8.3: Add 0030-linstor-Fix-to-add-sync-to-DD.patch #70
Conversation
SOURCES/0002-feat-drivers-add-CephFS-and-GlusterFS-drivers.patch
Outdated
Show resolved
Hide resolved
| @@ -74,6 +74,7 @@ Patch1026: 0026-feat-LVHDSR-add-a-way-to-modify-config-of-LVMs-60.patch | |||
| Patch1027: 0027-reflect-upstream-changes-in-our-tests.patch | |||
| Patch1028: 0028-CA-398425-correctly-check-for-multiple-targets-in-iS.patch | |||
| Patch1029: 0029-Synchronization-with-8.2-LINSTOR-before-a-stable-rel.patch | |||
| Patch1030: 0030-linstor-Fix-to-add-sync-to-DD.patch | |||
There was a problem hiding this comment.
Does this patch affect code which is common with other drivers, or is it code which only involves linstor?
There was a problem hiding this comment.
A few other drivers use this function, we will test the behavior with @Wescoeur before merging in sm.
It only make it so DD sync his block accesses before finishing.
Ensuring there is no inflight requests when dd has finished.
It should be a pretty minor behavior change.
There was a problem hiding this comment.
So it's something we might contribute upstream, too?
For what operations do drivers use dd?
There was a problem hiding this comment.
It's used to zero the footer of VHD specifically in the GC.
It is specifically an issue for a problem with the linstor driver.
There was a problem hiding this comment.
Will it have performance implications for other drivers, since it will force any current cache to be written? I'm not fond of diverging from upstream for upstream drivers.
Signed-off-by: Damien Thenot <[email protected]>
Nambrok
force-pushed
the
8.3-fix-dd-sync
branch
from
c52b72e to
1e9d2b6
Compare
Koji build(v8.3-v-linstor-testing, scratch): https://koji.xcp-ng.org/taskinfo?taskID=71212