Gotcha! The command that overwrote the batch was an XY_COLOR_BLT with RGB-only writes (no alpha). This is only used by intelClearWithBlit for GL_DEPTH_BUFFER (no stencil) clears. Looking over the code, the implication is that (x1, y1), (x2, y2) are beyond the bounds of the buffer.
diff --git a/src/mesa/drivers/dri/intel/intel_blit.c b/src/mesa/drivers/dri/intel/intel_blit.c index 1369e63..6269673 100644 --- a/src/mesa/drivers/dri/intel/intel_blit.c +++ b/src/mesa/drivers/dri/intel/intel_blit.c @@ -288,6 +288,8 @@ intelClearWithBlit(struct gl_context *ctx, GLbitfield mask) y1 = cy + irb->draw_y; x2 = cx + cw + irb->draw_x; y2 = cy + ch + irb->draw_y; + assert(x1 >= 0 && y1 >= 0); + assert(x2 <= region->pitch && y2 * region->pitch * region->cpp <= region->bo->size); pitch = region->pitch; cpp = region->cpp; -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/980325 Title: [i965gm] False GPU lockup render.IPEHR: 0x7b000000 To manage notifications about this bug go to: https://bugs.launchpad.net/xserver-xorg-video-intel/+bug/980325/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs