| Message ID | 20180628082615.10805-1-chris@chris-wilson.co.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Quoting Patchwork (2018-06-28 09:55:21) > == Series Details == > > Series: drm/i915: Show vma allocator stack when in doubt > URL : https://patchwork.freedesktop.org/series/45562/ > State : failure > > == Summary == > > = CI Bug Log - changes from CI_DRM_4396 -> Patchwork_9460 = > > == Summary - FAILURE == > > Serious unknown changes coming with Patchwork_9460 absolutely need to be > verified manually. > > If you think the reported changes have nothing to do with the changes > introduced in Patchwork_9460, please notify your bug team to allow them > to document this new failure mode, which will reduce false positives in CI. > > External URL: https://patchwork.freedesktop.org/api/1.0/series/45562/revisions/1/mbox/ > > == Possible new issues == > > Here are the unknown changes that may have been introduced in Patchwork_9460: > > === IGT changes === > > ==== Possible regressions ==== > > igt@gem_exec_gttfill@basic: > {fi-kbl-x1275}: PASS -> DMESG-WARN > fi-kbl-7500u: PASS -> DMESG-WARN > fi-kbl-guc: PASS -> DMESG-WARN > fi-skl-6600u: PASS -> DMESG-WARN > fi-kbl-7560u: PASS -> DMESG-WARN > fi-cfl-s3: PASS -> DMESG-WARN > fi-skl-6770hq: PASS -> DMESG-WARN > fi-skl-6700k2: PASS -> DMESG-WARN > fi-skl-6260u: PASS -> DMESG-WARN > fi-kbl-r: PASS -> DMESG-WARN > fi-bdw-5557u: PASS -> DMESG-WARN > fi-glk-j4005: PASS -> DMESG-WARN > fi-kbl-7567u: PASS -> DMESG-WARN > fi-skl-guc: PASS -> DMESG-WARN > fi-cfl-8700k: PASS -> DMESG-WARN > fi-whl-u: PASS -> DMESG-WARN > fi-cfl-guc: PASS -> DMESG-WARN > fi-bxt-j4205: PASS -> DMESG-WARN > fi-skl-6700hq: PASS -> DMESG-WARN Allocation not allowed, hmm. Let's see just how much the stack can take! -Chris
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index e82aa804cdba..222787106169 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -21,7 +21,7 @@ * IN THE SOFTWARE. * */ - + #include "i915_vma.h" #include "i915_drv.h" @@ -30,6 +30,45 @@ #include <drm/drm_gem.h> +#if defined(CONFIG_DRM_I915_DEBUG_GEM) && defined(CONFIG_DRM_DEBUG_MM) + +#include <linux/stackdepot.h> + +static void vma_print_allocator(struct i915_vma *vma, const char *reason) +{ + unsigned long entries[16]; + struct stack_trace trace = { + .entries = entries, + .max_entries = ARRAY_SIZE(entries), + }; + char *buf; + + if (!vma->node.stack) { + DRM_DEBUG_DRIVER("vma.node [%08llx + %08llx] %s: unknown owner\n", + vma->node.start, vma->node.size, reason); + return; + } + + buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!buf) + return; + + depot_fetch_stack(vma->node.stack, &trace); + snprint_stack_trace(buf, PAGE_SIZE, &trace, 0); + DRM_DEBUG_DRIVER("vma.node [%08llx + %08llx] %s: inserted at %s\n", + vma->node.start, vma->node.size, reason, buf); + + kfree(buf); +} + +#else + +static void vma_print_allocator(struct i915_vma *vma, const char *reason) +{ +} + +#endif + static void i915_vma_retire(struct i915_gem_active *active, struct i915_request *rq) { @@ -875,8 +914,10 @@ int i915_vma_unbind(struct i915_vma *vma) } GEM_BUG_ON(i915_vma_is_active(vma)); - if (i915_vma_is_pinned(vma)) + if (i915_vma_is_pinned(vma)) { + vma_print_allocator(vma, "is pinned"); return -EBUSY; + } if (!drm_mm_node_allocated(&vma->node)) return 0;
At the moment, gem_exec_gttfill fails with a sporadic EBUSY due to us wanting to unbind a pinned batch. Let's dump who first bound that vma to see if that helps us identify who still unexpectedly has it pinned. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/i915_vma.c | 45 +++++++++++++++++++++++++++++++-- 1 file changed, 43 insertions(+), 2 deletions(-)