Re: [musl] remove SIGSTKFLT on mips?

[Rich Felker <dalias@libc.org>]

On Thu, May 21, 2020 at 03:21:03PM -0400, Rich Felker wrote:
> On Thu, May 21, 2020 at 09:51:29AM -0700, enh wrote:
> > (context: https://github.com/landley/toybox/issues/194)
> > 
> > seems like musl defines SIGSTKFLT for mips as signal 7. the kernel
> > disagrees and says SIGEMT is 7, and that mips doesn't have a
> > SIGSTKFLT:
> > 
> > arch/mips/include/uapi/asm/signal.h:#define SIGEMT 7
> > 
> > git blame says SIGSTKFLT was introduced in the original commit:
> > 
> > commit 6315004f6102dca44c4ba50654a36967b8b9c2a6
> > Author: Rich Felker <dalias@aerifal.cx>
> > Date:   Wed Jul 11 04:22:13 2012 -0400
> > 
> >     initial version of mips (o32) port, based on work by Richard
> > Pennington (rdp)
> > 
> >     basically, this version of the code was obtained by starting with
> >     rdp's work from his ellcc source tree, adapting it to musl's build
> >     system and coding style, auditing the bits headers for discrepencies
> >     with kernel definitions or glibc/LSB ABI or large file issues, fixing
> >     up incompatibility with the old binutils from aboriginal linux, and
> >     adding some new special cases to deal with the oddities of sigaction
> >     and pipe syscall interfaces on mips.
> > 
> >     at present, minimal test programs work, but some interfaces are broken
> >     or missing. threaded programs probably will not link.
> 
> I have the two attached patches pending push if there's no objection.

> >From a19640a8912e9876f293267986875742867faa54 Mon Sep 17 00:00:00 2001
> From: Rich Felker <dalias@aerifal.cx>
> Date: Thu, 21 May 2020 13:14:40 -0400
> Subject: [PATCH 1/2] handle possibility that SIGEMT replaces SIGSTKFLT in
>  strsignal
> 
> presently all archs define SIGSTKFLT but this is not correct. change
> strsignal as a prerequisite for fixing that.
> ---
>  src/string/strsignal.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/src/string/strsignal.c b/src/string/strsignal.c
> index 96bfe841..922d3271 100644
> --- a/src/string/strsignal.c
> +++ b/src/string/strsignal.c
> @@ -31,7 +31,11 @@ static const char map[] = {
>  	[SIGPIPE]   = 13,
>  	[SIGALRM]   = 14,
>  	[SIGTERM]   = 15,
> +#if defined(SIGSTKFLT)
>  	[SIGSTKFLT] = 16,
> +#elif defined(SIGEMT)
> +	[SIGEMT]    = 16,
> +#endif

Note here that 16 does not imply the value is 16, just that the
remapping for archs that don't use the common signal order needs to
get the 16'th string here:

>  	[SIGCHLD]   = 17,
>  	[SIGCONT]   = 18,
>  	[SIGSTOP]   = 19,
> @@ -70,7 +74,11 @@ static const char strings[] =
>  	"Broken pipe\0"
>  	"Alarm clock\0"
>  	"Terminated\0"
> +#if defined(SIGSTKFLT)
>  	"Stack fault\0"
> +#elif defined(SIGEMT)
> +	"Emulator trap\0"
> +#endif

Which is conceptually slightly wrong since there needs to be a dummy
entry here if neither is defined, so as not to mess up the numbering.
I'll fix that before pushing even though in practice all archs have
one or the other (or the code wouldn't have compiled before).

Rich