[Mesa-dev] [PATCH] glsl/linker: Check the subroutine associated functions names

[Vadim Shovkoplias]

Hi Ian,

Thanks for the explanation. I've just sent a patch for review with the
compile time check https://patchwork.freedesktop.org/patch/255542/

Regards,
Vadym

пт, 5 окт. 2018 г. в 16:55, Ian Romanick <idr at freedesktop.org>:

> On 10/03/2018 11:52 PM, Iago Toral wrote:
> > On Wed, 2018-10-03 at 16:24 +0300, Vadim Shovkoplias wrote:
> >> Hi Iago,
> >>
> >> I also think that compiler is the best place for the fix. But from my
> >> understanding compiler fix will be limited to distinct shader objects
> >> (not the shader stage).
> >> In GLSL spec mentioned: "A program will fail to compile or link if any
> >> *shader or stage* contains two or more functions with the same name if
> >> the name is
> >> associated with a subroutine type".
> >> So if I understand the spec correctly such restriction for the shader
> >> stage can be fixed only in linker. E.g. consider the use case when
> >> fragment shader is linked from
> >> two fragment shader objects. One object contains function foo() which
> >> is associated with subroutine type and second shader object has some
> >> regular foo() function
> >> with the same name.  I suppose compiler fix won't be able to detect
> this.
> >
> > Yes, that is correct, good point.
> >
> >> IMHO the best way to fix this is to implement 2 patches:
> >> 1) One patch for linker to implement restriction for the shader stages
> >> (already done in this patch)
> >> 2) Another one for compiler to check the restriction for distinct
> >> shader objects.
> >>
> >> What do you think ?
> >
> > Since we need to keep the link-time check, any compile-time checks we
> > add would be redundant with that, so I think we should just go with the
> > link-time check only (your original patch)
>
> The spec quotation is pretty clear: "A program will fail to compile ...
> if any shader ... contains two or more functions with the same name if
> the name is associated with a subroutine type."  To be strictly correct,
> both checks are necessary.  Before doing that extra work, someone should
> check that glslang has the compile-time check.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20181009/6d3ecba4/attachment.html>