@@ -1123,6 +1123,10 @@ int intel_guc_submission_enable(struct intel_guc *guc)
enum intel_engine_id id;
int err;
+ err = i915_inject_load_error(gt->i915, -ENXIO);
+ if (err)
+ return err;
+
/*
* We're using GuC work items for submitting work through GuC. Since
* we're coalescing multiple requests from a single context into a
@@ -139,6 +139,10 @@ int intel_huc_auth(struct intel_huc *huc)
GEM_BUG_ON(!intel_uc_fw_is_loaded(&huc->fw));
GEM_BUG_ON(intel_huc_is_authenticated(huc));
+ ret = i915_inject_load_error(gt->i915, -ENXIO);
+ if (ret)
+ goto fail;
+
ret = intel_guc_auth_huc(guc,
intel_guc_ggtt_offset(guc, huc->rsa_data));
if (ret) {
@@ -158,13 +162,11 @@ int intel_huc_auth(struct intel_huc *huc)
}
huc->fw.status = INTEL_UC_FIRMWARE_RUNNING;
-
return 0;
fail:
+ i915_probe_error(gt->i915, "HuC: Authentication failed %d\n", ret);
huc->fw.status = INTEL_UC_FIRMWARE_FAIL;
-
- DRM_ERROR("HuC: Authentication failed %d\n", ret);
return ret;
}
@@ -41,6 +41,10 @@ static int __intel_uc_reset_hw(struct intel_uc *uc)
int ret;
u32 guc_status;
+ ret = i915_inject_load_error(gt->i915, -ENXIO);
+ if (ret)
+ return ret;
+
ret = intel_reset_guc(gt);
if (ret) {
DRM_ERROR("Failed to reset GuC, ret = %d\n", ret);
@@ -209,6 +213,10 @@ static int guc_enable_communication(struct intel_guc *guc)
GEM_BUG_ON(guc_communication_enabled(guc));
+ ret = i915_inject_load_error(i915, -ENXIO);
+ if (ret)
+ return ret;
+
ret = intel_guc_ct_enable(&guc->ct);
if (ret)
return ret;
@@ -340,7 +348,7 @@ void intel_uc_fini(struct intel_uc *uc)
intel_guc_fini(guc);
}
-static void __uc_sanitize(struct intel_uc *uc)
+static int __uc_sanitize(struct intel_uc *uc)
{
struct intel_guc *guc = &uc->guc;
struct intel_huc *huc = &uc->huc;
@@ -350,7 +358,7 @@ static void __uc_sanitize(struct intel_uc *uc)
intel_huc_sanitize(huc);
intel_guc_sanitize(guc);
- __intel_uc_reset_hw(uc);
+ return __intel_uc_reset_hw(uc);
}
void intel_uc_sanitize(struct intel_uc *uc)
@@ -378,6 +386,10 @@ static int uc_init_wopcm(struct intel_uc *uc)
GEM_BUG_ON(!(size & GUC_WOPCM_SIZE_MASK));
GEM_BUG_ON(size & ~GUC_WOPCM_SIZE_MASK);
+ err = i915_inject_load_error(gt->i915, -ENXIO);
+ if (err)
+ return err;
+
mask = GUC_WOPCM_SIZE_MASK | GUC_WOPCM_SIZE_LOCKED;
err = intel_uncore_write_and_verify(uncore, GUC_WOPCM_SIZE, size, mask,
size | GUC_WOPCM_SIZE_LOCKED);
@@ -434,7 +446,7 @@ int intel_uc_init_hw(struct intel_uc *uc)
* Always reset the GuC just before (re)loading, so
* that the state and timing are fairly predictable
*/
- ret = __intel_uc_reset_hw(uc);
+ ret = __uc_sanitize(uc);
if (ret)
goto err_out;
@@ -504,7 +516,7 @@ int intel_uc_init_hw(struct intel_uc *uc)
if (GEM_WARN_ON(ret == -EIO))
ret = -EINVAL;
- dev_err(i915->drm.dev, "GuC initialization failed %d\n", ret);
+ i915_probe_error(i915, "GuC initialization failed %d\n", ret);
return ret;
}
@@ -383,6 +383,10 @@ static int uc_fw_xfer(struct intel_uc_fw *uc_fw, struct intel_gt *gt,
u64 offset;
int ret;
+ ret = i915_inject_load_error(gt->i915, -ETIMEDOUT);
+ if (ret)
+ return ret;
+
intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
/* Set the source address for the uCode */
@@ -443,8 +447,13 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, struct intel_gt *gt,
/* make sure the status was cleared the last time we reset the uc */
GEM_BUG_ON(intel_uc_fw_is_loaded(uc_fw));
+ err = i915_inject_load_error(gt->i915, -ENOEXEC);
+ if (err)
+ return err;
+
if (!intel_uc_fw_is_available(uc_fw))
return -ENOEXEC;
+
/* Call custom loader */
intel_uc_fw_ggtt_bind(uc_fw, gt);
err = uc_fw_xfer(uc_fw, gt, wopcm_offset, dma_flags);
@@ -464,13 +473,10 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, struct intel_gt *gt,
return 0;
fail:
+ i915_probe_error(gt->i915, "Failed to load %s firmware %s (%d)\n",
+ intel_uc_fw_type_repr(uc_fw->type), uc_fw->path,
+ err);
uc_fw->status = INTEL_UC_FIRMWARE_FAIL;
- DRM_DEBUG_DRIVER("%s fw load failed\n",
- intel_uc_fw_type_repr(uc_fw->type));
-
- DRM_WARN("%s: Failed to load firmware %s (error %d)\n",
- intel_uc_fw_type_repr(uc_fw->type), uc_fw->path, err);
-
return err;
}
@@ -37,12 +37,12 @@ struct intel_gt;
#define INTEL_UC_FIRMWARE_URL "https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/i915"
enum intel_uc_fw_status {
- INTEL_UC_FIRMWARE_FAIL = -3, /* failed to xfer or init/auth the fw */
- INTEL_UC_FIRMWARE_MISSING = -2, /* blob not found on the system */
- INTEL_UC_FIRMWARE_NOT_SUPPORTED = -1, /* no uc HW */
+ INTEL_UC_FIRMWARE_NOT_SUPPORTED = -1, /* no uc HW or disabled */
INTEL_UC_FIRMWARE_UNINITIALIZED = 0, /* used to catch checks done too early */
INTEL_UC_FIRMWARE_SELECTED, /* selected the blob we want to load */
+ INTEL_UC_FIRMWARE_MISSING, /* blob not found on the system */
INTEL_UC_FIRMWARE_AVAILABLE, /* blob found and copied in mem */
+ INTEL_UC_FIRMWARE_FAIL, /* failed to xfer or init/auth the fw */
INTEL_UC_FIRMWARE_TRANSFERRED, /* dma xfer done */
INTEL_UC_FIRMWARE_RUNNING /* init/auth done */
};
@@ -83,18 +83,18 @@ static inline
const char *intel_uc_fw_status_repr(enum intel_uc_fw_status status)
{
switch (status) {
- case INTEL_UC_FIRMWARE_FAIL:
- return "FAIL";
- case INTEL_UC_FIRMWARE_MISSING:
- return "MISSING";
case INTEL_UC_FIRMWARE_NOT_SUPPORTED:
return "N/A";
case INTEL_UC_FIRMWARE_UNINITIALIZED:
return "UNINITIALIZED";
case INTEL_UC_FIRMWARE_SELECTED:
return "SELECTED";
+ case INTEL_UC_FIRMWARE_MISSING:
+ return "MISSING";
case INTEL_UC_FIRMWARE_AVAILABLE:
return "AVAILABLE";
+ case INTEL_UC_FIRMWARE_FAIL:
+ return "FAIL";
case INTEL_UC_FIRMWARE_TRANSFERRED:
return "TRANSFERRED";
case INTEL_UC_FIRMWARE_RUNNING:
@@ -1243,7 +1243,7 @@ int i915_gem_init_hw(struct drm_i915_private *i915)
/* We can't enable contexts until all firmware is loaded */
ret = intel_uc_init_hw(>->uc);
if (ret) {
- DRM_ERROR("Enabling uc failed (%d)\n", ret);
+ i915_probe_error(i915, "Enabling uc failed (%d)\n", ret);
goto out;
}
Inject probe errors into intel_uc_init_hw to make sure we correctly handle any uC initialization failure. To avoid complains from CI about injected errors use i915_probe_error to lower message level. v2: _sanitize instead _reset to correctly handle Gen9 retries v3: reorder fw status codes as failed fw is still available add more failure points Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> #v1 --- .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 4 ++++ drivers/gpu/drm/i915/gt/uc/intel_huc.c | 8 +++++--- drivers/gpu/drm/i915/gt/uc/intel_uc.c | 20 +++++++++++++++---- drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c | 18 +++++++++++------ drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h | 14 ++++++------- drivers/gpu/drm/i915/i915_gem.c | 2 +- 6 files changed, 45 insertions(+), 21 deletions(-)