Bug#219352: xmms libc crash

2003-11-12 Thread Felix Seeger
On Thursday 13 November 2003 04:12, Daniel Jacobowitz wrote:
> On Thu, Nov 13, 2003 at 02:54:40AM +0100, Felix Seeger wrote:
> > open("/usr/lib/tls/libGL.so.1", O_RDONLY) = 8
> > read(8, "[EMAIL PROTECTED]"..., 512)
> > = 512
> > fstat64(8, {st_mode=S_IFREG|0755, st_size=430820, ...}) = 0
> > writev(2, [{"Inconsistency detected by ld.so:"..., 33},
> > {"../sysdeps/generic/dl-tls.c", 27}, {": ", 2}, {"72", 2}, {": ", 2},
> > {"_dl_next_tls_modid", 18}, {": ", 2}, {"Assertion `", 11}, {"result <=
> > _rtld_local._dl_tls_ma"..., 41}, {"\' failed!\n", 10}], 10Inconsistency
> > detected by ld.so: ../sysdeps/generic/dl-tls.c: 72: _dl_next_tls_modid:
> > Assertion `result <= _rtld_local._dl_tls_max_dtv_idx' failed!
> > ) = 148
> > exit_group(127) = ?
>
> Then this bug is almost certainly related to the nvidia-glx drivers.
> Either as a libc bug or a TLS problem; it's hard to say without
> investigating more but that may let Goto-san reproduce it?

Yes, if I move libGL.so... to another place I get:
libmikmod.so.2: cannot open shared object file: No such file or directory
/usr/lib/tls/libGLcore.so.1: undefined symbol: __gl_tls_var0
/usr/lib/tls/libGLcore.so.1: undefined symbol: __gl_tls_var0

But xmms starts.

strings /usr/lib/tls/libGLcore.so.1 | grep nvidia
nvidia id: NVIDIA OpenGL Core Shared Library (libGLcore) (ELF TLS)  1.0-4496  
Wed Jul 16 19:52:36 PDT 2003

thanks
Felix





Bug#219352: xmms libc crash

2003-11-12 Thread Felix Seeger
On Thursday 13 November 2003 04:12, Daniel Jacobowitz wrote:
> On Thu, Nov 13, 2003 at 02:54:40AM +0100, Felix Seeger wrote:
> > open("/usr/lib/tls/libGL.so.1", O_RDONLY) = 8
> > read(8, "[EMAIL PROTECTED]"..., 512)
> > = 512
> > fstat64(8, {st_mode=S_IFREG|0755, st_size=430820, ...}) = 0
> > writev(2, [{"Inconsistency detected by ld.so:"..., 33},
> > {"../sysdeps/generic/dl-tls.c", 27}, {": ", 2}, {"72", 2}, {": ", 2},
> > {"_dl_next_tls_modid", 18}, {": ", 2}, {"Assertion `", 11}, {"result <=
> > _rtld_local._dl_tls_ma"..., 41}, {"\' failed!\n", 10}], 10Inconsistency
> > detected by ld.so: ../sysdeps/generic/dl-tls.c: 72: _dl_next_tls_modid:
> > Assertion `result <= _rtld_local._dl_tls_max_dtv_idx' failed!
> > ) = 148
> > exit_group(127) = ?
>
> Then this bug is almost certainly related to the nvidia-glx drivers.
> Either as a libc bug or a TLS problem; it's hard to say without
> investigating more but that may let Goto-san reproduce it?

Yes, if I move libGL.so... to another place I get:
libmikmod.so.2: cannot open shared object file: No such file or directory
/usr/lib/tls/libGLcore.so.1: undefined symbol: __gl_tls_var0
/usr/lib/tls/libGLcore.so.1: undefined symbol: __gl_tls_var0

But xmms starts.

strings /usr/lib/tls/libGLcore.so.1 | grep nvidia
nvidia id: NVIDIA OpenGL Core Shared Library (libGLcore) (ELF TLS)  1.0-4496  
Wed Jul 16 19:52:36 PDT 2003

thanks
Felix



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]



Bug#219352: xmms libc crash

2003-11-12 Thread Felix Seeger
On Thursday 13 November 2003 02:11, GOTO Masanori wrote:
> At Tue, 11 Nov 2003 20:23:58 +0100,
> Felix Seeger wrote:
>
> > On Tuesday 11 November 2003 18:19, Juergen Kreileder wrote:
> >
> > > Jeff Bailey <[EMAIL PROTECTED]> writes:
> > >
> > > > On Wed, Nov 12, 2003 at 12:49:38AM +0900, GOTO Masanori wrote:
> > > >
> > > >> I also tested on both 2.4 and 2.6 kernel, even with removing
> > > >> ~/.xmms.  However I cannot reproduce it...  Which CPU do you use?
> >
> > I use an AMD Athlon XP 2500+ on an Asus nforce2 board with nvidia
> > drivers.
 
> >
> > > > I also cannot reproduce 2.4 kernel on k7 using Debian's package, and
> > > > Pentium 2 Xeon, using Debian's 2.6.0-test9 kernel package.
> > >
> > >
> > > The original bug report says "Unless libmikmod2 is installed [...]".
> > > (xmms recommends libmikmod2.)
> > >
> > > If I move /usr/lib/libmikmod.so.2* out of the way I can reproduce
> > > this problem with 2.6.0-test9-mm2:
> >
> > Yes, if I install libmikmod2 xmms starts up normally.
>
> 
> This problem is occured under:
> 
>   - kernel 2.6.0-test9
>   - glibc 2.3.2.ds1-9
2.3.2.ds1-10, but I think it also happend with -9

>   - CPU is not related?
Don't know

>   - xmms 1.2.8-2
>   - libmikmod 3.1.10-5
If not installed

>   - you might not install libc6-i686
yes, not installed
>   - we use unstable sid.
yes

> I use such environment, but I cannot reproduce this problem...  Please
> check your environment settings, and if you can, please track with gdb
> and strace.  This bug may be downgraded to important...

I remove libmikmod2 and run xmms with gdb, but when I type bt, there isn't 
one. Here is the strace output from the libmikmod warning on:

write(2, "libmikmod.so.2: cannot open shar"..., 74libmikmod.so.2: cannot open 
shared object file: No such file or directory
) = 74
stat64("/usr/lib/xmms/Input/libcdaudio.so", {st_mode=S_IFREG|0644, 
st_size=55440, ...}) = 0
open("/usr/lib/xmms/Input/libcdaudio.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0 2\0\000"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=55440, ...}) = 0
old_mmap(NULL, 56736, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x4109c000
old_mmap(0x410a9000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 
0xd000) = 0x410a9000
close(8)= 0
stat64("/usr/lib/xmms/Input/libtonegen.so", {st_mode=S_IFREG|0644, 
st_size=8508, ...}) = 0
open("/usr/lib/xmms/Input/libtonegen.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\340\t\0"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=8508, ...}) = 0
old_mmap(NULL, 11584, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x410aa000
old_mmap(0x410ac000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 
0x1000) = 0x410ac000
close(8)= 0
getdents64(7, /* 0 entries */, 131072)  = 0
close(7)= 0
open("/usr/lib/xmms/Effect", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY) = 7
fstat64(7, {st_mode=S_IFDIR|0755, st_size=144, ...}) = 0
fcntl64(7, F_SETFD, FD_CLOEXEC) = 0
getdents64(7, /* 5 entries */, 131072)  = 144
stat64("/usr/lib/xmms/Effect/.", {st_mode=S_IFDIR|0755, st_size=144, ...}) = 0
stat64("/usr/lib/xmms/Effect/..", {st_mode=S_IFDIR|0755, st_size=176, ...}) = 
0
stat64("/usr/lib/xmms/Effect/libvoice.so", {st_mode=S_IFREG|0644, 
st_size=5064, ...}) = 0
open("/usr/lib/xmms/Effect/libvoice.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\240\7\0"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=5064, ...}) = 0
old_mmap(NULL, 8128, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x410ad000
old_mmap(0x410ae000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 0) 
= 0x410ae000
close(8)= 0
stat64("/usr/lib/xmms/Effect/libstereo.so", {st_mode=S_IFREG|0644, 
st_size=8620, ...}) = 0
open("/usr/lib/xmms/Effect/libstereo.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360\r\0"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=8620, ...}) = 0
old_mmap(NULL, 11692, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x410af000
old_mmap(0x410b1000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 
0x1000) = 0x410b1000
close(8)= 0
stat64("/usr/lib/xmms/Effect/libecho.so", {st_mode=S_IFREG|0644, 
st_size=12400, ...}) = 0
open("/usr/lib/xmms/Effect/libecho.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0

Bug#219352: xmms libc crash

2003-11-12 Thread Felix Seeger
On Thursday 13 November 2003 02:11, GOTO Masanori wrote:
> At Tue, 11 Nov 2003 20:23:58 +0100,
> Felix Seeger wrote:
>
> > On Tuesday 11 November 2003 18:19, Juergen Kreileder wrote:
> >
> > > Jeff Bailey <[EMAIL PROTECTED]> writes:
> > >
> > > > On Wed, Nov 12, 2003 at 12:49:38AM +0900, GOTO Masanori wrote:
> > > >
> > > >> I also tested on both 2.4 and 2.6 kernel, even with removing
> > > >> ~/.xmms.  However I cannot reproduce it...  Which CPU do you use?
> >
> > I use an AMD Athlon XP 2500+ on an Asus nforce2 board with nvidia
> > drivers.
 
> >
> > > > I also cannot reproduce 2.4 kernel on k7 using Debian's package, and
> > > > Pentium 2 Xeon, using Debian's 2.6.0-test9 kernel package.
> > >
> > >
> > > The original bug report says "Unless libmikmod2 is installed [...]".
> > > (xmms recommends libmikmod2.)
> > >
> > > If I move /usr/lib/libmikmod.so.2* out of the way I can reproduce
> > > this problem with 2.6.0-test9-mm2:
> >
> > Yes, if I install libmikmod2 xmms starts up normally.
>
> 
> This problem is occured under:
> 
>   - kernel 2.6.0-test9
>   - glibc 2.3.2.ds1-9
2.3.2.ds1-10, but I think it also happend with -9

>   - CPU is not related?
Don't know

>   - xmms 1.2.8-2
>   - libmikmod 3.1.10-5
If not installed

>   - you might not install libc6-i686
yes, not installed
>   - we use unstable sid.
yes

> I use such environment, but I cannot reproduce this problem...  Please
> check your environment settings, and if you can, please track with gdb
> and strace.  This bug may be downgraded to important...

I remove libmikmod2 and run xmms with gdb, but when I type bt, there isn't 
one. Here is the strace output from the libmikmod warning on:

write(2, "libmikmod.so.2: cannot open shar"..., 74libmikmod.so.2: cannot open 
shared object file: No such file or directory
) = 74
stat64("/usr/lib/xmms/Input/libcdaudio.so", {st_mode=S_IFREG|0644, 
st_size=55440, ...}) = 0
open("/usr/lib/xmms/Input/libcdaudio.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0 2\0\000"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=55440, ...}) = 0
old_mmap(NULL, 56736, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x4109c000
old_mmap(0x410a9000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 
0xd000) = 0x410a9000
close(8)= 0
stat64("/usr/lib/xmms/Input/libtonegen.so", {st_mode=S_IFREG|0644, 
st_size=8508, ...}) = 0
open("/usr/lib/xmms/Input/libtonegen.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\340\t\0"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=8508, ...}) = 0
old_mmap(NULL, 11584, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x410aa000
old_mmap(0x410ac000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 
0x1000) = 0x410ac000
close(8)= 0
getdents64(7, /* 0 entries */, 131072)  = 0
close(7)= 0
open("/usr/lib/xmms/Effect", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY) = 7
fstat64(7, {st_mode=S_IFDIR|0755, st_size=144, ...}) = 0
fcntl64(7, F_SETFD, FD_CLOEXEC) = 0
getdents64(7, /* 5 entries */, 131072)  = 144
stat64("/usr/lib/xmms/Effect/.", {st_mode=S_IFDIR|0755, st_size=144, ...}) = 0
stat64("/usr/lib/xmms/Effect/..", {st_mode=S_IFDIR|0755, st_size=176, ...}) = 
0
stat64("/usr/lib/xmms/Effect/libvoice.so", {st_mode=S_IFREG|0644, 
st_size=5064, ...}) = 0
open("/usr/lib/xmms/Effect/libvoice.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\240\7\0"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=5064, ...}) = 0
old_mmap(NULL, 8128, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x410ad000
old_mmap(0x410ae000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 0) 
= 0x410ae000
close(8)= 0
stat64("/usr/lib/xmms/Effect/libstereo.so", {st_mode=S_IFREG|0644, 
st_size=8620, ...}) = 0
open("/usr/lib/xmms/Effect/libstereo.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360\r\0"..., 512) = 
512
fstat64(8, {st_mode=S_IFREG|0644, st_size=8620, ...}) = 0
old_mmap(NULL, 11692, PROT_READ|PROT_EXEC, MAP_PRIVATE, 8, 0) = 0x410af000
old_mmap(0x410b1000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 8, 
0x1000) = 0x410b1000
close(8)= 0
stat64("/usr/lib/xmms/Effect/libecho.so", {st_mode=S_IFREG|0644, 
st_size=12400, ...}) = 0
open("/usr/lib/xmms/Effect/libecho.so", O_RDONLY) = 8
read(8, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0

Bug#219352: xmms libc crash

2003-11-11 Thread Felix Seeger
On Tuesday 11 November 2003 18:19, Juergen Kreileder wrote:
> Jeff Bailey <[EMAIL PROTECTED]> writes:
> > On Wed, Nov 12, 2003 at 12:49:38AM +0900, GOTO Masanori wrote:
> >> I also tested on both 2.4 and 2.6 kernel, even with removing
> >> ~/.xmms.  However I cannot reproduce it...  Which CPU do you use?
I use an AMD Athlon XP 2500+ on an Asus nforce2 board with nvidia drivers.

> > I also cannot reproduce 2.4 kernel on k7 using Debian's package, and
> > Pentium 2 Xeon, using Debian's 2.6.0-test9 kernel package.
>
> The original bug report says "Unless libmikmod2 is installed [...]".
> (xmms recommends libmikmod2.)
>
> If I move /usr/lib/libmikmod.so.2* out of the way I can reproduce
> this problem with 2.6.0-test9-mm2:
Yes, if I install libmikmod2 xmms starts up normally.

> % xmms
> libmikmod.so.2: cannot open shared object file: No such file or directory
> Inconsistency detected by ld.so: ../sysdeps/generic/dl-tls.c: 72:
> _dl_next_tls_modid: Assertion `result <= _rtld_local._dl_tls_max_dtv_idx'
> failed!
>
>
> I also can reproduce the problem mentioned above:
>
> % /lib/ld-2.3.2.so /lib/libc-2.3.2.so
> Inconsistency detected by ld.so: rtld.c: 1259: dl_main: Assertion
> `_rtld_local._dl_rtld_map.l_prev->l_next ==
> _rtld_local._dl_rtld_map.l_next' failed! % /lib/ld-2.3.2.so
> /lib/tls/libc-2.3.2.so
[...]

> BTW, why aren't the executable bits set for libc-2.3.2.so in the
> current glibc packages?

This is new since the -ds libc packages I think. Until now I thought it is to 
hide the bug ;)

thanks
have fun
Felix




Bug#219352: xmms libc crash

2003-11-11 Thread Felix Seeger
On Tuesday 11 November 2003 18:19, Juergen Kreileder wrote:
> Jeff Bailey <[EMAIL PROTECTED]> writes:
> > On Wed, Nov 12, 2003 at 12:49:38AM +0900, GOTO Masanori wrote:
> >> I also tested on both 2.4 and 2.6 kernel, even with removing
> >> ~/.xmms.  However I cannot reproduce it...  Which CPU do you use?
I use an AMD Athlon XP 2500+ on an Asus nforce2 board with nvidia drivers.

> > I also cannot reproduce 2.4 kernel on k7 using Debian's package, and
> > Pentium 2 Xeon, using Debian's 2.6.0-test9 kernel package.
>
> The original bug report says "Unless libmikmod2 is installed [...]".
> (xmms recommends libmikmod2.)
>
> If I move /usr/lib/libmikmod.so.2* out of the way I can reproduce
> this problem with 2.6.0-test9-mm2:
Yes, if I install libmikmod2 xmms starts up normally.

> % xmms
> libmikmod.so.2: cannot open shared object file: No such file or directory
> Inconsistency detected by ld.so: ../sysdeps/generic/dl-tls.c: 72:
> _dl_next_tls_modid: Assertion `result <= _rtld_local._dl_tls_max_dtv_idx'
> failed!
>
>
> I also can reproduce the problem mentioned above:
>
> % /lib/ld-2.3.2.so /lib/libc-2.3.2.so
> Inconsistency detected by ld.so: rtld.c: 1259: dl_main: Assertion
> `_rtld_local._dl_rtld_map.l_prev->l_next ==
> _rtld_local._dl_rtld_map.l_next' failed! % /lib/ld-2.3.2.so
> /lib/tls/libc-2.3.2.so
[...]

> BTW, why aren't the executable bits set for libc-2.3.2.so in the
> current glibc packages?

This is new since the -ds libc packages I think. Until now I thought it is to 
hide the bug ;)

thanks
have fun
Felix


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]