From: Rafał Sapała <rafal.a.sap...@intel.com> It is possible to hit a race condition in create_from_prime, when trying to import a BO that's currently being freed. In case of prime sharing we'll succesfully get a handle, but fail on get_tiling call, potentially confusing the caller (and requiring different locking scheme than with sharing using flink). Wrap fd_to_handle with struct_mutex to force a more consistent behaviour between prime/flink, convert fprintf to DBG when handling errors.
Testcase: igt/drm_import_export/import-close-race-prime Signed-off-by: Rafał Sapała <rafal.a.sap...@intel.com> Signed-off-by: Michał Winiarski <michal.winiar...@intel.com> --- intel/intel_bufmgr_gem.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/intel/intel_bufmgr_gem.c b/intel/intel_bufmgr_gem.c index b1c3b3a..ed4ffd2 100644 --- a/intel/intel_bufmgr_gem.c +++ b/intel/intel_bufmgr_gem.c @@ -2728,14 +2728,19 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s struct drm_i915_gem_get_tiling get_tiling; drmMMListHead *list; + pthread_mutex_lock(&bufmgr_gem->lock); ret = drmPrimeFDToHandle(bufmgr_gem->fd, prime_fd, &handle); + if (ret) { + DBG("create_from_prime: failed to obtain handle from fd: %s\n", strerror(errno)); + pthread_mutex_unlock(&bufmgr_gem->lock); + return NULL; + } /* * See if the kernel has already returned this buffer to us. Just as * for named buffers, we must not create two bo's pointing at the same * kernel object */ - pthread_mutex_lock(&bufmgr_gem->lock); for (list = bufmgr_gem->named.next; list != &bufmgr_gem->named; list = list->next) { @@ -2747,12 +2752,6 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s } } - if (ret) { - fprintf(stderr,"ret is %d %d\n", ret, errno); - pthread_mutex_unlock(&bufmgr_gem->lock); - return NULL; - } - bo_gem = calloc(1, sizeof(*bo_gem)); if (!bo_gem) { pthread_mutex_unlock(&bufmgr_gem->lock); @@ -2793,6 +2792,7 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s DRM_IOCTL_I915_GEM_GET_TILING, &get_tiling); if (ret != 0) { + DBG("create_from_prime: failed to get tiling: %s\n", strerror(errno)); drm_intel_gem_bo_unreference(&bo_gem->bo); return NULL; } -- 2.4.3 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx