Re: [PATCH] s390/uapi: Remove obsolete unistd_32.h from Kbuild file

[Heiko Carstens]

On Wed, Jul 08, 2026 at 02:47:01PM +0200, Alexander Gordeev wrote:
> On Wed, Jul 01, 2026 at 03:47:27PM +0200, Thomas Huth wrote:
> 
> Hi Thomas,
> 
> > From: Thomas Huth <thuth@xxxxxxxxxx>
> > 
> > unistd_32.h is not build anymore since commit 4ac286c4a8d90
> > ("s390/syscalls: Switch to generic system call table generation").
> > Thus drop the superfluous line from the Kbuild file now, too.
> > 
> > Signed-off-by: Thomas Huth <thuth@xxxxxxxxxx>
> > ---
> >  arch/s390/include/uapi/asm/Kbuild | 1 -
> >  1 file changed, 1 deletion(-)
> > 
> > diff --git a/arch/s390/include/uapi/asm/Kbuild b/arch/s390/include/uapi/asm/Kbuild
> > index 46c1ff0b842a1..378151cef3cce 100644
> > --- a/arch/s390/include/uapi/asm/Kbuild
> > +++ b/arch/s390/include/uapi/asm/Kbuild
> > @@ -1,4 +1,3 @@
> >  # SPDX-License-Identifier: GPL-2.0
> >  
> > -generated-y += unistd_32.h
> >  generated-y += unistd_64.h
> 
> On repeated builds it results in:
> 
>   SYSHDR  arch/s390/include/generated/uapi/asm/unistd_32.h
>   REMOVE  arch/s390/include/generated/uapi/asm/unistd_32.h

How did you achieve that unistd_32.h is actually generated? The required
syscall table information to generate such a header file is not present.
This looks very odd to me.

The above should be impossible with current upstream code, and I cannot
reproduce it. Any chance you applied Thomas' patch on top of an old commit
which still includes compat support?

> A complete removal would require changes to arch/s390/kernel/syscalls/Makefile
> at least,

I don't think so.

> but we still  might want to have unistd_32.h around.

No, that doesn't make sense. compat support is completely gone, including all
header files. The patch looks fine as it is for me, but we need to sort out
what you reported.