patch to fix PR83712
Vladimir Makarov
vmakarov@redhat.com
Sat Mar 10 16:40:00 GMT 2018
A few people reported that the patch broke i686. I am going to work on
the patch more. Meanwhile I've reverted the patch.
On 03/09/2018 11:16 AM, Vladimir Makarov wrote:
> The following patch fixes
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83712
>
> It is another "cannot find a spill reg for reload" problem. LRA has
> already a code splitting hard reg live ranges to avoid such problem.Â
> This code is in LRA inheritance pass. Unfortunately, the code does
> splitting for small class pseudos only. This PR is a more complicated
> code and it is hard to adapt the inheritance sub-pass to reliably
> solve such problems.
>
> To fix the PR, I added a sub-pass which works in very rare cases after
> we already found that we have no hard regs for a reload pseudo. It
> tries to split a hard reg live range for the pseudo. After that it
> tries again to assign a hard reg to the pseudo. The patch changes
> LRA-subpass flow for this. I hope that the patch will finally solved
> all such problems but I am not sure to be completely certain.
>
> The patch was bootstrapped and tested on x86-64 and ppc64.
>
> Committed as rev. 258390.
>
More information about the Gcc-patches
mailing list