Re: [PATCH 4/5] uapi: always define F_GETLK64/F_SETLK64/F_SETLKW64 in fcntl.h

[Guo Ren]

On Tue, Jan 11, 2022 at 11:33 PM Arnd Bergmann <arnd@xxxxxxxx> wrote:
>
> On Tue, Jan 11, 2022 at 9:35 AM Christoph Hellwig <hch@xxxxxx> wrote:
> >
> > The fcntl F_GETLK64/F_SETLK64/F_SETLKW64 are only implemented for the
> > 32-bit syscall APIs, but we also need them for compat handling on 64-bit
> > builds.  Redefining them is error prone (as shown by the example that
> > parisc gets it wrong currently), so we should use the same defines for
> > both case.  In theory we could try to hide them from userspace, but
> > given that only MIPS actually gets that right, while the asm-generic
> > version used by most architectures relies on a Kconfig symbol that can't
> > be relied on to be set properly by userspace is a clear indicator to not
> > bother.
> >
> > Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> > ---
>
> > diff --git a/include/uapi/asm-generic/fcntl.h b/include/uapi/asm-generic/fcntl.h
> > index 98f4ff165b776..43d7c44031be0 100644
> > --- a/include/uapi/asm-generic/fcntl.h
> > +++ b/include/uapi/asm-generic/fcntl.h
> > @@ -116,13 +116,11 @@
> >  #define F_GETSIG       11      /* for sockets. */
> >  #endif
> >
> > -#ifndef CONFIG_64BIT
> >  #ifndef F_GETLK64
> >  #define F_GETLK64      12      /*  using 'struct flock64' */
> >  #define F_SETLK64      13
> >  #define F_SETLKW64     14
> >  #endif
> > -#endif
> >
> >  #ifndef F_SETOWN_EX
> >  #define F_SETOWN_EX    15
>
> This is a very subtle change to the exported UAPI header contents:
> On 64-bit architectures, the three unusable numbers are now always
> shown, rather than depending on a user-controlled symbol.
>
> This is probably what we want here for compatibility reasons, but I think
> it should be explained in the changelog text, and I'd like Jeff or Bruce
> to comment on it as well: the alternative here would be to make the
> uapi definition depend on __BITS_PER_LONG==32, which is

__BITS_PER_LONG==32 || __KERNEL__  just for kernel use in compat.

> technically the right thing to do but more a of a change.
>
>        Arnd



-- 
Best Regards
 Guo Ren

ML: https://lore.kernel.org/linux-csky/