We missed setting the CCS mode during resume and engine resets. Create a workaround to be added in the engine's workaround list. This workaround sets the XEHP_CCS_MODE value at every reset.
The issue can be reproduced by running:
$ clpeak --kernel-latency
Without resetting the CCS mode, we encounter a fence timeout:
Fence expiration time out i915-0000:03:00.0:clpeak[2387]:2!
Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/10895 Fixes: 6db31251bb26 ("drm/i915/gt: Enable only one CCS for compute workload") Reported-by: Gnattu OC gnattuoc@me.com Signed-off-by: Andi Shyti andi.shyti@linux.intel.com Cc: Chris Wilson chris.p.wilson@linux.intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Cc: Matt Roper matthew.d.roper@intel.com Cc: stable@vger.kernel.org # v6.2+ --- Hi Gnattu,
thanks again for reporting this issue and for your prompt replies on the issue. Would you give this patch a chance?
Andi
drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.c | 6 +++--- drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.h | 2 +- drivers/gpu/drm/i915/gt/intel_workarounds.c | 4 +++- 3 files changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.c b/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.c index 044219c5960a..99b71bb7da0a 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.c +++ b/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.c @@ -8,14 +8,14 @@ #include "intel_gt_ccs_mode.h" #include "intel_gt_regs.h"
-void intel_gt_apply_ccs_mode(struct intel_gt *gt) +unsigned int intel_gt_apply_ccs_mode(struct intel_gt *gt) { int cslice; u32 mode = 0; int first_ccs = __ffs(CCS_MASK(gt));
if (!IS_DG2(gt->i915)) - return; + return 0;
/* Build the value for the fixed CCS load balancing */ for (cslice = 0; cslice < I915_MAX_CCS; cslice++) { @@ -35,5 +35,5 @@ void intel_gt_apply_ccs_mode(struct intel_gt *gt) XEHP_CCS_MODE_CSLICE_MASK); }
- intel_uncore_write(gt->uncore, XEHP_CCS_MODE, mode); + return mode; } diff --git a/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.h b/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.h index 9e5549caeb26..55547f2ff426 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.h +++ b/drivers/gpu/drm/i915/gt/intel_gt_ccs_mode.h @@ -8,6 +8,6 @@
struct intel_gt;
-void intel_gt_apply_ccs_mode(struct intel_gt *gt); +unsigned int intel_gt_apply_ccs_mode(struct intel_gt *gt);
#endif /* __INTEL_GT_CCS_MODE_H__ */ diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 68b6aa11bcf7..58693923bf6c 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -2703,6 +2703,7 @@ add_render_compute_tuning_settings(struct intel_gt *gt, static void ccs_engine_wa_mode(struct intel_engine_cs *engine, struct i915_wa_list *wal) { struct intel_gt *gt = engine->gt; + u32 mode;
if (!IS_DG2(gt->i915)) return; @@ -2719,7 +2720,8 @@ static void ccs_engine_wa_mode(struct intel_engine_cs *engine, struct i915_wa_li * After having disabled automatic load balancing we need to * assign all slices to a single CCS. We will call it CCS mode 1 */ - intel_gt_apply_ccs_mode(gt); + mode = intel_gt_apply_ccs_mode(gt); + wa_masked_en(wal, XEHP_CCS_MODE, mode); }
/*