test-cancel1 failure

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

test-cancel1 failure

Andreas Jaeger

I configure on my x86-64 systems with:
/cvs/libc/configure --prefix=/usr --enable-add-ons --enable-kernel=2.6.16 --disable-profile

And just got on two different machines this failure:
GCONV_PATH=/abuild/aj/glibc/10.1/iconvdata LC_ALL=C   /abuild/aj/glibc/10.1/elf/ld-linux-x86-64.so.2 --library-path /abuild/aj/glibc/10.1:/abuild/aj/glibc/10.1/math:/abuild/aj/glibc/10.1/elf:/abuild/aj/glibc/10.1/dlfcn:/abuild/aj/glibc/10.1/nss:/abuild/aj/glibc/10.1/nis:/abuild/aj/glibc/10.1/rt:/abuild/aj/glibc/10.1/resolv:/abuild/aj/glibc/10.1/crypt:/abuild/aj/glibc/10.1/nptl /abuild/aj/glibc/10.1/nptl/tst-cancel1  > /abuild/aj/glibc/10.1/nptl/tst-cancel1.out
Didn't expect signal from child: got `Aborted'

strace gives:
open("/lib64/libgcc_s.so.1", O_RDONLY)  = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340\35\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=55616, ...}) = 0
mmap(NULL, 1100424, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2aaaaaad3000
mprotect(0x2aaaaaae0000, 1044480, PROT_NONE) = 0
mmap(0x2aaaaabdf000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0xc000) = 0x2aaaaabdf000
close(3)                                = 0
munmap(0x2aaaaaaab000, 161794)          = 0
tgkill(32406, 32407, SIGRTMIN)          = 0
futex(0x408009d0, FUTEX_WAIT, 32407, NULL) = -1 EINTR (Interrupted system call)
+++ killed by SIGABRT +++

All the other tests pass.  Does anybody have an idea?

Andreas
--
 Andreas Jaeger, [hidden email], http://www.suse.de/~aj/
  SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
   GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126

attachment0 (194 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: test-cancel1 failure

Klaus Dittrich
Andreas Jaeger wrote:

> I configure on my x86-64 systems with:
> /cvs/libc/configure --prefix=/usr --enable-add-ons --enable-kernel=2.6.16 --disable-profile
>
> And just got on two different machines this failure:
> GCONV_PATH=/abuild/aj/glibc/10.1/iconvdata LC_ALL=C   /abuild/aj/glibc/10.1/elf/ld-linux-x86-64.so.2 --library-path /abuild/aj/glibc/10.1:/abuild/aj/glibc/10.1/math:/abuild/aj/glibc/10.1/elf:/abuild/aj/glibc/10.1/dlfcn:/abuild/aj/glibc/10.1/nss:/abuild/aj/glibc/10.1/nis:/abuild/aj/glibc/10.1/rt:/abuild/aj/glibc/10.1/resolv:/abuild/aj/glibc/10.1/crypt:/abuild/aj/glibc/10.1/nptl /abuild/aj/glibc/10.1/nptl/tst-cancel1  > /abuild/aj/glibc/10.1/nptl/tst-cancel1.out
> Didn't expect signal from child: got `Aborted'
>
> strace gives:
> open("/lib64/libgcc_s.so.1", O_RDONLY)  = 3
> read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340\35\0"..., 832) = 832
> fstat(3, {st_mode=S_IFREG|0755, st_size=55616, ...}) = 0
> mmap(NULL, 1100424, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2aaaaaad3000
> mprotect(0x2aaaaaae0000, 1044480, PROT_NONE) = 0
> mmap(0x2aaaaabdf000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0xc000) = 0x2aaaaabdf000
> close(3)                                = 0
> munmap(0x2aaaaaaab000, 161794)          = 0
> tgkill(32406, 32407, SIGRTMIN)          = 0
> futex(0x408009d0, FUTEX_WAIT, 32407, NULL) = -1 EINTR (Interrupted system call)
> +++ killed by SIGABRT +++
>
> All the other tests pass.  Does anybody have an idea?
>
> Andreas


Same observation on X86-32. (2xP4, kernel 2.6.17.3)
--
Klaus
Reply | Threaded
Open this post in threaded view
|

Re: test-cancel1 failure

Jakub Jelinek
In reply to this post by Andreas Jaeger
On Thu, Sep 21, 2006 at 05:52:25AM +0200, Andreas Jaeger wrote:
>
> I configure on my x86-64 systems with:
> /cvs/libc/configure --prefix=/usr --enable-add-ons --enable-kernel=2.6.16 --disable-profile
>
> And just got on two different machines this failure:
> GCONV_PATH=/abuild/aj/glibc/10.1/iconvdata LC_ALL=C   /abuild/aj/glibc/10.1/elf/ld-linux-x86-64.so.2 --library-path /abuild/aj/glibc/10.1:/abuild/aj/glibc/10.1/math:/abuild/aj/glibc/10.1/elf:/abuild/aj/glibc/10.1/dlfcn:/abuild/aj/glibc/10.1/nss:/abuild/aj/glibc/10.1/nis:/abuild/aj/glibc/10.1/rt:/abuild/aj/glibc/10.1/resolv:/abuild/aj/glibc/10.1/crypt:/abuild/aj/glibc/10.1/nptl /abuild/aj/glibc/10.1/nptl/tst-cancel1  > /abuild/aj/glibc/10.1/nptl/tst-cancel1.out
> Didn't expect signal from child: got `Aborted'

Does your libgcc_s.so.1 support DW_CFA_val_{offset{,_sf},expression}?
On i?86/x86-64 for glibc 2.5 you need an unwinder (and gdb)
that can handle those.

See
http://gcc.gnu.org/ml/gcc-patches/2006-03/msg00208.html
http://sources.redhat.com/ml/gdb-patches/2006-03/msg00076.html
http://sources.redhat.com/ml/binutils/2006-03/msg00043.html
http://sources.redhat.com/ml/binutils/2006-03/msg00009.html

        Jakub
Reply | Threaded
Open this post in threaded view
|

Re: test-cancel1 failure

Andreas Jaeger
Jakub Jelinek <[hidden email]> writes:

> On Thu, Sep 21, 2006 at 05:52:25AM +0200, Andreas Jaeger wrote:
>>
>> I configure on my x86-64 systems with:
>> /cvs/libc/configure --prefix=/usr --enable-add-ons --enable-kernel=2.6.16 --disable-profile
>>
>> And just got on two different machines this failure:
>> GCONV_PATH=/abuild/aj/glibc/10.1/iconvdata LC_ALL=C   /abuild/aj/glibc/10.1/elf/ld-linux-x86-64.so.2 --library-path /abuild/aj/glibc/10.1:/abuild/aj/glibc/10.1/math:/abuild/aj/glibc/10.1/elf:/abuild/aj/glibc/10.1/dlfcn:/abuild/aj/glibc/10.1/nss:/abuild/aj/glibc/10.1/nis:/abuild/aj/glibc/10.1/rt:/abuild/aj/glibc/10.1/resolv:/abuild/aj/glibc/10.1/crypt:/abuild/aj/glibc/10.1/nptl /abuild/aj/glibc/10.1/nptl/tst-cancel1  > /abuild/aj/glibc/10.1/nptl/tst-cancel1.out
>> Didn't expect signal from child: got `Aborted'
>
> Does your libgcc_s.so.1 support DW_CFA_val_{offset{,_sf},expression}?
> On i?86/x86-64 for glibc 2.5 you need an unwinder (and gdb)
> that can handle those.
>
> See
> http://gcc.gnu.org/ml/gcc-patches/2006-03/msg00208.html
> http://sources.redhat.com/ml/gdb-patches/2006-03/msg00076.html
> http://sources.redhat.com/ml/binutils/2006-03/msg00043.html
> http://sources.redhat.com/ml/binutils/2006-03/msg00009.html
>
> Jakub
I doubt that I have the gcc patch in my toolchain.

thanks for the info,
Andreas
--
 Andreas Jaeger, [hidden email], http://www.suse.de/~aj/
  SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
   GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126

attachment0 (194 bytes) Download Attachment