diff options
| author | H.J. Lu <hjl.tools@gmail.com> | 2021-03-05 07:26:42 -0800 |
|---|---|---|
| committer | H.J. Lu <hjl.tools@gmail.com> | 2021-03-29 07:40:17 -0700 |
| commit | 7ebba91361badf7531d4e75050627a88d424872f (patch) | |
| tree | d99781a37b47b95441ad358d119ec3741960d405 /sysdeps/x86_64/sysdep.h | |
| parent | 91264fe3577fe887b4860923fa6142b5274c8965 (diff) | |
| download | glibc-7ebba91361badf7531d4e75050627a88d424872f.tar.gz | |
x86-64: Add AVX optimized string/memory functions for RTM
Since VZEROUPPER triggers RTM abort while VZEROALL won't, select AVX
optimized string/memory functions with
xtest
jz 1f
vzeroall
ret
1:
vzeroupper
ret
at function exit on processors with usable RTM, but without 256-bit EVEX
instructions to avoid VZEROUPPER inside a transactionally executing RTM
region.
Diffstat (limited to 'sysdeps/x86_64/sysdep.h')
| -rw-r--r-- | sysdeps/x86_64/sysdep.h | 22 |
1 files changed, 22 insertions, 0 deletions
diff --git a/sysdeps/x86_64/sysdep.h b/sysdeps/x86_64/sysdep.h index d07b8f0aaf..7bebdeb210 100644 --- a/sysdeps/x86_64/sysdep.h +++ b/sysdeps/x86_64/sysdep.h @@ -95,6 +95,28 @@ lose: \ #define R14_LP r14 #define R15_LP r15 +/* Zero upper vector registers and return with xtest. NB: Use VZEROALL + to avoid RTM abort triggered by VZEROUPPER inside transactionally. */ +#define ZERO_UPPER_VEC_REGISTERS_RETURN_XTEST \ + xtest; \ + jz 1f; \ + vzeroall; \ + ret; \ +1: \ + vzeroupper; \ + ret + +/* Zero upper vector registers and return. */ +#ifndef ZERO_UPPER_VEC_REGISTERS_RETURN +# define ZERO_UPPER_VEC_REGISTERS_RETURN \ + VZEROUPPER; \ + ret +#endif + +#ifndef VZEROUPPER_RETURN +# define VZEROUPPER_RETURN VZEROUPPER; ret +#endif + #else /* __ASSEMBLER__ */ /* Long and pointer size in bytes. */ |