* [RFC] Add 4/8 bytes to 'struct k_sigaction'...
@ 2005-01-06 4:36 Steven J. Hill
2005-01-06 15:48 ` Christoph Hellwig
0 siblings, 1 reply; 4+ messages in thread
From: Steven J. Hill @ 2005-01-06 4:36 UTC (permalink / raw)
To: linux-mips
In order support easy building of IRIX emulation support as either
a module or statically, I would like to simply have the 'sa_restorer'
function pointer be defined for all big endian architectures. This
means extra storage space, but I wanted comments first. The problem
is that if you configure for IRIX, build and then decide that you
want to build it as a module, CONFIG_BINFMT_IRIX_MODULE gets defined
instead of CONFIG_BINFMT_IRIX and hence 'signal.h' changes and the
whole kernel has to be rebuilt. If we simply have it be defined for
big endian architectures regardless, then there is no recompile. Al
Viro did comment that Christoph Hellwig may be moving stuff like
'struct sighand_struct' out of 'sched.h' along with some other things
which might make this below a non-issue, but that is far off. Comments
before I commit?
-Steve
Index: signal.h
===================================================================
RCS file: /home/cvs/linux/include/asm-mips/signal.h,v
retrieving revision 1.17
diff -u -r1.17 signal.h
--- signal.h 30 Sep 2003 14:27:29 -0000 1.17
+++ signal.h 6 Jan 2005 04:21:58 -0000
@@ -135,7 +135,7 @@
struct k_sigaction {
struct sigaction sa;
-#ifdef CONFIG_BINFMT_IRIX
+#if !defined(CONFIG_CPU_LITTLE_ENDIAN)
void (*sa_restorer)(void);
#endif
};
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [RFC] Add 4/8 bytes to 'struct k_sigaction'...
2005-01-06 4:36 [RFC] Add 4/8 bytes to 'struct k_sigaction' Steven J. Hill
@ 2005-01-06 15:48 ` Christoph Hellwig
2005-01-06 15:55 ` sjhill
2005-01-06 16:12 ` Thiemo Seufer
0 siblings, 2 replies; 4+ messages in thread
From: Christoph Hellwig @ 2005-01-06 15:48 UTC (permalink / raw)
To: Steven J. Hill; +Cc: linux-mips
> --- signal.h 30 Sep 2003 14:27:29 -0000 1.17
> +++ signal.h 6 Jan 2005 04:21:58 -0000
> @@ -135,7 +135,7 @@
>
> struct k_sigaction {
> struct sigaction sa;
> -#ifdef CONFIG_BINFMT_IRIX
> +#if !defined(CONFIG_CPU_LITTLE_ENDIAN)
> void (*sa_restorer)(void);
> #endif
#ifdef __mipseb__ maybe?
Is IRIX emulation even working?
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [RFC] Add 4/8 bytes to 'struct k_sigaction'...
2005-01-06 15:48 ` Christoph Hellwig
@ 2005-01-06 15:55 ` sjhill
2005-01-06 16:12 ` Thiemo Seufer
1 sibling, 0 replies; 4+ messages in thread
From: sjhill @ 2005-01-06 15:55 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Steven J. Hill, linux-mips
>
> #ifdef __mipseb__ maybe?
>
I thought about that too. I spoke with Ralf on IRC privately and
there are race conditions associated with removing the module. That
aside, I was concerned about how people felt about the extra 4 or 8
bytes.
> Is IRIX emulation even working?
>
That is what I am working on.
-Steve
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [RFC] Add 4/8 bytes to 'struct k_sigaction'...
2005-01-06 15:48 ` Christoph Hellwig
2005-01-06 15:55 ` sjhill
@ 2005-01-06 16:12 ` Thiemo Seufer
1 sibling, 0 replies; 4+ messages in thread
From: Thiemo Seufer @ 2005-01-06 16:12 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Steven J. Hill, linux-mips
Christoph Hellwig wrote:
> > --- signal.h 30 Sep 2003 14:27:29 -0000 1.17
> > +++ signal.h 6 Jan 2005 04:21:58 -0000
> > @@ -135,7 +135,7 @@
> >
> > struct k_sigaction {
> > struct sigaction sa;
> > -#ifdef CONFIG_BINFMT_IRIX
> > +#if !defined(CONFIG_CPU_LITTLE_ENDIAN)
> > void (*sa_restorer)(void);
> > #endif
>
> #ifdef __mipseb__ maybe?
AFAICS most parts of the kernel seem to prefer CONFIG_CPU_LITTLE_ENDIAN
over compiler-dependent macros. (And IIRC it would need to be __MIPSEB.)
Thiemo
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2005-01-06 16:13 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-01-06 4:36 [RFC] Add 4/8 bytes to 'struct k_sigaction' Steven J. Hill
2005-01-06 15:48 ` Christoph Hellwig
2005-01-06 15:55 ` sjhill
2005-01-06 16:12 ` Thiemo Seufer
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.