Messages in this thread | | | Date | Tue, 2 Dec 2014 15:38:24 +0100 | Subject | Re: [PATCH 2/2] fs: Wire up sendfd() syscall (all architectures) | From | Geert Uytterhoeven <> |
| |
Hi Alex,
On Tue, Dec 2, 2014 at 3:31 PM, Alex Dubov <alex.dubov@gmail.com> wrote: > On Tue, Dec 2, 2014 at 10:42 PM, Michal Simek <monstr@monstr.eu> wrote: >> On 12/02/2014 09:01 AM, Geert Uytterhoeven wrote: >>> This really needs a CC to linux-arch (added). >>> >>> On Tue, Dec 2, 2014 at 5:35 AM, Alex Dubov <alex.dubov@gmail.com> wrote: >>>> Signed-off-by: Alex Dubov <oakad@yahoo.com> >>>> --- > >> The same for microblaze here >> arch/microblaze/include/asm/unistd.h > > This invites the question as to why the __NR_syscalls macro is not > defined in uapi/asm/unistd.h on those architectures, where it will be > easier to spot? After all, asm/unistd.h includes uapi/asm/unistd.h > unconditionally.
Because it's not part of the ABI?
There may be multiple ABIs, with multiple syscall ranges. Userspace only needs to know if a syscall is available, not what the valid syscall number range is. The kernel does need to know the size of the full syscall table.
Gr{oetje,eeting}s,
Geert
-- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
| |