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251205-net-mptcp-misc-fixes-6-19-rc1-v1-2-9e4781a6c1b8@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Before this patch, the kernel was saving any flags set by the userspace,
even unknown ones. This doesn't cause critical issues because the kernel
is only looking at specific ones. But on the other hand, endpoints dumps
could tell the userspace some recent flags seem to be supported on older
kernel versions.
Instead, ignore all unknown flags when parsing them. By doing that, the
userspace can continue to set unsupported flags, but it has a way to
verify what is supported by the kernel.
Note that it sounds better to continue accepting unsupported flags not
to change the behaviour, but also that eases things on the userspace
side by adding "optional" endpoint types only supported by newer kernel
versions without having to deal with the different kernel versions.
A note for the backports: there will be conflicts in mptcp.h on older
versions not having the mentioned flags, the new line should still be
added last, and the '5' needs to be adapted to have the same value as
the last entry.
Fixes: 01cacb00b3 ("mptcp: add netlink-based PM")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251205-net-mptcp-misc-fixes-6-19-rc1-v1-1-9e4781a6c1b8@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The wireguard YNL conversion was missing the customary .deps entry.
NIPA doesn't catch this but my CentOS 9 system complains:
wireguard-user.c:72:10: error: ‘WGALLOWEDIP_A_FLAGS’ undeclared here
wireguard-user.c:58:67: error: parameter 1 (‘value’) has incomplete type
58 | const char *wireguard_wgallowedip_flags_str(enum wgallowedip_flag value)
| ~~~~~~~~~~~~~~~~~~~~~~^~~~~
And similarly does Ubuntu 22.04.
One extra complication here is that we renamed the header guard,
so we need to compat with both old and new guard define.
Reviewed-by: Asbjørn Sloth Tønnesen <ast@fiberby.net>
Link: https://patch.msgid.link/20251207013848.1692990-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Recent commit 68e83f3472 ("tools: ynl-gen: add regeneration comment")
added a hint how to regenerate the code to the headers. Update
the new headers from this release cycle to also include it.
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20251207004740.1657799-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Since the only crypto functions used by the mptcp code are the SHA-256
library functions and crypto_memneq(), select only the options needed
for those: CRYPTO_LIB_SHA256 and CRYPTO_LIB_UTILS.
Previously, CRYPTO was selected instead of CRYPTO_LIB_UTILS. That does
pull in CRYPTO_LIB_UTILS as well, but it's unnecessarily broad.
Years ago, the CRYPTO_LIB_* options were visible only when CRYPTO. That
may be another reason why CRYPTO is selected here. However, that was
fixed years ago, and the libraries can now be selected directly.
Signed-off-by: Eric Biggers <ebiggers@kernel.org>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Link: https://patch.msgid.link/20251204054417.491439-1-ebiggers@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Otherwise the lock is susceptible to ever-changing false-sharing due to
unrelated changes. This in particular popped up here where an unrelated
change improved performance:
https://lore.kernel.org/oe-lkp/202511281306.51105b46-lkp@intel.com/
Stabilize it with an explicit annotation which also has a side effect
of furher improving scalability:
> in our oiginal report, 284922f4c5 has a 6.1% performance improvement comparing
> to parent 17d85f33a8.
> we applied your patch directly upon 284922f4c5. as below, now by
> "284922f4c5 + your patch"
> we observe a 12.8% performance improvements (still comparing to 17d85f33a8).
Note nothing was done for the other fields, so some fluctuation is still
possible.
Tested-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20251203100122.291550-1-mjguzik@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Replace the hand-rolled %02x formatting of the Z packet warning in the
BYD driver with the %*ph format specifier. %*ph is the preferred helper
for printing a buffer in hexadecimal and makes the logging clearer and
more consistent.
Signed-off-by: Vivek BalachandharTN <vivek.balachandhar@gmail.com>
Link: https://patch.msgid.link/20251202033120.2264474-1-vivek.balachandhar@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
-----BEGIN PGP SIGNATURE-----
iQGzBAABCgAdFiEE6fsu8pdIjtWE/DpLiiy9cAdyT1EFAmk3lJoACgkQiiy9cAdy
T1HZhAv+IreQpPeRVeq8hd8JYfsys75bo6Dc4eElyO4mLBirMS2wcfM54kCm1Uxr
9dOt/q+nD6YI6L0IYRMwGSawWFqs76qpz5ZxZzg9xQ2AS3jWCPXK1cdtIOZFahJ0
v14Pvo0tF8QmSxqKtTQ10FojWIS9Y9feqWvXW+BbVd2w1zk7iQ//9vnygfzu27o2
DTWqkmEOirt08AUDuROr5vvhl6Ew4J6Dej6jd3C7GuZ5vB0o425vxk3gFE7dI2d1
1h+YeyiLLmgAKyZ9QZAqLVXqgsBz1SflycR6UpW4gvNeTkDxI1MbXsIUz+f/xReJ
SUM6t7gp3hO3XCtW+wYkDq4spIjspJVUNeM1YK0Ip4x5TRy5Jcoe0ikeh/DbET+N
W7mijp+LLTrWG9B1hIXH+0fATuH5dolGmhADOrVqvHL4A8oIp16HyGEjg3dRpMZ+
AsDyjICmv1qXQNX/lvEJtXAQOSyOoO6ouCuavNBXwJVTsSqDrcW5pmifSkF+ZAV8
5PGQNI+2
=PmvL
-----END PGP SIGNATURE-----
Merge tag 'v6.19-rc-part1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client updates from Steve French:
- multichannel fixes, including enabling ability to change multichannel
settings with remount
- debugging improvements: adding additional tracepoints, improving log
messages
- cleanup, including restructuring some of the transport layer for the
client to make it clearer, and cleanup of status code table to be
more consistent with protocol documentation
- fixes for reads that start beyond end of file use cases
- fix to backoff reconnects to reduce reconnect storms
- locking improvement for getting mid entries
- fixes for missing status code error mappings
- performance improvement for status code to error mappings
* tag 'v6.19-rc-part1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6: (22 commits)
smb/client: update some SMB2 status strings
cifs: Remove dead function prototypes
smb/client: add two elements to smb2_error_map_table array
smb: rename to STATUS_SMB_NO_PREAUTH_INTEGRITY_HASH_OVERLAP
smb/client: remove unused elements from smb2_error_map_table array
smb/client: reduce loop count in map_smb2_to_linux_error() by half
smb: client: Add tracepoint for krb5 auth
smb: client: improve error message when creating SMB session
smb: client: relax session and tcon reconnect attempts
cifs: Fix handling of a beyond-EOF DIO/unbuffered read over SMB2
cifs: client: allow changing multichannel mount options on remount
cifs: Do some preparation prior to organising the function declarations
cifs: Add a tracepoint to log EIO errors
cifs: Don't need state locking in smb2_get_mid_entry()
cifs: Remove the server pointer from smb_message
cifs: Fix specification of function pointers
cifs: Replace SendReceiveBlockingLock() with SendReceive() plus flags
cifs: Clean up some places where an extra kvec[] was required for rfc1002
cifs: Make smb1's SendReceive() wrap cifs_send_recv()
cifs: Remove the RFC1002 header from smb_hdr
...
In the DSP event handling code, a put_user() loop copies event data.
When the user buffer size is not aligned to 4 bytes, it could overwrite
beyond the buffer boundary.
Fix by adding a bounds check before put_user().
Suggested-by: Takashi Iwai <tiwai@suse.de>
Fixes: 634ec0b290 ("ALSA: firewire-motu: notify event for parameter change in register DSP model")
Signed-off-by: Junrui Luo <moonafterrain@outlook.com>
Link: https://patch.msgid.link/SYBPR01MB788112C72AF8A1C8C448B4B8AFA3A@SYBPR01MB7881.ausprd01.prod.outlook.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The global gpio_shared_lock has caused some issues when recursively
removing GPIO shared proxies. The thing is: we don't really need it.
Once created from an init-call, the shared GPIO data structures are
never removed, there's no need to protect the linked lists of entries
and references. All we need to protect is the shared GPIO descriptor
(which we already do with a per-entry mutex) and the auxiliary device
data in struct gpio_shared_ref.
Remove the global gpio_shared_lock and use a per-reference mutex to
protect shared references when adding/removing the auxiliary devices and
their GPIO lookup entries.
Link: https://lore.kernel.org/r/20251206-gpio-shared-teardown-fixes-v1-4-35ac458cfce1@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Dropping the last reference to the internal struct device should be the
last thing we do so delete the device first and then uninit it which
also involves the final put_device().
Fixes: a060b8c511 ("gpiolib: implement low-level, shared GPIO support")
Link: https://lore.kernel.org/r/20251206-gpio-shared-teardown-fixes-v1-3-35ac458cfce1@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
It's possible that not all proxy entries will be set up when the device
gets removed so check if they are before trying to dereference members
which are still NULL. This can happen if some consumers never requested
their shared GPIOs.
Fixes: a060b8c511 ("gpiolib: implement low-level, shared GPIO support")
Link: https://lore.kernel.org/r/20251206-gpio-shared-teardown-fixes-v1-2-35ac458cfce1@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Don't consider disabled devices when traversing the device-tree looking
for shared GPIOs. Even if they do share a phandle to a pin, they can
never be instantiated and request it.
Fixes: a060b8c511 ("gpiolib: implement low-level, shared GPIO support")
Link: https://lore.kernel.org/r/20251203092309.34737-1-bartosz.golaszewski@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
