[Mesa-dev] [Bug 109532] ir_variable has maximum access out of bounds -- but it's not out of bounds

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Sat Feb 23 01:57:39 UTC 2019


https://bugs.freedesktop.org/show_bug.cgi?id=109532

--- Comment #38 from Ilia Mirkin <imirkin at alum.mit.edu> ---
(In reply to asimiklit from comment #36)
> Ilia could you please share your thoughts about this solution?
> Do you agree with Ian suggestion as to disallow the elimination from the
> beginning of the array?
> I am asking about it to understand have we consensus as to a direction of
> the fix or not.

Sounds good to me... I think you can do something like

layout(binding=n) foo[6]

means that n..n+5 are all taken and assigned to the relevant items.

However if you just have "foo[6]", without the explicit binding, you can (and
should) eliminate any pre- and post- elements, i.e. before the first and after
the last. However if foo[1] and foo[5] are used, then they should be 4 binding
points away from one another.

Hope that makes sense.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20190223/5e19da3f/attachment.html>


More information about the mesa-dev mailing list