Hi Sasha,
On Thu, Dec 20, 2018 at 07:26:15PM +0000, Sasha Levin wrote:
Hi,
[This is an automated email]
This commit has been processed because it contains a "Fixes:" tag, fixing commit: 432c6bacbd0c MIPS: Use per-mm page to execute branch delay slot instructions.
The bot has tested the following trees: v4.19.10, v4.14.89, v4.9.146,
Neat! I like the idea of this automation :)
v4.19.10: Build OK! v4.14.89: Build OK! v4.9.146: Failed to apply! Possible dependencies: 05ce77249d50 ("userfaultfd: non-cooperative: add madvise() event for MADV_DONTNEED request") 163e11bc4f6e ("userfaultfd: hugetlbfs: UFFD_FEATURE_MISSING_HUGETLBFS") 67dece7d4c58 ("x86/vdso: Set vDSO pointer only after success") 72f87654c696 ("userfaultfd: non-cooperative: add mremap() event") 893e26e61d04 ("userfaultfd: non-cooperative: Add fork() event") 897ab3e0c49e ("userfaultfd: non-cooperative: add event for memory unmaps") 9cd75c3cd4c3 ("userfaultfd: non-cooperative: add ability to report non-PF events from uffd descriptor") d811914d8757 ("userfaultfd: non-cooperative: rename *EVENT_MADVDONTNEED to *EVENT_REMOVE")
This list includes the correct soft dependency - commit 897ab3e0c49e ("userfaultfd: non-cooperative: add event for memory unmaps") which added an extra argument to mmap_region().
How should we proceed with this patch?
The backport to v4.9 should simply drop the last argument (NULL) in the call to mmap_region().
Is there some way I can indicate this sort of thing in future patches so that the automation can spot that I already know it won't apply cleanly to a particular range of kernel versions? Or even better, that I could indicate what needs to be changed when backporting to those kernel versions?
Thanks, Paul
On Fri, Dec 21, 2018 at 09:16:37PM +0000, Paul Burton wrote:
Hi Sasha,
On Thu, Dec 20, 2018 at 07:26:15PM +0000, Sasha Levin wrote:
Hi,
[This is an automated email]
This commit has been processed because it contains a "Fixes:" tag, fixing commit: 432c6bacbd0c MIPS: Use per-mm page to execute branch delay slot instructions.
The bot has tested the following trees: v4.19.10, v4.14.89, v4.9.146,
Neat! I like the idea of this automation :)
Thank you! :)
v4.19.10: Build OK! v4.14.89: Build OK! v4.9.146: Failed to apply! Possible dependencies: 05ce77249d50 ("userfaultfd: non-cooperative: add madvise() event for MADV_DONTNEED request") 163e11bc4f6e ("userfaultfd: hugetlbfs: UFFD_FEATURE_MISSING_HUGETLBFS") 67dece7d4c58 ("x86/vdso: Set vDSO pointer only after success") 72f87654c696 ("userfaultfd: non-cooperative: add mremap() event") 893e26e61d04 ("userfaultfd: non-cooperative: Add fork() event") 897ab3e0c49e ("userfaultfd: non-cooperative: add event for memory unmaps") 9cd75c3cd4c3 ("userfaultfd: non-cooperative: add ability to report non-PF events from uffd descriptor") d811914d8757 ("userfaultfd: non-cooperative: rename *EVENT_MADVDONTNEED to *EVENT_REMOVE")
This list includes the correct soft dependency - commit 897ab3e0c49e ("userfaultfd: non-cooperative: add event for memory unmaps") which added an extra argument to mmap_region().
How should we proceed with this patch?
The backport to v4.9 should simply drop the last argument (NULL) in the call to mmap_region().
Is there some way I can indicate this sort of thing in future patches so that the automation can spot that I already know it won't apply cleanly to a particular range of kernel versions? Or even better, that I could indicate what needs to be changed when backporting to those kernel versions?
The "official" way of doing that is described here:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Docu...
However, most people just either add a comment in the commit message, or reply to the patch mail (or the "FAILED:" mail from Greg) saying how to fix this. Either way works really.
Greg will also usually look up these automated mails when he adds stuff to -stable, so if you describe how to deal with it here (like you did above) is enough as well.
-- Thanks, Sasha
linux-stable-mirror@lists.linaro.org
-
Paul Burton -
Sasha Levin