This patch fixes a NULL pointer dereference in a test case which allocates piles of buffers without ever freeing them, and repeatedly tries to upload/fetch data to/from those buffers, without doing any other work (such as rendering).
The same problem exists in a variety of other places, some of which are more difficult to recover from. This patch does not try to fix any of those. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=82683 Signed-off-by: Kenneth Graunke <kenn...@whitecape.org> --- src/mesa/drivers/dri/i965/intel_buffer_objects.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/src/mesa/drivers/dri/i965/intel_buffer_objects.c b/src/mesa/drivers/dri/i965/intel_buffer_objects.c index 96dacde..d063d92 100644 --- a/src/mesa/drivers/dri/i965/intel_buffer_objects.c +++ b/src/mesa/drivers/dri/i965/intel_buffer_objects.c @@ -252,7 +252,10 @@ intel_bufferobj_subdata(struct gl_context * ctx, if (brw->has_llc) { if (offset + size <= intel_obj->gpu_active_start || intel_obj->gpu_active_end <= offset) { - drm_intel_gem_bo_map_unsynchronized(intel_obj->buffer); + if (drm_intel_gem_bo_map_unsynchronized(intel_obj->buffer) < 0) { + _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBufferSubData (map failed)"); + return; + } memcpy(intel_obj->buffer->virtual + offset, data, size); drm_intel_bo_unmap(intel_obj->buffer); -- 2.1.0 _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
