This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
RE: [PATCHv4, MIPS] Add support for O32 FPXX and program header based ABI information
- From: Matthew Fortune <Matthew dot Fortune at imgtec dot com>
- To: Joseph Myers <joseph at codesourcery dot com>
- Cc: "'libc-alpha at sourceware dot org'" <libc-alpha at sourceware dot org>, "Moore, Catherine (Catherine_Moore at mentor dot com)" <Catherine_Moore at mentor dot com>, 'Andrew Pinski' <pinskia at gmail dot com>, "Rich Felker (dalias at libc dot org)" <dalias at libc dot org>, Rich Fuhler <Rich dot Fuhler at imgtec dot com>, "'macro at codesourcery dot com'" <macro at codesourcery dot com>
- Date: Wed, 31 Dec 2014 22:28:46 +0000
- Subject: RE: [PATCHv4, MIPS] Add support for O32 FPXX and program header based ABI information
- Authentication-results: sourceware.org; auth=none
- References: <6D39441BF12EF246A7ABCE6654B0235320F8DFA4 at LEMAIL01 dot le dot imgtec dot org> <alpine dot DEB dot 2 dot 10 dot 1412311718240 dot 26199 at digraph dot polyomino dot org dot uk>
Joseph Myers <joseph@codesourcery.com> writes:
> On Tue, 23 Dec 2014, Matthew Fortune wrote:
>
> > I'm still running the glibc testsuite on a native MIPS machine to
> > double check results. I'm running with glibc configured for
> > mips2/mips3 on a board with no-fpu and see 31 failures for O32, 32 for
> > N64 and a bizarre set of failures about "allocate_stack: Assertion
> `size != 0' failed."
> > for N32. None of these look anything to do with this patch so I'm
> > collecting reference results before my patch to make sure.
>
> That's more than the known set for previous releases listed at
> <https://sourceware.org/glibc/wiki/Release/2.20>, but still should be
> independent of this patch.
It looks like I had some sort of inconsistency in the build the re-run was
much cleaner. For reference the failures I am saw were:
gcc version 4.9.1 (Debian 4.9.1-19)
(GNU Binutils for Debian) 2.24.90.20141023
hard-float, big endian
= FAIL for all ABIs =
FAIL: conform/ISO/math.h/linknamespace
FAIL: conform/POSIX/fcntl.h/conform
FAIL: conform/POSIX/math.h/linknamespace
FAIL: conform/POSIX/mqueue.h/conform
FAIL: conform/POSIX/signal.h/conform
FAIL: conform/POSIX/termios.h/conform
FAIL: conform/POSIX2008/termios.h/conform
FAIL: conform/UNIX98/math.h/linknamespace
FAIL: conform/UNIX98/termios.h/conform
FAIL: conform/XPG3/math.h/linknamespace
FAIL: conform/XPG4/math.h/linknamespace
FAIL: elf/check-abi-libc
FAIL: elf/check-execstack
FAIL: elf/check-localplt
FAIL: elf/tst-audit1
FAIL: elf/tst-audit2
FAIL: elf/tst-audit8
FAIL: elf/tst-audit9
FAIL: inet/test-ifaddrs
FAIL: inet/test_ifindex
FAIL: math/atest-exp2
FAIL: math/atest-sincos
FAIL: rt/tst-cputimer1
= Fail for O32/N32 =
FAIL: conform/POSIX2008/mqueue.h/conform
FAIL: conform/UNIX98/mqueue.h/conform
FAIL: conform/XOPEN2K8/mqueue.h/conform
= Fail for O32 =
FAIL: conform/POSIX/sys/stat.h/conform
FAIL: conform/POSIX2008/sys/stat.h/conform
FAIL: conform/UNIX98/sys/stat.h/conform
FAIL: conform/XOPEN2K/sys/stat.h/conform
FAIL: conform/XOPEN2K8/sys/stat.h/conform
>
> > Any further feedback welcome... Assuming tests do turn out with the
> > same PASS rate then I'd like to commit?
>
> OK with changes as noted below:
>
> > +#if __mips_fpr != 0 || _MIPS_SPFPSET != 16 #error This test requires
> > +-mfpxx -mno-odd-spreg #endif
>
> Missing preprocessor indentation, "# error". Likewise in other tests.
Done throughout.
>
> > +static int mode[6] = { 0,
> > + PR_FP_MODE_FR,
> > + PR_FP_MODE_FR | PR_FP_MODE_FRE,
> > + PR_FP_MODE_FR,
> > + 0,
> > + PR_FP_MODE_FR | PR_FP_MODE_FRE };
>
> Should have braces on separate lines, along the lines of:
>
> static int mode[6] =
> {
> 0,
> ...
> };
>
> Likewise in other tests.
Done throughout.
> > + if ((float)i != f || (double)i != d)
>
> Missing spaces in casts, "(float) i", "(double) i".
Also done.
This is now committed. Thanks for the reviews.
I just saw your comment about a NEWS entry. I'll have to do that in
a follow up commit. I meant to ask you if a NEWS entry was
appropriate here. I'll post some text for review but it will have to
be Friday now.
Thanks,
Matthew