[Mesa-stable] [Mesa-dev] [PATCH 01/12] i965/fs: Unpack count argument to 64-bit shift ops on Atom

Matt Turner mattst88 at gmail.com
Wed Nov 29 02:53:18 UTC 2017


On Tue, Nov 28, 2017 at 12:35 PM, Ian Romanick <idr at freedesktop.org> wrote:
> Also... it looks like only a few patches in this series got reviewed.
> Did the whole thing actually land?

Yes, I felt safe pushing the majority of it without review because of
the extensive unit tests I wrote.

I also felt justified in pushing the majority of it without review
because like how no one reviews piglit patches, no one reviews
validation code in brw_eu_validate.c or its unit tests in
test_eu_validate.cpp.

The series adding the first non-trivial validation code was sent in
Nov 2016 (See git log --stat --pretty=fuller
f57bdd48498f~4..f57bdd48498f). The first few simple patches were
reviewed relatively quickly, but two months later without comment (and
a number of in-person requests for review) I decided to just push the
actual meat of the series without review. Again, I felt safe doing
that because of the extensive unit tests. The lack of regressions, but
more so the problems caught and prevented by this code since it landed
confirm to me that it was the right decision.

I get not wanting to review tedious code like this. Believe me, it was
tedious to write! I've just accepted that patches to this code will
not be reviewed. That's okay by me. The unit tests probably do a
better job at shaking out bugs than a reviewer anyway.


More information about the mesa-stable mailing list