Re: [PATCH] x86: Compile some check-function-bodies tests with -mtune=generic
"H.J. Lu" <[email protected]>
| Newsgroups | gmane.comp.gcc.patches |
|---|---|
| Message-ID | <CAMe9rOpKGKwLe=e8Q4A7k20c43+mdD-B9JH6MPToDZxbQ8wiPw@mail.gmail.com> |
On Thu, Aug 20, 2026 at 9:14 AM Liu, Hongtao <[email protected]> wrote: > > > > > -----Original Message----- > > From: H.J. Lu <[email protected]> > > Sent: Thursday, August 20, 2026 7:52 AM > > To: GCC Patches <[email protected]>; Uros Bizjak > > <[email protected]>; Liu, Hongtao <[email protected]> > > Subject: [PATCH] x86: Compile some check-function-bodies tests with - > > mtune=generic > > > > Compile some check-function-bodies tests with -mtune=generic to avoid > > > > FAIL: gcc.target/i386/memset-pr120683-10.c check-function-bodies foo > > FAIL: gcc.target/i386/memset-pr120683-12.c check-function-bodies foo > > FAIL: gcc.target/i386/pr122343-1b.c check-function-bodies foo > > FAIL: gcc.target/i386/pr122343-5b.c check-function-bodies foo > > > > due to the generated sequence > > > > movq $0, 29(%rdi) > > movq $0, 37(%rdi) > > movq $0, 45(%rdi) > > movq $0, 53(%rdi) > > movq $0, (%rdi) > > movq $0, 8(%rdi) > > movq $0, 16(%rdi) > > movq $0, 24(%rdi) > > > > vs the expected sequence > > > > movq $0, 48(%rdi) > > movq $0, (%rdi) > > movq $0, 8(%rdi) > > movq $0, 16(%rdi) > > movq $0, 24(%rdi) > > movq $0, 32(%rdi) > > movq $0, 40(%rdi) > > movq $0, 53(%rdi) > > > > when a different tuning is enabled by default. > Can we check movq $0, [1-9]*(%rdi), use regex for the disp? regex won't check random instructions added by some commits and there are other different code sequences generated by different default tuning which doesn't work with regex. > > > > PR testsuite/126923 > > * gcc.target/i386/memset-pr120683-10.c (dg-options): Add -mtune=generic. > > * gcc.target/i386/memset-pr120683-12.c (dg-options): Likewise. > > * gcc.target/i386/pr122343-1b.c (dg-options): Likewise. > > * gcc.target/i386/pr122343-1b.c (dg-options): Likewise. > > > > -- > > H.J. -- H.J.