[Bug target/61193] New: ABI incompatibility between POWER and Z HTM builtins and intrinsics

[Bug target/61193] New: ABI incompatibility between POWER and Z HTM builtins and intrinsics

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61193

            Bug ID: 61193
           Summary: ABI incompatibility between POWER and Z HTM builtins
                    and intrinsics
           Product: gcc
           Version: 4.8.3
            Status: UNCONFIRMED
          Severity: blocker
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: bergner at gcc dot gnu.org

An ABI incompatibility between POWER and Z HTM builtins and intrinsics has been
found.  The incompatibility is on mainline, 4.9 and 4.8.  We're still waiting
to hear what the XL team has implemented for their return status.  Does the XL
on POWER and Z match the POWER or Z sematics?  Does XL on POWER match GCC on
POWER and XL on Z match GCC on Z?  We're tracking that answer down now.  The
intent was that common code on all platforms and compilers would have the same
ABI.

[Bug target/61193] ABI incompatibility between POWER and Z HTM builtins and intrinsics

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61193

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dje at gcc dot gnu.org,
                   |                            |jakub at gcc dot gnu.org,
                   |                            |krebbel at gcc dot gnu.org,
                   |                            |rguenth at gcc dot gnu.org
   Target Milestone|---                         |4.8.3

[Bug target/61193] ABI incompatibility between POWER and Z HTM builtins and intrinsics

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61193

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |powerpc*-linux s390-linux
           Priority|P3                          |P1

[Bug target/61193] ABI incompatibility between POWER and Z HTM builtins and intrinsics

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61193

--- Comment #2 from Peter Bergner <bergner at gcc dot gnu.org> ---
Author: bergner
Date: Fri May 16 03:44:19 2014
New Revision: 210487

URL: http://gcc.gnu.org/viewcvs?rev=210487&root=gcc&view=rev
Log:
    PR target/61193
    * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
    (__TM_simple_begin): Use it.
    (__TM_begin): Likewise.

Modified:
    branches/gcc-4_9-branch/gcc/ChangeLog
    branches/gcc-4_9-branch/gcc/config/rs6000/htmxlintrin.h

[Bug target/61193] ABI incompatibility between POWER and Z HTM builtins and intrinsics

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61193

--- Comment #3 from Peter Bergner <bergner at gcc dot gnu.org> ---
Author: bergner
Date: Fri May 16 03:46:23 2014
New Revision: 210488

URL: http://gcc.gnu.org/viewcvs?rev=210488&root=gcc&view=rev
Log:
    Backport from mainline
    2014-05-15  Peter Bergner  <bergner@vnet.ibm.com>

    PR target/61193
    * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
    (__TM_simple_begin): Use it.
    (__TM_begin): Likewise.

Modified:
    branches/gcc-4_8-branch/gcc/ChangeLog
    branches/gcc-4_8-branch/gcc/config/rs6000/htmxlintrin.h

[Bug target/61193] ABI incompatibility between POWER and Z HTM builtins and intrinsics

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61193

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2014-05-16
           Assignee|unassigned at gcc dot gnu.org      |bergner at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #4 from Peter Bergner <bergner at gcc dot gnu.org> ---
After discussing this with the XL compiler team, the S390 GCC team and
the Power GCC team, we have decided to leave the incompatibility between
Power and S390.  However, the XL and GCC compilers will be compatible
with each other when targeting the same processor target.  To mitigate
the incompatibility somewhat, we have decided to add a macro to the
powerpc*-linux's htmintrin.h file that defines what the "sucessful"
return status value of the __TM_simple_begin() and __TM_begin() intrinsic
function is.  This macro is already defined in the S390's htmintrin.h
header file and is used by the S390 to determine whether the transaction
was successfully started or not.  By adding the same macro on Power, we
allow users to write common code between Power and S390, even though our
successful return status values are different.

For example, the following code can be used on Power and S390, even
though the actual value returned by __TM_simple_begin() is different.

  if ((tx_state = __TM_simple_begin ()) == _HTM_TBEGIN_STARTED)
    {
      /* Transaction State Initiated.  */
      ...
    }
  else
    {
      /* Transaction State Failed.  */
      ...
    }

[Bug target/61193] ABI incompatibility between POWER and Z HTM builtins and intrinsics

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61193

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
                URL|                            |https://gcc.gnu.org/ml/gcc-
                   |                            |patches/2014-05/msg01239.ht
                   |                            |ml
         Resolution|---                         |FIXED

--- Comment #5 from Peter Bergner <bergner at gcc dot gnu.org> ---
Fixes committed to mainline and the 4.9 and 4.8 branches.