From: Stuart Summers <stuart.summers@intel.com>
Cc: intel-xe@lists.freedesktop.org, matthew.brost@intel.com,
farah.kassabri@intel.com,
Stuart Summers <stuart.summers@intel.com>
Subject: [PATCH 0/9] Add TLB invalidation abstraction
Date: Mon, 25 Aug 2025 17:57:12 +0000 [thread overview]
Message-ID: <20250825175721.89129-1-stuart.summers@intel.com> (raw)
This is a new collection of patches from Matt that has
been floating around internally and on the mailing list.
The goal here is to abstract the actual mechanism of
the invalidation from the higher level invalidation triggers
(like page table updates).
Most of these were brought in unmodified by Matt, but
I've done some minor rebase work here and there and
added my signoff where those rebases seemed a little
more extensive.
Tested on BMG locally.
v9: Use tlb_inval_reset in TLB inval tear down sequence.
v8: Fix documentation failures in CI and rebase
v7: Add a little more documentation around the TLB worker
cancel on teardown and move that cancellation to a
drm teardown helper.
v6: Fix for UAF in timer.c due to outstanding TLB inval
on teardown.
v5: Make sure seqno_lock covers the prep and send in
the later patches (Matt)
v4: Replace CT lock with seqno_lock
v3: Minor spelling fixes and added R-B's per updates on
on the mailing list
v2: Start the series with a new patch to drop the
explicit CT lock (Matt)
Pull in the remaining patches from [1]
[1] https://patchwork.freedesktop.org/series/151670/#rev1
Matthew Brost (7):
drm/xe: s/tlb_invalidation/tlb_inval
drm/xe: Add xe_tlb_inval structure
drm/xe: Add xe_gt_tlb_invalidation_done_handler
drm/xe: Decouple TLB invalidations from GT
drm/xe: Prep TLB invalidation fence before sending
drm/xe: Add helpers to send TLB invalidations
drm/xe: Split TLB invalidation code in frontend and backend
Stuart Summers (2):
drm/xe: Move explicit CT lock in TLB invalidation sequence
drm/xe: Cancel pending TLB inval workers on teardown
drivers/gpu/drm/xe/Makefile | 5 +-
drivers/gpu/drm/xe/xe_device_types.h | 4 +-
drivers/gpu/drm/xe/xe_exec_queue.c | 2 +-
drivers/gpu/drm/xe/xe_ggtt.c | 4 +-
drivers/gpu/drm/xe/xe_gt.c | 8 +-
drivers/gpu/drm/xe/xe_gt_pagefault.c | 1 -
drivers/gpu/drm/xe/xe_gt_tlb_inval_job.h | 34 -
drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c | 604 ------------------
drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h | 40 --
.../gpu/drm/xe/xe_gt_tlb_invalidation_types.h | 32 -
drivers/gpu/drm/xe/xe_gt_types.h | 33 +-
drivers/gpu/drm/xe/xe_guc_ct.c | 8 +-
drivers/gpu/drm/xe/xe_guc_tlb_inval.c | 242 +++++++
drivers/gpu/drm/xe/xe_guc_tlb_inval.h | 19 +
drivers/gpu/drm/xe/xe_lmtt.c | 12 +-
drivers/gpu/drm/xe/xe_migrate.h | 10 +-
drivers/gpu/drm/xe/xe_pci.c | 6 +-
drivers/gpu/drm/xe/xe_pci_types.h | 2 +-
drivers/gpu/drm/xe/xe_pt.c | 63 +-
drivers/gpu/drm/xe/xe_svm.c | 3 +-
drivers/gpu/drm/xe/xe_tlb_inval.c | 434 +++++++++++++
drivers/gpu/drm/xe/xe_tlb_inval.h | 46 ++
..._gt_tlb_inval_job.c => xe_tlb_inval_job.c} | 154 +++--
drivers/gpu/drm/xe/xe_tlb_inval_job.h | 33 +
drivers/gpu/drm/xe/xe_tlb_inval_types.h | 130 ++++
drivers/gpu/drm/xe/xe_trace.h | 24 +-
drivers/gpu/drm/xe/xe_vm.c | 66 +-
drivers/gpu/drm/xe/xe_vm.h | 4 +-
28 files changed, 1098 insertions(+), 925 deletions(-)
delete mode 100644 drivers/gpu/drm/xe/xe_gt_tlb_inval_job.h
delete mode 100644 drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c
delete mode 100644 drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h
delete mode 100644 drivers/gpu/drm/xe/xe_gt_tlb_invalidation_types.h
create mode 100644 drivers/gpu/drm/xe/xe_guc_tlb_inval.c
create mode 100644 drivers/gpu/drm/xe/xe_guc_tlb_inval.h
create mode 100644 drivers/gpu/drm/xe/xe_tlb_inval.c
create mode 100644 drivers/gpu/drm/xe/xe_tlb_inval.h
rename drivers/gpu/drm/xe/{xe_gt_tlb_inval_job.c => xe_tlb_inval_job.c} (50%)
create mode 100644 drivers/gpu/drm/xe/xe_tlb_inval_job.h
create mode 100644 drivers/gpu/drm/xe/xe_tlb_inval_types.h
--
2.34.1
next reply other threads:[~2025-08-25 17:57 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-08-25 17:57 Stuart Summers [this message]
2025-08-25 17:57 ` [PATCH 1/9] drm/xe: Move explicit CT lock in TLB invalidation sequence Stuart Summers
2025-08-25 17:57 ` [PATCH 2/9] drm/xe: Cancel pending TLB inval workers on teardown Stuart Summers
2025-08-25 18:06 ` Summers, Stuart
2025-08-25 18:20 ` Matthew Brost
2025-08-25 18:23 ` Summers, Stuart
2025-08-25 18:32 ` Summers, Stuart
2025-08-25 17:57 ` [PATCH 3/9] drm/xe: s/tlb_invalidation/tlb_inval Stuart Summers
2025-08-25 17:57 ` [PATCH 4/9] drm/xe: Add xe_tlb_inval structure Stuart Summers
2025-08-25 17:57 ` [PATCH 5/9] drm/xe: Add xe_gt_tlb_invalidation_done_handler Stuart Summers
2025-08-25 17:57 ` [PATCH 6/9] drm/xe: Decouple TLB invalidations from GT Stuart Summers
2025-08-25 17:57 ` [PATCH 7/9] drm/xe: Prep TLB invalidation fence before sending Stuart Summers
2025-08-25 17:57 ` [PATCH 8/9] drm/xe: Add helpers to send TLB invalidations Stuart Summers
2025-08-25 17:57 ` [PATCH 9/9] drm/xe: Split TLB invalidation code in frontend and backend Stuart Summers
2025-08-25 19:09 ` ✗ CI.checkpatch: warning for Add TLB invalidation abstraction (rev9) Patchwork
2025-08-25 19:10 ` ✓ CI.KUnit: success " Patchwork
2025-08-25 20:09 ` ✓ Xe.CI.BAT: " Patchwork
2025-08-26 6:18 ` ✓ Xe.CI.Full: " Patchwork
-- strict thread matches above, loose matches on Subject: below --
2025-08-26 18:29 [PATCH 0/9] Add TLB invalidation abstraction Stuart Summers
2025-08-20 23:30 Stuart Summers
2025-08-20 22:45 Stuart Summers
2025-08-13 19:47 stuartsummers
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=20250825175721.89129-1-stuart.summers@intel.com \
--to=stuart.summers@intel.com \
--cc=farah.kassabri@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).