-
Notifications
You must be signed in to change notification settings - Fork 1.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
RFC/WIP: Revert Consolidate arc_buf allocation checks #12227
RFC/WIP: Revert Consolidate arc_buf allocation checks #12227
Conversation
This revert should be safe and I think doing so would be a good idea. This isn't a change we need, and nothing has really fundamentally changed in the way this area of the code was intended to work. The only wrinkle with reverting this you've already addressed and that was the zstd conflict. Thanks for keeping those patches separate at least for the moment since it made it easier to verify it was a clean revert. The minor cstyle error will need to be sorted but aside from that this functionally looks good. |
Great! Do you want me to squash these, or just add more discussion in the commits (and rebase)? |
@aerusso why don't we keep them split up for now. Can you just update your top commit to resolve the csyle warning it introduced and update the PR. I'll see about rounding up some more reviewers. When we merge this I can squash everything. |
In preparation for reverting 13fac09, we partially revert 10b3c7f, within dbuf_alloc_arcbuf_from_arcbuf, which otherwise would conflict. Signed-off-by: Antonio Russo <[email protected]>
This reverts commit 13fac09. Per the discussion in openzfs#11531, the reverted commit---which intended only to be a cleanup commit---introduced a subtle, unintended change in behavior. Suggested-by: @chrisrd Suggested-by: [email protected] Signed-off-by: Antonio Russo <[email protected]>
eba0e14
to
a87bc85
Compare
This commit reverts the partial reversion of 10b3c7f, transposed to account for the un-consolidation of dbuf_alloc_arcbuf_from_arcbuf. Signed-off-by: Antonio Russo <[email protected]>
a87bc85
to
ad96131
Compare
This reverts commit 13fac09. Per the discussion in #11531, the reverted commit---which intended only to be a cleanup commit---introduced a subtle, unintended change in behavior. Care was taken to partially revert and then reapply 10b3c7f which would otherwise have caused a conflict. These changes were squashed in to this commit. Reviewed-by: Brian Behlendorf <[email protected]> Suggested-by: @chrisrd Suggested-by: [email protected] Signed-off-by: Antonio Russo <[email protected]> Closes #11531 Closes #12227
This reverts commit 13fac09. Per the discussion in openzfs#11531, the reverted commit---which intended only to be a cleanup commit---introduced a subtle, unintended change in behavior. Care was taken to partially revert and then reapply 10b3c7f which would otherwise have caused a conflict. These changes were squashed in to this commit. Reviewed-by: Brian Behlendorf <[email protected]> Suggested-by: @chrisrd Suggested-by: [email protected] Signed-off-by: Antonio Russo <[email protected]> Closes openzfs#11531 Closes openzfs#12227
This reverts commit 13fac09. Per the discussion in #11531, the reverted commit---which intended only to be a cleanup commit---introduced a subtle, unintended change in behavior. Care was taken to partially revert and then reapply 10b3c7f which would otherwise have caused a conflict. These changes were squashed in to this commit. Reviewed-by: Brian Behlendorf <[email protected]> Suggested-by: @chrisrd Suggested-by: [email protected] Signed-off-by: Antonio Russo <[email protected]> Closes #11531 Closes #12227
This reverts commit 13fac09. Per the discussion in #11531, the reverted commit---which intended only to be a cleanup commit---introduced a subtle, unintended change in behavior. Care was taken to partially revert and then reapply 10b3c7f which would otherwise have caused a conflict. These changes were squashed in to this commit. Reviewed-by: Brian Behlendorf <[email protected]> Suggested-by: @chrisrd Suggested-by: [email protected] Signed-off-by: Antonio Russo <[email protected]> Closes #11531 Closes #12227
Motivation and Context
The discussion in #11531 suggests that 13fac09 may be implicated in that bug.
I would like comments on the safety of revert the aforementioned commit.
Description
Because 10b3c7f also touches the same code, this patch set carefully reverts 10b3c7f (limited to
module/zfs/dbuf.c
), then actually reverts 13fac09, and then finally reverts the first commit.How Has This Been Tested?
It passes the ZTS when rebased on 2.0.4 and the MR for 2.0.5.
Types of changes
Checklist:
Signed-off-by
.