Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [RFC] drm/xe/ras: Report one SIGID per device memory error
@ 2026-11-09 16:17 Rodrigo Vivi
  2026-11-09 17:08 ` ✓ CI.KUnit: success for " Patchwork
                   ` (2 more replies)
  0 siblings, 3 replies; 4+ messages in thread
From: Rodrigo Vivi @ 2026-11-09 16:17 UTC (permalink / raw)
  To: intel-xe
  Cc: Rodrigo Vivi, Riana Tauro, Tejas Upadhyay, Himal Prasad Ghimiray,
	Badal Nilawar, Raag Jadav, Mallesh Koujalagi, Aravind Iddamsetty

handle_page_offline() logged each step as SIGID 202 RECOVERABLE, so a
single double-bit ECC error produced several 202 lines whose severity
disagreed with the CPER record emitted by handle_device_memory_errors().

Log those steps without a SIGID and leave the caller's 202 with CPER as
the only report. Firmware failures keep SIGID 105.

Cc: Riana Tauro <riana.tauro@intel.com>
Cc: Tejas Upadhyay <tejas.upadhyay@intel.com>
Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Cc: Badal Nilawar <badal.nilawar@intel.com>
Cc: Raag Jadav <raag.jadav@intel.com>
Cc: Mallesh Koujalagi <mallesh.koujalagi@intel.com>
Cc: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---

AI generated and untested.

Aiming to solve the discussion:
https://lore.kernel.org/intel-xe/520a56ea-40a6-47bd-84a0-be5cf604e5e7@intel.com/#r

I agree with Badal that we should reduce the noise in the mempage-offline
that is only bringing multiple entries of the same and then emmit only a single
SIGID and a single CPER for it.

 drivers/gpu/drm/xe/xe_ras.c | 32 +++++++++++++-------------------
 1 file changed, 13 insertions(+), 19 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_ras.c b/drivers/gpu/drm/xe/xe_ras.c
index 2c01c183d02b..5b13daf972ba 100644
--- a/drivers/gpu/drm/xe/xe_ras.c
+++ b/drivers/gpu/drm/xe/xe_ras.c
@@ -269,11 +269,9 @@ static int send_page_offline_cmd(struct xe_device *xe, u64 page_address,
 	request.action = action;
 
 	if (action == XE_RAS_PAGE_ACTION_OFFLINE)
-		xe_log_err(xe, DEVICE_MEMORY, 0, "Requesting firmware to offline page 0x%llx\n",
-			   page_address);
+		xe_dbg(xe, "Requesting firmware to offline page 0x%llx\n", page_address);
 	else
-		xe_log_err(xe, DEVICE_MEMORY, 0, "Requesting firmware to remove page 0x%llx from queue\n",
-			   page_address);
+		xe_dbg(xe, "Requesting firmware to remove page 0x%llx from queue\n", page_address);
 
 	xe_sysctrl_create_command(&command, XE_SYSCTRL_GROUP_GFSP, XE_SYSCTRL_CMD_PAGE_OFFLINE,
 				  &request, sizeof(request), &response, sizeof(response));
@@ -326,7 +324,7 @@ static int handle_page_offline(struct xe_device *xe, u64 page_address, bool noti
 	/* User policy set to disable page offlining */
 	if (xe_ras_get_disable_vram_page_offline(xe)) {
 		action = XE_RAS_PAGE_ACTION_REMOVE;
-		xe_log_err(xe, DEVICE_MEMORY, -EOPNOTSUPP, "User policy set to disable page offlining\n");
+		xe_info(xe, "User policy set to disable page offlining\n");
 		goto notify;
 	}
 
@@ -346,14 +344,12 @@ static int handle_page_offline(struct xe_device *xe, u64 page_address, bool noti
 	switch (ret) {
 	case 0:
 		action = XE_RAS_PAGE_ACTION_REMOVE;
-		xe_log_err(xe, DEVICE_MEMORY, ret,
-			   "Poison detected at physical address 0x%llx, page soft-offlined\n",
-			   page_address);
+		xe_info(xe, "Poison detected at physical address 0x%llx, page soft-offlined\n",
+			page_address);
 		break;
 	case -EIO:
-		xe_log_err(xe, DEVICE_MEMORY, ret,
-			   "Poison detected at physical address 0x%llx, page belongs to critical BO and cannot be soft-offlined\n",
-			   page_address);
+		xe_err(xe, "Poison detected at physical address 0x%llx, page belongs to critical BO and cannot be soft-offlined\n",
+		       page_address);
 		return ret;
 	case -EEXIST:
 		if (xa_load(&state->offlined_pages, page_address >> XE_PTE_SHIFT))
@@ -361,18 +357,16 @@ static int handle_page_offline(struct xe_device *xe, u64 page_address, bool noti
 		else
 			action = XE_RAS_PAGE_ACTION_OFFLINE;
 
-		xe_log_err(xe, DEVICE_MEMORY, ret,
-			   "Double-bit ECC error detected at physical address 0x%llx, page soft-offlined\n",
-			   page_address);
+		xe_info(xe, "Double-bit ECC error detected at physical address 0x%llx, page soft-offlined\n",
+			page_address);
 		break;
 	case -ENOMEM:
-		xe_log_err(xe, DEVICE_MEMORY, ret,
-			   "Poison detected at physical address 0x%llx, failed to soft-offline\n",
-			   page_address);
+		xe_err(xe, "Poison detected at physical address 0x%llx, failed to soft-offline\n",
+		       page_address);
 		return ret;
 	default:
-		xe_log_err(xe, DEVICE_MEMORY, ret, "Failed to handle address fault at physical address 0x%llx\n",
-			   page_address);
+		xe_err(xe, "Failed to handle address fault at physical address 0x%llx: %d\n",
+		       page_address, ret);
 		return 0;
 	}
 
-- 
2.55.0


^ permalink raw reply related	[flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-11-10 16:42 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-09 16:17 [RFC] drm/xe/ras: Report one SIGID per device memory error Rodrigo Vivi
2026-11-09 17:08 ` ✓ CI.KUnit: success for " Patchwork
2026-11-09 17:39 ` ✓ Xe.CI.BAT: " Patchwork
2026-11-10 16:41 ` [RFC] " Rodrigo Vivi

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