On Sat, Jul 27, 2013 at 03:09:48PM +0200, David Herrmann wrote: > Hi > > On Sat, Jul 27, 2013 at 3:06 PM, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote: > > On Sat, Jul 27, 2013 at 01:38:42PM +0200, David Herrmann wrote: > >> i915 is the last user of the weird search+get_block drm_mm API. Convert it > >> to an explicit kmalloc()+insert_node(). This drops the last user of the > >> node-cache in drm_mm. We can remove it now in a follow-up patch. > >> > >> v2: > >> - simplify error path in i915_setup_compression() > > > > You only applied it to the err path, I was expecting err_fb to do the > > kfree(compressed_llb) as well. Feel free to rename those to err_fb and > > err_llb respectively if you think that helps. > > I wasn't sure about that second error-path. Some kernel subsystems > tend to inline error-paths that are only taken once (which is the case > for compressed_llb). But if that's not the case for DRM, I will surely > fix that, too. My concern in this case is to keep the two path similar; we use the same allocation style so we should use the same cleanup. Otherwise, the next time I look at the code, I wil wonder why they are different. :) -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel