[Intel-gfx] i915 vs checkpatch

Arkadiusz Hiler arkadiusz.hiler at intel.com
Thu Mar 1 09:47:06 UTC 2018


Hey all,

Since not so long ago our CI is running and reporting sparse and
checkpatch. Sparse is doing just fine but I had to disable checkpatch
for the time being - too much "false" positives causing people to
complain. It's simply confusing to see one thing in the code, and
fitting your change in only to get a report that it's wrong.

We are explicitly going against couple of the recommendations it tries
to enforce, e.g. not using BIT macro, splitting quoted strings:
https://lists.freedesktop.org/archives/intel-gfx/2018-February/156613.html

IMHO we should make a couple of decisions here:
 1. Do we really want to use the checkpatch / have CI reports?
 2. Which of the checkpatch checks we want to disabled for i915?
 3. How strongly do we want to enforce the rest?
 4. Do we want to change what's already in the tree, for compliance?

Recent-ish drm-tip, vanilla checkpatch on i915 reports:
total: 399 errors, 3573 warnings, 209374 lines checked

-- 
Cheers,
Arek




More information about the Intel-gfx mailing list