Re: [PULL 2/2] linux-user: fix guards for the fsmount(2) syscall series

On 7/27/26 16:15, Stefan Hajnoczi wrote:
> On Sun, Jul 26, 2026 at 3:36 PM Helge Deller <[email protected]> wrote:
>>
>> From: Matt Turner <[email protected]>
>>
>> The fsopen(), fsconfig(), fsmount() and fspick() implementations are
>> guarded by defined(NR_fsopen) rather than defined(__NR_fsopen). No such
>> macro exists, so the guard is never true and the entire series compiles
>> out. Guests calling any of the four get -ENOSYS, which for example makes
>> systemd's credential setup fail with EXIT_CREDENTIALS for most units.
>>
>> The strace bits for fsconfig() have the same typo.
> 
> Please take a look at the following CI failure:
> 
> ../linux-user/strace.c: In function ‘print_fsconfig_cmd_name’:
> ../linux-user/strace.c:4352:10: error: ‘FSCONFIG_SET_FLAG’ undeclared
> (first use in this function)
> 4352 | case FSCONFIG_SET_FLAG:
> | ^~~~~~~~~~~~~~~~~
> 
> https://gitlab.com/qemu-project/qemu/-/jobs/15553379213#L2358
> 
> Once this has been fixed, please send a v2 pull request. Thanks!

I've sent a new pull request.
I added a dependency on the availability of the FSCONFIG_SET_FLAG define.
This is okay for now, but maybe a better fix is possible?

Usually this define should be available, but it doesn't seem to
be in that chroot image.
I'm sure I'm missing something, but is it somewhere documented how to get that chroot
image locally so that I can reproduce the problem?

Helge