linux-next: test failure after merge of the landlock tree

linux-next: test failure after merge of the landlock tree

[Mark Brown]

Hi all,

After merging the landlock tree, today's linux-next arm64 kunit failed
like this:

[00:12:55] ============================================================
[00:12:55] ================ landlock_cap (5 subtests) =================
[00:12:55] # test_cap_to_bit_invalid: EXPECTATION FAILED at security/landlock/cap.c:113
[00:12:55] Expected kunit_suppressed_warning_count(__kunit_suppress) == 1, but
[00:12:55]     kunit_suppressed_warning_count(__kunit_suppress) == 2 (0x2)
[00:12:55] [FAILED] test_cap_to_bit_invalid
[00:12:55]     # module: landlock
[00:12:55] # landlock_cap: pass:4 fail:1 skip:0 total:5
[00:12:55] # Totals: pass:4 fail:1 skip:0 total:5
[00:12:55] ================== [FAILED] landlock_cap ===================
[00:12:55] ============================================================

Caused by commit

   b6a56b30b9d5a (landlock: Enforce capability restrictions)

Due to the prior build failures I have merged the version from
next-20261001 instead.

Re: linux-next: test failure after merge of the landlock tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 202 bytes --]

On Sat, Oct 10, 2026 at 01:31:03AM +0100, Mark Brown wrote:
> Hi all,
> 
> After merging the landlock tree, today's linux-next arm64 kunit failed
> like this:

This issue is still present today.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: linux-next: test failure after merge of the landlock tree

[Mickaël Salaün]

On Mon, Oct 19, 2026 at 07:17:09PM +0100, Mark Brown wrote:
> On Sat, Oct 10, 2026 at 01:31:03AM +0100, Mark Brown wrote:
> > Hi all,
> > 
> > After merging the landlock tree, today's linux-next arm64 kunit failed
> > like this:
> 
> This issue is still present today.

I removed these wrong kunit checks Friday so it should be good now, but
I notice that linux-next still has the next-20261001/landlock merge
commit instead of the latest changes from mic/next (which should not
conflict).  Could you please confirm that a normal merge works?

Regards,
 Mickaël

Re: linux-next: test failure after merge of the landlock tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 663 bytes --]

On Mon, Oct 19, 2026 at 11:11:20PM +0200, Mickaël Salaün wrote:
> On Mon, Oct 19, 2026 at 07:17:09PM +0100, Mark Brown wrote:

> > This issue is still present today.

> I removed these wrong kunit checks Friday so it should be good now, but
> I notice that linux-next still has the next-20261001/landlock merge
> commit instead of the latest changes from mic/next (which should not
> conflict).  Could you please confirm that a normal merge works?

It was broken with whatever I fetched yesterday hence me reporting it
still broken and merging the old version, I've just started the merge
for today so I'll find out this afternoon if the fix worked.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: linux-next: test failure after merge of the landlock tree

[Mickaël Salaün]

On Tue, Oct 20, 2026 at 10:54:29AM +0100, Mark Brown wrote:
> On Mon, Oct 19, 2026 at 11:11:20PM +0200, Mickaël Salaün wrote:
> > On Mon, Oct 19, 2026 at 07:17:09PM +0100, Mark Brown wrote:
> 
> > > This issue is still present today.
> 
> > I removed these wrong kunit checks Friday so it should be good now, but

In fact, I updated mic/next Wednesday with a new patch series addressing
these issues.  I only added two (unrelated) commits on top of that
Friday.

> > I notice that linux-next still has the next-20261001/landlock merge
> > commit instead of the latest changes from mic/next (which should not
> > conflict).  Could you please confirm that a normal merge works?
> 
> It was broken with whatever I fetched yesterday hence me reporting it
> still broken and merging the old version, I've just started the merge
> for today so I'll find out this afternoon if the fix worked.

The new commits from Wednesday don't add the bogus kunit lines (i.e.
there is no kunit_suppressed_warning_count() calls anymore), so it's not
possible to get the same error.

Merging mic/next into the current linux-next works for me, and
./tools/testing/kunit/kunit.py run --alltests works too.  Please let me
know if anything is wrong.

Regards,
 Mickaël

Re: linux-next: test failure after merge of the landlock tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 709 bytes --]

On Tue, Oct 20, 2026 at 02:28:48PM +0200, Mickaël Salaün wrote:
> On Tue, Oct 20, 2026 at 10:54:29AM +0100, Mark Brown wrote:

> > > I removed these wrong kunit checks Friday so it should be good now, but

> In fact, I updated mic/next Wednesday with a new patch series addressing
> these issues.  I only added two (unrelated) commits on top of that
> Friday.

What I have today is f349c073889ff7694b470ca34dae9bd94d47b0cc which is
"landlock: Fix superblock use-after-free in release_inode()" rebased on
the 16th (I just tried refetching and there was no change).  It seems
very surprising that there might have been an infrastructure issue
fetching the tree all of Thursday, Friday and Monday?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: linux-next: test failure after merge of the landlock tree

[Mickaël Salaün]

On Tue, Oct 20, 2026 at 02:02:42PM +0100, Mark Brown wrote:
> On Tue, Oct 20, 2026 at 02:28:48PM +0200, Mickaël Salaün wrote:
> > On Tue, Oct 20, 2026 at 10:54:29AM +0100, Mark Brown wrote:
> 
> > > > I removed these wrong kunit checks Friday so it should be good now, but
> 
> > In fact, I updated mic/next Wednesday with a new patch series addressing
> > these issues.  I only added two (unrelated) commits on top of that
> > Friday.
> 
> What I have today is f349c073889ff7694b470ca34dae9bd94d47b0cc which is
> "landlock: Fix superblock use-after-free in release_inode()" rebased on
> the 16th (I just tried refetching and there was no change).  It seems
> very surprising that there might have been an infrastructure issue
> fetching the tree all of Thursday, Friday and Monday?

Commit 0da3cb3db77e3fe2202ddb637f6d27668dc360f7 appears in the
transparency log:
https://git.kernel.org/pub/scm/infra/transparency-logs/gitolite/git/1.git/commit/?id=0da8a8d3dc37ec9a71b76f9739957297d64d4b73

Re: linux-next: test failure after merge of the landlock tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1125 bytes --]

On Tue, Oct 20, 2026 at 03:16:51PM +0200, Mickaël Salaün wrote:
> On Tue, Oct 20, 2026 at 02:02:42PM +0100, Mark Brown wrote:

> > What I have today is f349c073889ff7694b470ca34dae9bd94d47b0cc which is
> > "landlock: Fix superblock use-after-free in release_inode()" rebased on
> > the 16th (I just tried refetching and there was no change).  It seems
> > very surprising that there might have been an infrastructure issue
> > fetching the tree all of Thursday, Friday and Monday?

> Commit 0da3cb3db77e3fe2202ddb637f6d27668dc360f7 appears in the
> transparency log:
> https://git.kernel.org/pub/scm/infra/transparency-logs/gitolite/git/1.git/commit/?id=0da8a8d3dc37ec9a71b76f9739957297d64d4b73

This probably needs reporting to helpdesk (CCed), when I fetch:

   https://git.kernel.org/pub/scm/linux/kernel/git/mic/linux.git next

or go to:

   https://git.kernel.org/pub/scm/linux/kernel/git/mic/linux.git/commit/?id=f349c073889ff7694b470ca34dae9bd94d47b0cc

what I see is the commit referenced above being the head of your next
branch, not this other one you are mentioning.  I'm on the fra mirror.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: linux-next: test failure after merge of the landlock tree

[Mickaël Salaün]

On Tue, Oct 20, 2026 at 02:33:26PM +0100, Mark Brown wrote:
> On Tue, Oct 20, 2026 at 03:16:51PM +0200, Mickaël Salaün wrote:
> > On Tue, Oct 20, 2026 at 02:02:42PM +0100, Mark Brown wrote:
> 
> > > What I have today is f349c073889ff7694b470ca34dae9bd94d47b0cc which is
> > > "landlock: Fix superblock use-after-free in release_inode()" rebased on
> > > the 16th (I just tried refetching and there was no change).  It seems
> > > very surprising that there might have been an infrastructure issue
> > > fetching the tree all of Thursday, Friday and Monday?
> 
> > Commit 0da3cb3db77e3fe2202ddb637f6d27668dc360f7 appears in the
> > transparency log:
> > https://git.kernel.org/pub/scm/infra/transparency-logs/gitolite/git/1.git/commit/?id=0da8a8d3dc37ec9a71b76f9739957297d64d4b73
> 
> This probably needs reporting to helpdesk (CCed), when I fetch:
> 
>    https://git.kernel.org/pub/scm/linux/kernel/git/mic/linux.git next
> 
> or go to:
> 
>    https://git.kernel.org/pub/scm/linux/kernel/git/mic/linux.git/commit/?id=f349c073889ff7694b470ca34dae9bd94d47b0cc
> 
> what I see is the commit referenced above being the head of your next
> branch, not this other one you are mentioning.  I'm on the fra mirror.

This head commit correct.  I mean that commit 0da3cb3db77e ("landlock:
Enforce capability restrictions") was the one that fixed the kunit issue
and it was part of the commits I pushed Wednesday.  Anyway, it looks
good now, thanks.

Re: linux-next: test failure after merge of the landlock tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 752 bytes --]

On Tue, Oct 20, 2026 at 03:16:51PM +0200, Mickaël Salaün wrote:
> On Tue, Oct 20, 2026 at 02:02:42PM +0100, Mark Brown wrote:

> > What I have today is f349c073889ff7694b470ca34dae9bd94d47b0cc which is
> > "landlock: Fix superblock use-after-free in release_inode()" rebased on
> > the 16th (I just tried refetching and there was no change).  It seems
> > very surprising that there might have been an infrastructure issue
> > fetching the tree all of Thursday, Friday and Monday?

> Commit 0da3cb3db77e3fe2202ddb637f6d27668dc360f7 appears in the
> transparency log:
> https://git.kernel.org/pub/scm/infra/transparency-logs/gitolite/git/1.git/commit/?id=0da8a8d3dc37ec9a71b76f9739957297d64d4b73

What I have has made it through KUnit OK.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]