On Sat, Jan 11, 2025 at 1:11 PM Jeff Law <jeffreya...@gmail.com> wrote: > > > > On 1/11/25 2:08 PM, Andrew Pinski (QUIC) wrote: > >> -----Original Message----- > >> From: Jeff Law <jeffreya...@gmail.com> > >> Sent: Saturday, January 11, 2025 8:12 AM > >> To: Andrew Pinski (QUIC) <quic_apin...@quicinc.com>; gcc- > >> patc...@gcc.gnu.org > >> Subject: Re: [PATCH] final: Fix get_attr_length for asm goto > >> [PR118411] > >> > >> > >> > >> On 1/11/25 2:52 AM, Andrew Pinski wrote: > >>> The problem is for inline-asm goto, the outer rtl insn type is > >> a > >>> jump_insn and get_attr_length does not handle ASM > >> specially unlike if > >>> the outer rtl insn type was just insn. > >>> > >>> This fixes the issue by adding support for both CALL_INSN > >> and > >>> JUMP_INSN with asm. > >>> > >>> OK? Bootstrapped and tested on x86_64-linux-gnu. > >>> > >>> PR middle-end/118411 > >>> > >>> gcc/ChangeLog: > >>> > >>> * final.cc (get_attr_length_1): Handle asm for CALL_INSN > >>> and JUMP_INSNs. > >> OK > > > > Can I apply this to both the GCC 14, 13 and 12 branches too? Even though > > this is not a regression; it seems like an important bug fix for asm goto. > > Especially since the Linux kernel uses it and might run into this issue. > Seems pretty safe to me and while not a regression, it's a case where we > generate incorrect output. So yea, fine by me. Consider giving it a > week or so on the trunk just in case...
Applied finally to the open branches. > > jeff >
