This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] powerpc: Enable demuxed sysv IPC syscalls
- From: Arnd Bergmann <arnd at arndb dot de>
- To: libc-alpha at sourceware dot org
- Cc: Michael Ellerman <mpe at ellerman dot id dot au>, Andreas Schwab <schwab at linux-m68k dot org>, "Paul E. Murphy" <murphyp at linux dot vnet dot ibm dot com>, Tulio Magno Quites Machado Filho <tuliom at linux dot vnet dot ibm dot com>, Sam Bobroff <sam dot bobroff at au1 dot ibm dot com>
- Date: Wed, 16 Dec 2015 10:57:21 +0100
- Subject: Re: [PATCH] powerpc: Enable demuxed sysv IPC syscalls
- Authentication-results: sourceware.org; auth=none
- References: <5660A8D0 dot 5090003 at linux dot vnet dot ibm dot com> <2535039 dot fp0alOV0kv at wuerfel> <1450257707 dot 4370 dot 2 dot camel at ellerman dot id dot au>
On Wednesday 16 December 2015 20:21:47 Michael Ellerman wrote:
> On Wed, 2015-12-16 at 10:00 +0100, Arnd Bergmann wrote:
> > On Wednesday 16 December 2015 09:42:53 Andreas Schwab wrote:
> > > Michael Ellerman <mpe@ellerman.id.au> writes:
>
> > > > Having said that it doesn't seem like it buys us that much. Just that userspace
> > > > no longer has to worry about passing IPC_64, or is there more to it that I'm
> > > > missing? I guess that's a worthwhile cleanup though.
> > >
> > > By removing the requirement to add IPC_64 you remove the need for a
> > > special wrapper, and the syscalls can be implemented by just adding it
> > > to syscalls.list. That should be the goal for every new syscall. Also,
> > > it doesn't make sense to offer obsolete interfaces (the non-IPC_64
> > > variants) through such syscalls.
> >
> > Right, the other point is that we can do it consistently for all
> > architectures that currently don't have separate ipc syscalls but
> > may want to add them.
>
> OK, sounds good.
>
> > I made a table when this first came up:
> > https://docs.google.com/spreadsheets/d/18GxXEHE2ywnSr-SPoGFd1ABz6wEM1ex-JMu5lEraaH8
> >
> > So the affected architectures are cris, frv, m32r, m68k, mips-o32, mn10300,
> > s390, sh, sparc and x86.
>
> Eek, powerpc is red, bad!
>
> So the only complication with reverting the patch is that we've since added
> sys_mlock2 after the IPC calls, as NR 378.
>
> I guess because we haven't released a kernel with mlock2 using that number I'll
> just renumber it, as if the IPC calls were never there.
I think it would be better to leave the numbers as they are and bring them back
with the name numbers when we get there, just to avoid any incompatibilities with
people using mlock2 on prerelease kernels.
Arnd