oshmem/mca/sshmem: Fix build with --enable-mem-debug
#7926
Conversation
|
@bertwesarg We posted similar fixes. You are missing at least one place in the UCX backend, see https://github.com/open-mpi/ompi/pull/7927/files#diff-ae1fde1f4c92b725fdcf638c9bd2b73bR101 |
|
The IBM CI (GNU/Scale) build failed! Please review the log, linked below. Gist: https://gist.github.com/d320d8698a93bc0c86aeaac867ccb6a4 |
|
The IBM CI (XL) build failed! Please review the log, linked below. Gist: https://gist.github.com/8eccf19eaaee76904a277167da37f0d8 |
|
The IBM CI (PGI) build failed! Please review the log, linked below. Gist: https://gist.github.com/e80a71467c7be5ee4be2da1bcaf4731b |
`--enable-mem-debug` `#define`s `realloc`/`free` as macros, though macros are also matched if they appear in references to members. Rename the members to avoid this matching. See open-mpi#6995 Signed-off-by: Bert Wesarg <[email protected]>
bertwesarg
force-pushed
the
fix-mem-debug-build
branch
from
fb516fa to
3111877
Compare
|
@bertwesarg @devreal Can you guys figure out which of this PR and #7927 should get merged? They're fairly straightforward and have been sitting around forever. Thanks! |
|
Not sure why the build checker failed, can someone restart them please? I'd favor this PR over #7927 because the |
|
bot:retest |
|
The signed-off checker seems to be stuck. Other than that this PR is good to go |
|
bot:ompi:retest |
|
hung again :/ |
|
The bot: commands do not cause the signed-off-by checker to run again. This is a long-standing issue with GitHub that @jsquyres has been trying to figure out for some time. The only solution is to re-force push to cause the checker to refire. |
|
I asked @bertwesarg but he seems to be on vacation until the end of the week. |
|
Well this was an unanticipated side effect... Since @bertwesarg is out, and I apparently don't have write perms on the branch for this PR in his fork, I created #8107 to replace this PR (i.e., trigger all the CI again, including the signed-off-by-checker). The intent was to "just get this done" -- close out this PR and replace it with #8107. Amusingly, Github was smarter than that: Github triggered all the CI on the new PR #8107, but since the git hash on this PR is identical to the git hash on #8107, all the CI tests -- including the signed off by checker -- updated over here, too. So when CI finishes, we can just merge this PR (because there's another issue linked to it), which will also auto-close #8107. |
|
bot:aws:recheck |
|
bot:ompi:retest |
|
Looks like, this missed the 4.1.0 release. |
7 participants
--enable-mem-debug#definesrealloc/freeas macros, though macrosare also matched if they appear in references to members. Rename the
members to avoid this matching.
See #6995