[Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Philip Müller]

Hi Greg,

the issue might be due to this patch:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/tree/releases/5.4.269/af_unix-fix-lockdep-positive-in-sk_diag_dump_icons.patch

2024-02-23T15:39:05.6297767Z   CC      kernel/sys_ni.o
2024-02-23T15:39:05.7583048Z security/apparmor/af_unix.c: In function 
‘unix_state_double_lock’:
2024-02-23T15:39:05.7586076Z security/apparmor/af_unix.c:583:17: error: 
too few arguments to function ‘unix_state_lock_nested’
2024-02-23T15:39:05.7588374Z   583 | 
unix_state_lock_nested(sk2);
2024-02-23T15:39:05.7589913Z       |                 ^~~~~~~~~~~~~~~~~~~~~~
2024-02-23T15:39:05.7591564Z In file included from 
security/apparmor/include/af_unix.h:15,
2024-02-23T15:39:05.7593341Z                  from 
security/apparmor/af_unix.c:17:
2024-02-23T15:39:05.7594989Z ./include/net/af_unix.h:77:20: note: 
declared here
2024-02-23T15:39:05.7596733Z    77 | static inline void 
unix_state_lock_nested(struct sock *sk,
2024-02-23T15:39:05.7598516Z       | 
^~~~~~~~~~~~~~~~~~~~~~
2024-02-23T15:39:05.7600862Z security/apparmor/af_unix.c:586:17: error: 
too few arguments to function ‘unix_state_lock_nested’
2024-02-23T15:39:05.7603177Z   586 | 
unix_state_lock_nested(sk1);
2024-02-23T15:39:05.7605189Z       |                 ^~~~~~~~~~~~~~~~~~~~~~
2024-02-23T15:39:05.7606765Z ./include/net/af_unix.h:77:20: note: 
declared here
2024-02-23T15:39:05.7608497Z    77 | static inline void 
unix_state_lock_nested(struct sock *sk,
2024-02-23T15:39:05.7610208Z       | 
^~~~~~~~~~~~~~~~~~~~~~
2024-02-23T15:39:05.8002385Z make[2]: *** [scripts/Makefile.build:262: 
security/apparmor/af_unix.o] Error 1
2024-02-23T15:39:05.8005077Z make[2]: *** Waiting for unfinished jobs....
2024-02-23T15:39:05.8094726Z   CC      crypto/scatterwalk.o
2024-02-23T15:39:05.9082621Z   CC [M]  fs/btrfs/sysfs.o
2024-02-23T15:39:06.2502316Z   CC      kernel/nsproxy.o
2024-02-23T15:39:06.4094246Z make[1]: *** [scripts/Makefile.build:497: 
security/apparmor] Error 2
2024-02-23T15:39:06.4207119Z make: *** [Makefile:1750: security] Error 2
2024-02-23T15:39:06.4208636Z   CC      kernel/notifier.o
2024-02-23T15:39:06.4210296Z make: *** Waiting for unfinished jobs....
2024-02-23T15:39:06.8604827Z   CC      crypto/proc.o

-- 
Best, Philip

Re: [Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Greg Kroah-Hartman]

On Fri, Feb 23, 2024 at 11:18:40PM +0700, Philip Müller wrote:
> Hi Greg,
> 
> the issue might be due to this patch:
> https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/tree/releases/5.4.269/af_unix-fix-lockdep-positive-in-sk_diag_dump_icons.patch
> 
> 2024-02-23T15:39:05.6297767Z   CC      kernel/sys_ni.o
> 2024-02-23T15:39:05.7583048Z security/apparmor/af_unix.c: In function
> ‘unix_state_double_lock’:
> 2024-02-23T15:39:05.7586076Z security/apparmor/af_unix.c:583:17: error: too
> few arguments to function ‘unix_state_lock_nested’
> 2024-02-23T15:39:05.7588374Z   583 | unix_state_lock_nested(sk2);
> 2024-02-23T15:39:05.7589913Z       |                 ^~~~~~~~~~~~~~~~~~~~~~
> 2024-02-23T15:39:05.7591564Z In file included from
> security/apparmor/include/af_unix.h:15,
> 2024-02-23T15:39:05.7593341Z                  from
> security/apparmor/af_unix.c:17:
> 2024-02-23T15:39:05.7594989Z ./include/net/af_unix.h:77:20: note: declared
> here
> 2024-02-23T15:39:05.7596733Z    77 | static inline void
> unix_state_lock_nested(struct sock *sk,
> 2024-02-23T15:39:05.7598516Z       | ^~~~~~~~~~~~~~~~~~~~~~
> 2024-02-23T15:39:05.7600862Z security/apparmor/af_unix.c:586:17: error: too
> few arguments to function ‘unix_state_lock_nested’
> 2024-02-23T15:39:05.7603177Z   586 | unix_state_lock_nested(sk1);
> 2024-02-23T15:39:05.7605189Z       |                 ^~~~~~~~~~~~~~~~~~~~~~
> 2024-02-23T15:39:05.7606765Z ./include/net/af_unix.h:77:20: note: declared
> here
> 2024-02-23T15:39:05.7608497Z    77 | static inline void
> unix_state_lock_nested(struct sock *sk,
> 2024-02-23T15:39:05.7610208Z       | ^~~~~~~~~~~~~~~~~~~~~~
> 2024-02-23T15:39:05.8002385Z make[2]: *** [scripts/Makefile.build:262:
> security/apparmor/af_unix.o] Error 1
> 2024-02-23T15:39:05.8005077Z make[2]: *** Waiting for unfinished jobs....
> 2024-02-23T15:39:05.8094726Z   CC      crypto/scatterwalk.o
> 2024-02-23T15:39:05.9082621Z   CC [M]  fs/btrfs/sysfs.o
> 2024-02-23T15:39:06.2502316Z   CC      kernel/nsproxy.o
> 2024-02-23T15:39:06.4094246Z make[1]: *** [scripts/Makefile.build:497:
> security/apparmor] Error 2
> 2024-02-23T15:39:06.4207119Z make: *** [Makefile:1750: security] Error 2
> 2024-02-23T15:39:06.4208636Z   CC      kernel/notifier.o
> 2024-02-23T15:39:06.4210296Z make: *** Waiting for unfinished jobs....
> 2024-02-23T15:39:06.8604827Z   CC      crypto/proc.o

Odd line wrapping :(

Anyway, what config options are you using?  I can't see this here and
none of the CI systems caught it either.

thanks,

greg k-h

Re: [Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Philip Müller]

On 24/02/2024 16:11, Greg Kroah-Hartman wrote:
> 
> Odd line wrapping :(
> 
> Anyway, what config options are you using?  I can't see this here and
> none of the CI systems caught it either.
> 
> thanks,
> 
> greg k-h

The config should be this one:

https://gitlab.manjaro.org/packages/core/linux54/-/blob/cb28712a0dd32966c25b3d1463db29fec4c4157f/config

Latest functional build was 5.4.268:

https://github.com/manjaro-kernels/linux54/releases

Full build log of 5.4.269 can be found here:

https://pipelinesghubeus25.actions.githubusercontent.com/IX0HkTGyy7sqAEfMIwU74grK3vJhhYEQ6R3IOeb6PZvzd3ERQo/_apis/pipelines/1/runs/193/signedlogcontent/2?urlExpires=2024-02-26T04%3A26%3A26.8692425Z&urlSigningMethod=HMACV1&urlSignature=Wmey7xXmswXsqnn%2Fd1SD2cQ%2FM2A62bQ8rW3M0sOJntc%3D

I will check if reverting the patch I think might create this issue 
might fix it. Let's see ...

-- 
Best, Philip

Re: [Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Philip Müller]

On 26/02/2024 11:28, Philip Müller wrote:
> Full build log of 5.4.269 can be found here:
> 
> https://pipelinesghubeus25.actions.githubusercontent.com/IX0HkTGyy7sqAEfMIwU74grK3vJhhYEQ6R3IOeb6PZvzd3ERQo/_apis/pipelines/1/runs/193/signedlogcontent/2?urlExpires=2024-02-26T04%3A26%3A26.8692425Z&urlSigningMethod=HMACV1&urlSignature=Wmey7xXmswXsqnn%2Fd1SD2cQ%2FM2A62bQ8rW3M0sOJntc%3D

Seems github has some session tokens which expire. Log download should 
be possible thru here:

https://github.com/manjaro-kernels/linux54/suites/21057183037/logs?attempt=1

-- 
Best, Philip

Re: [Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Greg Kroah-Hartman]

On Mon, Feb 26, 2024 at 11:28:33AM +0700, Philip Müller wrote:
> On 24/02/2024 16:11, Greg Kroah-Hartman wrote:
> > 
> > Odd line wrapping :(
> > 
> > Anyway, what config options are you using?  I can't see this here and
> > none of the CI systems caught it either.
> > 
> > thanks,
> > 
> > greg k-h
> 
> The config should be this one:
> 
> https://gitlab.manjaro.org/packages/core/linux54/-/blob/cb28712a0dd32966c25b3d1463db29fec4c4157f/config

Ah, here's the config, let me go test...

Re: [Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Greg Kroah-Hartman]

On Tue, Feb 27, 2024 at 10:02:37AM +0100, Greg Kroah-Hartman wrote:
> On Mon, Feb 26, 2024 at 11:28:33AM +0700, Philip Müller wrote:
> > On 24/02/2024 16:11, Greg Kroah-Hartman wrote:
> > > 
> > > Odd line wrapping :(
> > > 
> > > Anyway, what config options are you using?  I can't see this here and
> > > none of the CI systems caught it either.
> > > 
> > > thanks,
> > > 
> > > greg k-h
> > 
> > The config should be this one:
> > 
> > https://gitlab.manjaro.org/packages/core/linux54/-/blob/cb28712a0dd32966c25b3d1463db29fec4c4157f/config
> 
> Ah, here's the config, let me go test...
> 
> 

This config builds just fine for me here.  Are you sure this isn't some
out-of-tree change you have in your tree?  I'm going to need more
information here as no one seems to be able to reproduce the issue...

thanks,

greg k-h

Re: [Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Philip Müller]

On 24/02/2024 16:11, Greg Kroah-Hartman wrote:
> the issue might be due to this patch:
> https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/tree/releases/5.4.269/af_unix-fix-lockdep-positive-in-sk_diag_dump_icons.patch

Yes, reverting this commit makes it build again on my end ...

-- 
Best, Philip

Re: [Regression] 5.4.269 fails to build due to security/apparmor/af_unix.c:583:17: error: too few arguments to function ‘unix_state_lock_nested’

[Greg Kroah-Hartman]

On Mon, Feb 26, 2024 at 10:05:16PM +0700, Philip Müller wrote:
> On 24/02/2024 16:11, Greg Kroah-Hartman wrote:
> > the issue might be due to this patch:
> > https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/tree/releases/5.4.269/af_unix-fix-lockdep-positive-in-sk_diag_dump_icons.patch
> 
> Yes, reverting this commit makes it build again on my end ...

Can you attach a .config file that shows this?  I can't seem to
duplicate it here.

thanks,

greg k-h