From: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
To: intel-xe@lists.freedesktop.org
Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
"Matthew Brost" <matthew.brost@intel.com>,
"Matthew Auld" <matthew.auld@intel.com>
Subject: [PATCH v2 0/4] drm/xe: Fix userptr races and missed validations
Date: Fri, 28 Feb 2025 08:30:54 +0100 [thread overview]
Message-ID: <20250228073058.59510-1-thomas.hellstrom@linux.intel.com> (raw)
A series that fixes a couple of userptr races and a missed validation.
Patch 1
- Fixes a missed userptr validation causing a potential infinite loop.
Patch 2
- Fixes a misplaced #endif in preparation for patch 3.
Patch 3
- Fixes a race between fault mode invalidation and unbind, where
the userptr invalidation could attempt to traverse a page-table
tree under modification
Patch 4
- Fixes a race between fault mode invalidation and multiple bind
operation, where the invalidation could potentially get lost.
v2:
- Fix an unintended Fallthrough (Matt Brost)
- Don't recommend backport of Patch2 for older kernels
than 6.12+ (Lucas)
- Update the locking requirement docs and asserts for
forced invalidations
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Matthew Brost (1):
drm/xe: Add staging tree for VM binds
Thomas Hellström (3):
drm/xe/vm: Validate userptr during gpu vma prefetching
drm/xe/vm: Fix a misplaced #endif
drm/xe: Fix fault mode invalidation with unbind
drivers/gpu/drm/xe/xe_pt.c | 96 ++++++++++++++++----------------
drivers/gpu/drm/xe/xe_pt_walk.c | 3 +-
drivers/gpu/drm/xe/xe_pt_walk.h | 4 ++
drivers/gpu/drm/xe/xe_vm.c | 96 ++++++++++++++++++++++----------
drivers/gpu/drm/xe/xe_vm.h | 10 +++-
drivers/gpu/drm/xe/xe_vm_types.h | 4 +-
6 files changed, 132 insertions(+), 81 deletions(-)
--
2.48.1
next reply other threads:[~2025-02-28 7:31 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-28 7:30 Thomas Hellström [this message]
2025-02-28 7:30 ` [PATCH v2 1/4] drm/xe/vm: Validate userptr during gpu vma prefetching Thomas Hellström
2025-03-04 17:48 ` Matthew Brost
2025-02-28 7:30 ` [PATCH v2 2/4] drm/xe/vm: Fix a misplaced #endif Thomas Hellström
2025-02-28 7:30 ` [PATCH v2 3/4] drm/xe: Fix fault mode invalidation with unbind Thomas Hellström
2025-02-28 7:30 ` [PATCH v2 4/4] drm/xe: Add staging tree for VM binds Thomas Hellström
2025-02-28 9:40 ` ✓ CI.Patch_applied: success for drm/xe: Fix userptr races and missed validations (rev2) Patchwork
2025-02-28 9:41 ` ✓ CI.checkpatch: " Patchwork
2025-02-28 9:42 ` ✓ CI.KUnit: " Patchwork
2025-02-28 9:59 ` ✓ CI.Build: " Patchwork
2025-02-28 10:01 ` ✓ CI.Hooks: " Patchwork
2025-02-28 10:02 ` ✓ CI.checksparse: " Patchwork
2025-02-28 10:21 ` ✓ Xe.CI.BAT: " Patchwork
2025-03-04 19:30 ` ✓ CI.Patch_applied: success for drm/xe: Fix userptr races and missed validations (rev3) Patchwork
2025-03-04 19:31 ` ✓ CI.checkpatch: " Patchwork
2025-03-04 19:32 ` ✓ CI.KUnit: " Patchwork
2025-03-04 19:49 ` ✓ CI.Build: " Patchwork
2025-03-04 19:51 ` ✓ CI.Hooks: " Patchwork
2025-03-04 19:53 ` ✓ CI.checksparse: " Patchwork
2025-03-04 20:11 ` ✓ Xe.CI.BAT: " Patchwork
2025-03-04 23:19 ` ✗ Xe.CI.Full: failure " Patchwork
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250228073058.59510-1-thomas.hellstrom@linux.intel.com \
--to=thomas.hellstrom@linux.intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.auld@intel.com \
--cc=matthew.brost@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.