Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2 0/6] Add support to handle memory double-bit ecc errors
@ 2026-09-07  9:47 Riana Tauro
  2026-09-07  9:47 ` [PATCH v2 1/6] drm/xe/xe_ras: Handle page offline requests for device memory " Riana Tauro
                   ` (8 more replies)
  0 siblings, 9 replies; 25+ messages in thread
From: Riana Tauro @ 2026-09-07  9:47 UTC (permalink / raw)
  To: intel-xe
  Cc: riana.tauro, anshuman.gupta, rodrigo.vivi, aravind.iddamsetty,
	badal.nilawar, raag.jadav, ravi.kishore.koppuravuri,
	mallesh.koujalagi, soham.purkait, tejas.upadhyay

This series introduces the firmware commands required to handle
uncorrectable memory double-bit ecc errors.

When a error is detected, the driver determines whether the affected
page can be offlined and sends an offline/decline request to firmware
via sysctrl interface. Pages belonging to critical Bos or stolen memory
require a Secondary Bus Reset(SBR) to recover.
Pages configured for log-only handling are declined.

For the rest of the valid pages, the first error is a poison error and a
decline is sent to firmware to avoid marking the page as bad. On the
second occurence at the same address, we request the firmware to offline
the page as this would indicate a Double-bit ECC error. Any subsequent
errors at the same address should be because of memory scrubber and a
decline is sent to firmware to remove the page from queue. This is done
by tracking the firmware offlined pages by Xe KMD.

At module load, the driver queries firmware for both the offline list
and the queue and processes them accordingly.

Rev2: Rebase
      fix sigid logs
      fix sashiko review comments

Riana Tauro (6):
  drm/xe/xe_ras: Handle page offline requests for device memory ecc
    errors
  drm/xe/xe_ras: Add support to query page offline queue and list
  drm/xe: Separate drm-ras netlink data from device and firmware RAS
    state
  drm/xe/xe_ras: Add function to get maximum pages firmware can store
  drm/xe/xe_ttm_vram: Report max_pages reported by firmware in debugfs
  drm/xe/xe_ras: Track offlined pages by firmware to avoid duplicates

 drivers/gpu/drm/xe/xe_device.c                |   4 +-
 drivers/gpu/drm/xe/xe_device_types.h          |  11 +-
 drivers/gpu/drm/xe/xe_drm_ras.c               |  16 +-
 drivers/gpu/drm/xe/xe_drm_ras_types.h         |   3 -
 drivers/gpu/drm/xe/xe_hw_error.c              |   6 +-
 drivers/gpu/drm/xe/xe_ras.c                   | 272 +++++++++++++++++-
 drivers/gpu/drm/xe/xe_ras.h                   |  13 +-
 drivers/gpu/drm/xe/xe_ras_types.h             |  93 ++++++
 drivers/gpu/drm/xe/xe_sysctrl_mailbox_types.h |   6 +
 drivers/gpu/drm/xe/xe_ttm_vram_mgr.c          |   9 +-
 drivers/gpu/drm/xe/xe_ttm_vram_mgr_types.h    |   2 -
 11 files changed, 395 insertions(+), 40 deletions(-)

-- 
2.47.1


^ permalink raw reply	[flat|nested] 25+ messages in thread

end of thread, other threads:[~2026-09-11  6:46 UTC | newest]

Thread overview: 25+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-09-07  9:47 [PATCH v2 0/6] Add support to handle memory double-bit ecc errors Riana Tauro
2026-09-07  9:47 ` [PATCH v2 1/6] drm/xe/xe_ras: Handle page offline requests for device memory " Riana Tauro
2026-09-07 10:14   ` sashiko-bot
2026-09-07 17:35   ` Ghimiray, Himal Prasad
2026-09-11  6:45     ` Tauro, Riana
2026-09-07  9:47 ` [PATCH v2 2/6] drm/xe/xe_ras: Add support to query page offline queue and list Riana Tauro
2026-09-07 10:00   ` sashiko-bot
2026-09-07 18:25   ` Ghimiray, Himal Prasad
2026-09-11  6:11     ` Tauro, Riana
2026-09-11  6:39       ` Ghimiray, Himal Prasad
2026-09-07  9:47 ` [PATCH v2 3/6] drm/xe: Separate drm-ras netlink data from device and firmware RAS state Riana Tauro
2026-09-07 19:04   ` Ghimiray, Himal Prasad
2026-09-07 19:21     ` Ghimiray, Himal Prasad
2026-09-09  5:43       ` Tauro, Riana
2026-09-07  9:47 ` [PATCH v2 4/6] drm/xe/xe_ras: Add function to get maximum pages firmware can store Riana Tauro
2026-09-07 19:21   ` Ghimiray, Himal Prasad
2026-09-07 19:24     ` Ghimiray, Himal Prasad
2026-09-09  5:23       ` Tauro, Riana
2026-09-07  9:47 ` [PATCH v2 5/6] drm/xe/xe_ttm_vram: Report max_pages reported by firmware in debugfs Riana Tauro
2026-09-07 19:26   ` Ghimiray, Himal Prasad
2026-09-07  9:47 ` [PATCH v2 6/6] drm/xe/xe_ras: Track offlined pages by firmware to avoid duplicates Riana Tauro
2026-09-07 10:02   ` sashiko-bot
2026-09-07 11:01 ` ✓ CI.KUnit: success for Add support to handle memory double-bit ecc errors (rev2) Patchwork
2026-09-07 11:45 ` ✓ Xe.CI.BAT: " Patchwork
2026-09-07 13:41 ` ✓ Xe.CI.FULL: " Patchwork

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox