[Mesa-dev] [PATCH 13/14] nir: propagate bitsize information in nir_search

[Iago Toral]

On Wed, 2016-03-16 at 09:48 +0100, Samuel Iglesias Gonsálvez wrote:
> 
> On 15/03/16 08:41, Iago Toral wrote:
> > On Mon, 2016-03-14 at 17:33 -0700, Jason Ekstrand wrote:
> [...]
> >> +   nir_alu_type type; + union { -      uint32_t u; -
> >> int32_t i; -      float f; +      uint64_t u; +      int64_t i; +
> >> double d; } data; } nir_search_constant;
> > 
> > Maybe we should rename these to u64, i64 as f64 like you suggested
> > for a similar case in another patch?
> > 
> 
> We decided to do the following changes to this patch series (among the
> others already agreed): rename these fields and add the inline helper
> functions to get type size and base type from nir_alu_type enum
> values. So the rest of the fp64 patches will have those changes
> already and everything would be easier to maintain.
> 
> One question for Jason: Should we squash all the patches of this
> series in a single commit as originally proposed? I'd rather have them
> separately so we understand what each patch does in case of a git
> bisect, however I don't have a strong opinion against squashing.

At the very least we should squash the one for the copy-propagation pass
with the new opcodes, otherwise we'd break the build.

The main concern with squashing everything is that it is going to be a
huge patch and it is going to be difficult to make sense of it like that
for anyone who tracks something down to it, but otherwise we are going
to break some things in between which is also not ideal...

Iago