Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] drm/xe/cri: Implement workaround Wa_15020021844
@ 2026-10-12 23:30 Matt Roper
  2026-10-13  0:30 ` ✓ CI.KUnit: success for " Patchwork
                   ` (3 more replies)
  0 siblings, 4 replies; 7+ messages in thread
From: Matt Roper @ 2026-10-12 23:30 UTC (permalink / raw)
  To: intel-xe; +Cc: matthew.d.roper, Balasubramani Vivekanandan

On Crescent Island platforms, the pcode is signaling "ready" to the
driver prematurely since one last sysctrl firmware initialization step
is still underway and not fully complete.  Since the driver's
initialization is racing with the end of the firmware's init, this can
lead to error messages like:

   xe 0000:be:00.0: [drm] *ERROR* sysctrl: Firmware error: 0x06
   xe 0000:be:00.0: [drm] *ERROR* sysctrl: Mailbox command failed: -EIO

The current recommenedation to work around this is for the driver to
wait an additional 0.5 seconds after getting the "ready" response,
before moving on with other initialization tasks.

Note that the official definition of this workaround is still being
finalized, so "Wa_15020021844" is a bit of a placeholder/guess.  It's
possible that the actual number could still change, although the request
to wait the additional half second is pretty well set at this point.

Cc: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
---
 drivers/gpu/drm/xe/xe_device_wa_oob.rules | 1 +
 drivers/gpu/drm/xe/xe_pcode.c             | 5 +++++
 2 files changed, 6 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_device_wa_oob.rules b/drivers/gpu/drm/xe/xe_device_wa_oob.rules
index d8dc41851425c..0488a20461d81 100644
--- a/drivers/gpu/drm/xe/xe_device_wa_oob.rules
+++ b/drivers/gpu/drm/xe/xe_device_wa_oob.rules
@@ -7,3 +7,4 @@
 14026633728	PLATFORM(CRESCENTISLAND)
 14026746987	PLATFORM(CRESCENTISLAND)
 14026779378	PLATFORM(CRESCENTISLAND)
+15020021844	PLATFORM(CRESCENTISLAND)
diff --git a/drivers/gpu/drm/xe/xe_pcode.c b/drivers/gpu/drm/xe/xe_pcode.c
index 266deecbb100a..897d44ecae42b 100644
--- a/drivers/gpu/drm/xe/xe_pcode.c
+++ b/drivers/gpu/drm/xe/xe_pcode.c
@@ -11,6 +11,7 @@
 
 #include <drm/drm_managed.h>
 
+#include "generated/xe_device_wa_oob.h"
 #include "regs/xe_pmt.h"
 #include "xe_assert.h"
 #include "xe_device.h"
@@ -20,6 +21,7 @@
 #include "xe_pm.h"
 #include "xe_printk.h"
 #include "xe_vsec.h"
+#include "xe_wa.h"
 
 /**
  * DOC: PCODE
@@ -320,6 +322,9 @@ int xe_pcode_ready(struct xe_device *xe, bool locked)
 		xe_log_err(tile, PCODE, ret, "initialization timedout after %ld seconds\n",
 			   timeout_us / USEC_PER_SEC);
 
+	if (!ret && XE_DEVICE_WA(xe, 15020021844))
+		msleep(500);
+
 	return ret;
 }
 
-- 
2.43.0


^ permalink raw reply related	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2026-10-13 15:14 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-12 23:30 [PATCH] drm/xe/cri: Implement workaround Wa_15020021844 Matt Roper
2026-10-13  0:30 ` ✓ CI.KUnit: success for " Patchwork
2026-10-13  4:03 ` ✓ Xe.CI.BAT: " Patchwork
2026-10-13  4:49 ` [PATCH] " Tauro, Riana
2026-10-13  5:22   ` Raag Jadav
2026-10-13 15:13     ` Matt Roper
2026-10-13 13:39 ` ✗ Xe.CI.FULL: failure for " Patchwork

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