Code review comment for lp:~kdub/mir/gbm-ext-v2

5) If we add const to this extension, we have some MirBuffer-related functions that have const and others that don't. This seems worse (in terms of update cost or mismatched signatures) than leaving as-is.
6) It is a precondition that the MirBuffer is a a GBM buffer. Stronger typing would get us in trouble with how the buffers get submitted.
7) Stride for buffer was rejected. https://code.launchpad.net/~kdub/mir/mir-buffer-get-stride/+merge/303135
8) Sounds like its alright as it is then? The mesa driver makes use of it now, if the mesa driver implements that logic, it can go away.