[Bug target/48227] "rep ret" generated for -march=core2

[pinskia at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=48227

Andrew Pinski  changed:

   What|Removed |Added

 Status|UNCONFIRMED |RESOLVED
 Resolution|--- |FIXED
   Target Milestone|--- |8.0
   See Also||https://gcc.gnu.org/bugzill
   ||a/show_bug.cgi?id=81616

--- Comment #6 from Andrew Pinski  ---
/* X86_TUNE_PAD_RETURNS: Place NOP before every RET that is a destination
   of conditional jump or directly preceded by other jump instruction.
   This is important for AND K8-AMDFAM10 because the branch prediction
   architecture expect at most one jump per 2 byte window.  Failing to
   pad returns leads to misaligned return stack.  */
DEF_TUNE (X86_TUNE_PAD_RETURNS, "pad_returns",
  m_ATHLON_K8 | m_AMDFAM10)

Core2 issue was Fixed by r0-125351 for 4.9.0.

GCC 8 removed it even for generic by r8-5077.

So all fixed.

[Bug target/48227] "rep ret" generated for -march=core2

[matt at godbolt dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=48227

--- Comment #5 from Matt Godbolt  ---
Seems to have been fixed in 4.9

[Bug target/48227] rep ret generated for -march=core2

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48227

--- Comment #4 from Andrew Pinski pinskia at gcc dot gnu.org 2011-12-15 
23:18:35 UTC ---
  /* X86_TUNE_PAD_RETURNS */
  m_CORE2I7 | m_AMD_MULTIPLE | m_GENERIC,


So we still do it for Core2 i7.

[Bug target/48227] rep ret generated for -march=core2

[zuxy.meng at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48227

--- Comment #3 from Zuxy zuxy.meng at gmail dot com 2011-04-06 13:42:21 UTC 
---
(In reply to comment #1)
 A good question is does it make a difference in actual performance numbers, it
 might still make a positive difference.  Until someone tries it out and sees
 the difference in performance, I am going to say we should keep it.

Actually GCC since 4.6 has stopped generating 'rep ret' for all Intel CPUs.

[Bug target/48227] rep ret generated for -march=core2

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48227

Andrew Pinski pinskia at gcc dot gnu.org changed:

   What|Removed |Added

 Target||x86_64-linux-gnu

--- Comment #1 from Andrew Pinski pinskia at gcc dot gnu.org 2011-04-01 
19:20:52 UTC ---
A good question is does it make a difference in actual performance numbers, it
might still make a positive difference.  Until someone tries it out and sees
the difference in performance, I am going to say we should keep it.

[Bug target/48227] rep ret generated for -march=core2

[ian at airs dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48227

--- Comment #2 from Ian Lance Taylor ian at airs dot com 2011-04-02 00:55:40 
UTC ---
If it doesn't make a difference in performance, we should get rid of it, so
that we can save a byte of code.