пт, 6 дек. 2024 г. в 19:32, Vladimir Makarov :
>
> The proposed patch can be a fix and you can commit it. The only request is
> not to close PR for now.
>
> LRA rematerialization sub-pass rematerializes insn containing only pseudos
> assigned to hard regs and should not change live-range of spil
The proposed patch can be a fix and you can commit it. The only request
is not to close PR for now.
LRA rematerialization sub-pass rematerializes insn containing only
pseudos assigned to hard regs and should not change live-range of
spilling pseudos. So sentence "Rematerialization sometimes
Denis Chertykov writes:
> The fix for PR116778:
>
Added Vlad to CC.
> [...]
>
> diff --git a/gcc/lra-lives.cc b/gcc/lra-lives.cc
> index 49134ade713..510f7d927ab 100644
> --- a/gcc/lra-lives.cc
> +++ b/gcc/lra-lives.cc
> @@ -62,9 +62,10 @@ int lra_hard_reg_usage[FIRST_PSEUDO_REGISTER];
> /* A
The fix for PR116778:
Brief:
The bug appears in LRA after rematerialization pass while creating live ranges.
File lra.cc:
*
/* Now we know what pseudos should be spilled. Try to
rematerialize them first. */
if (lra
