Re: [Qemu-devel] [PULL v3 47/55] linux headers: update against Linux 5.2-rc1

[Cornelia Huck <cohuck@redhat.com>]

On Wed, 22 May 2019 15:22:23 +0200
Aleksandar Markovic <aleksandar.m.mail@gmail.com> wrote:

> On May 22, 2019 2:24 PM, "Cornelia Huck" <cohuck@redhat.com> wrote:
> >
> > On Wed, 22 May 2019 14:10:39 +0200
> > Laurent Vivier <laurent@vivier.eu> wrote:
> >  
> > > On 22/05/2019 14:07, Cornelia Huck wrote:  
> > > > On Wed, 22 May 2019 13:47:25 +0200
> > > > Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
> > > >  
> > > >> On 5/21/19 5:28 PM, Cornelia Huck wrote:  
> > > >>> commit a188339ca5a396acc588e5851ed7e19f66b0ebd9
> > > >>>
> > > >>> Signed-off-by: Cornelia Huck <cohuck@redhat.com>
> > > >>> ---  
> > > >> [...]  
> > > >>>   #define __NR_mq_notify 184
> > > >>>   __SC_COMP(__NR_mq_notify, sys_mq_notify, compat_sys_mq_notify)
> > > >>>   #define __NR_mq_getsetattr 185
> > > >>> @@ -536,8 +567,10 @@ __SC_COMP(__NR_msgsnd, sys_msgsnd,  
> compat_sys_msgsnd)
> > > >>>   __SYSCALL(__NR_semget, sys_semget)
> > > >>>   #define __NR_semctl 191
> > > >>>   __SC_COMP(__NR_semctl, sys_semctl, compat_sys_semctl)
> > > >>> +#if defined(__ARCH_WANT_TIME32_SYSCALLS) || __BITS_PER_LONG != 32  
> > > >
> > > > Eww. It seems only aarch64 sets __ARCH_WANT_TIME32_SYSCALLS, and the
> > > > second condition probably catches others but not mipsel.
> > > >  
> > > >>>   #define __NR_semtimedop 192
> > > >>> -__SC_COMP(__NR_semtimedop, sys_semtimedop, compat_sys_semtimedop)
> > > >>> +__SC_COMP(__NR_semtimedop, sys_semtimedop, sys_semtimedop_time32)
> > > >>> +#endif
> > > >>>   #define __NR_semop 193
> > > >>>   __SYSCALL(__NR_semop, sys_semop)  
> > > >> [...]
> > > >>
> > > >> https://app.shippable.com/github/qemu/qemu/runs/1703/summary/console
> > > >>
> > > >> It seems this commit introduce a regression on mips32:
> > > >>
> > > >>    CC      mipsel-linux-user/linux-user/syscall.o
> > > >> ./linux-user/syscall.c: In function 'safe_semtimedop':
> > > >> ./linux-user/syscall.c:697:25: error: '__NR_semtimedop' undeclared
> > > >> (first use in this function)
> > > >>       return safe_syscall(__NR_##name, arg1, arg2, arg3, arg4); \  
> > > >
> > > > So, we unconditionally deal with this syscall, i.e. we assume it is
> > > > always present? (I'm not sure of the logic in linux-user code.)
> > > >  
> > >
> > > linux-user assumes it is present if __NR_msgsnd is present.  
> >
> > Hm. The kernel change seems to break that assumption. Does anyone with
> > mips knowledge have an idea whether that was intentional (and the
> > linux-user code needs to be changed), or whether that's an issue on the
> > kernel side?
> >  
> 
> Hi, Cornelia.
> 
> Thanks for your involving into this issue!
> 
> It could be that (not-originating-from-MIPS) kernel commit:
> 
> https://github.com/torvalds/linux/commit/1a787fc5ba18ac767e635c58d06a0b46876184e3
> 
> made a mess with system call availability for MIPS (I will forward this to
> MIPS kernel maintainer Paul Burton). My impression is that this was not
> intentional, and is a temporary instability of kernel interface.

I don't think that's the problematic commit; that one seems to be a
follow-up on c8ce48f06503 ("asm-generic: Make time32 syscall numbers
optional") for tools usage (we sync from the 'normal' headers).

The stated intention of the asm-generic commit is to keep 32 bit
architectures working as before via defining
__ARCH_WANT_TIME32_SYSCALLS, but it seems that was not done for mips
(but it should, right?)

> However, I think that QEMU nevertheless should not make the assumption that
> if __NR_MSGSND, than semtimedop() is present. It could be true, but it is
> still just self-imposed belief in QEMU, kernel never guarantied such things.

I'm not too familiar with that family of syscalls; is there a better
way to check for syscall availability here?

> The alternative way of invoking via IPCV6 (else part of “ifdef
> __NR_MSGSND”) should work for MIPS in the present stage of headers and
> kernel.

If my assumption above (mips skipped by accident) is correct, we need
to fix the kernel headers instead :/ -- unless we want to add a
temporary build fix.

> As a side note, perhaps we shoul update kernel headers only off of stable
> kernel releases.

In the past, we have even updated the kernel headers against
non-mainline (kvm) versions :)

Breakage here seems to be rare (and if this is a real kernel interface
bug, it'd be a good thing that we caught it); I believe getting support
for new features into QEMU quicker makes that a good trade-off.