Skip to contentSkip to conversation

glibc cvs: additional patche

parisc-linux Joel Soete 3 participants 35 messages Mar 7, 2004 – Apr 2, 2004

lore
Share & tools
lore

Joel Soete

author
#0 lore raw

Hi Carlos,

Since yesterday cvs update, i also need a patch similar to:
<http://sources.redhat.com/cgi-bin/cvsweb.cgi/libc/sysdeps/s390/s390-32/dl-machine.h.diff?r1=1.19&r2=1.20&cvsroot=glibc>

=========><=========

hppa/dl-machine.h 2004-03-07 12:34:06.000000000 +0100 renamed from hppa/dl-machine.h.orig 2004-03-07 12:26:36.000000000 +0100 +4 -4
@@ -252,7 +252,7 @@
                 }
               else
                {
-                 if (_dl_name_match_p (GL(dl_profile), l))
+                 if (_dl_name_match_p (GLRO(dl_profile), l))
                    {
                      /* This is the object we are looking for.  Say that
                         we really want profiling and the timers are
@@ -581,9 +581,9 @@
 static inline void __attribute__ ((unused))
 dl_platform_init (void)
 {
-       if (GL(dl_platform) != NULL && *GL(dl_platform) == '\0')
+       if (GLRO(dl_platform) != NULL && *GLRO(dl_platform) == '\0')
        /* Avoid an empty string which would disturb us.  */
-               GL(dl_platform) = NULL;
+               GLRO(dl_platform) = NULL;
 }
 
 #endif /* !dl_machine_h */
@@ -684,7 +684,7 @@
        break;
       if (__builtin_expect (sym->st_size > refsym->st_size, 0)
          || (__builtin_expect (sym->st_size < refsym->st_size, 0)
-             && __builtin_expect (GL(dl_verbose), 0)))
+             && __builtin_expect (GLRO(dl_verbose), 0)))
        {
          const char *strtab;
 

=========><=========

hth,
Joel

#1 lore raw
Quoted 1 line · Joel Soete

Since yesterday cvs update, i also need a patch similar to:

I see changes required to dl-runtime.c aswell.

c.

author
#2 lore raw
Quoted 2 lines

I see changes required to dl-runtime.c aswell.

Oops I miss somehing,

Thanks,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

#3 lore raw

[parisc-linux] New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 1 line · Joel Soete

Oops I miss somehing,

New patches here, passes testsuite without regression.

http://www.parisc-linux.org/~carlos/glibc-hppa-patches-2004-03-08.tar.gz

Joel,

Is there any chance that you could do some testing for me please?

1. Use debootstrap to build a chroot (/mnt/chroot)
2. Compile glibc with --prefix=/mnt/chroot
3. Install the new glibc there.
4. Test out running some stuff from the chroot.
o Try running Xnest
o Try running X applications through Xnest
o Try building a new kernel in the chroot.

Thanks!

c.

author
#4 lore raw

[parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 4 lines

New patches here, passes testsuite without regression.

http://www.parisc-linux.org/~carlos/glibc-hppa-patches-2004-03-08.tar.gz

Thanks :)

Quoted 3 lines

Joel,

Is there any chance that you could do some testing for me please?

Just let me find a free disk :)

I will advise ;)

Quoted 1 line

1. Use debootstrap to build a chroot (/mnt/chroot)

BTW which install should I install: man debootstrap give the fllowing example deboostrap woody ./woody-chroot http://ftp.cx.debian.org/debian

Well not familiar with debootstrap but I presume that ./woody-chroot would be here /mnt/chroot and ftp.cx.debian.org my nearest debian repository. But my question is: would you like I build a woody or would you prefer a sarge?

Quoted 3 lines

2. Compile glibc with --prefix=/mnt/chroot
3. Install the new glibc there.
4. Test out running some stuff from the chroot.

o Try running Xnest
o Try running X applications through Xnest
o Try building a new kernel in the chroot.

Clear

Quoted 1 line

Thanks!

wellcome,

Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

author
#5 lore raw

RE: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Carlos,

Quoted 7 lines
Quoted 3 lines

Joel,

Is there any chance that you could do some testing for me please?

Just let me find a free disk :)

I will advise ;)

Good new I have one (just met a small pb with se terminator: apparently fixed now).

Quoted 10 lines
Quoted 1 line

1. Use debootstrap to build a chroot (/mnt/chroot)

BTW which install should I install: man debootstrap give the fllowing example deboostrap woody ./woody-chroot http://ftp.cx.debian.org/debian

Well not familiar with debootstrap but I presume that ./woody-chroot would be here /mnt/chroot and ftp.cx.debian.org my nearest debian repository.

But my question is: would you like I build a woody or would you prefer a sarge?

For the moment I tested with sarge ended with:
"I: Base system installed successfully."

:)

let me know anyway if you would prefer woody, it doesn't required so much time :)

Quoted 1 line
Quoted 1 line

2. Compile glibc with --prefix=/mnt/chroot

I am going on...

Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

#6 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 5 lines · Joel Soete
Quoted 4 lines

BTW which install should I install: man debootstrap give the fllowing example deboostrap woody ./woody-chroot http://ftp.cx.debian.org/debian
But my question is: would you like I build a woody or would you prefer a sarge?

Quoted 7 lines

For the moment I tested with sarge ended with:
"I: Base system installed successfully."

:)

let me know anyway if you would prefer woody, it doesn't required so much time :)

Joel,

Both unstable, testing, and stable should technically be tested with the new glibc :) You can test with whichever you like right now.

Never *ever* install glibc on a running system, always install to the chroot, and jump into the chroot.

c.

author
#7 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Hi Carlos,

Quoted 5 lines

Both unstable, testing, and stable should technically be tested with the new glibc :) You can test with whichever you like right now.

Never *ever* install glibc on a running system, always install to the chroot, and jump into the chroot.

Yes I read the hot but interesting thread "Latest Glibc from CVS has segmentation problems"

Thanks for remind,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

author
#8 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

In reply to Carlos O'Donell

Hi Carlos,

Here are the news:

(first of all sorry for delay but I have temporarely have to exchange my N4k for a b180 so it take a bit more time)

The glibc cvs 2.3.3 cvs dated 20040308 + your patch compiles well (just this time I enable profiling) :).

I also install it on /mnt/chroot (according James method, just to test).

As I foreseen to use this disk to other test on the N4k, I grab the list of packages installed (dpkg --get-selection > /mnt/chroot/var/tmp/Selection.doc) and chroot to /mnt/chroot to install previously mentioned pkg list. All works fine (just 3 pkg failled to install because no /proc available in chroot).

In the same time I drive a make -k check and just get additional Errors:
[...]
make[2]: *** [/Develop/parisc-linux/build/glibc/libio/tst-mmap-eofsync.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/libio/tst-mmap-fflushsync.out] Error 1
make[1]: *** [libio/tests] Error 2
make[2]: [/Develop/parisc-linux/build/glibc/posix/annexc.out] Error 1 (ignored) make[2]: *** [/Develop/parisc-linux/build/glibc/elf/multiload.o] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/elf/origtest.o] Error 1 [...]

I have two hypothesis:
a. it is the first time I added enable-profile;
b. I forget to build first gcc(3.3) tool with your gcc patch (gcc/config/pa/fptr.c)

Now, 2.6.4-rc3-pa0 is just available, I will so try first to build it in the chroot disk...

hth,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

author
#9 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Hi Carlos,

Quoted 2 lines

Now, 2.6.4-rc3-pa0 is just available, I will so try first to build it in the chroot disk...

And finaly, it compiles well and even boot well with the original config.

(I just try to boot with chroot disk but it seems that I still forget some config as nic and getty on ttyS0 :( ; so I will continu tomorrow and try to restart X on this b180 (I don't yet look for a quick and dirty howto use Xnest) ).

hth,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

#10 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

In reply to Joel Soete
Quoted 2 lines · Joel Soete

The glibc cvs 2.3.3 cvs dated 20040308 + your patch compiles well (just this time I enable profiling) :).

Excellent.

Quoted 4 lines

As I foreseen to use this disk to other test on the N4k, I grab the list of packages installed (dpkg --get-selection > /mnt/chroot/var/tmp/Selection.doc) and chroot to /mnt/chroot to install previously mentioned pkg list. All works fine (just 3 pkg failled to install because no /proc available in chroot).

Try binding your chroot if you want to expose your process list to the chroot 'mount -o bind /proc /mnt/chroot/proc

Quoted 6 lines

In the same time I drive a make -k check and just get additional Errors:
[...]
make[2]: *** [/Develop/parisc-linux/build/glibc/libio/tst-mmap-eofsync.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/libio/tst-mmap-fflushsync.out] Error 1

Your kernel is too old, jejb fixed these.

Quoted 5 lines

make[1]: *** [libio/tests] Error 2
make[2]: [/Develop/parisc-linux/build/glibc/posix/annexc.out] Error 1 (ignored) make[2]: *** [/Develop/parisc-linux/build/glibc/elf/multiload.o] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/elf/origtest.o] Error 1 [...]

The 'annexc.out' is an expected failure. The other two failures are new and unknown, they aren't segfaulting which is good (Error 1), they just aren't completing for some reason. Most likely they are taking too long and timeout before the 2 second maximum. If they *are* timing out, set TIMEOUTFACTOR=10 or 100 in your environment, this scales the test timeout.

Quoted 2 lines

I have two hypothesis:
a. it is the first time I added enable-profile;

No, profiling works.

Quoted 1 line

b. I forget to build first gcc(3.3) tool with your gcc patch (gcc/config/pa/fptr.c)

No, I fixed this in glibc such that gcc doesn't have to be recompiled. If I hand't the userspace would have been *completely* broken, since function pointer comparisons wouldn't have worked.

Quoted 2 lines

Now, 2.6.4-rc3-pa0 is just available, I will so try first to build it in the chroot disk...

That should fix the mmap failures, please check to see if the other two failures are timeout related.

c.

author
#11 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

C.,

Quoted 3 lines
Quoted 3 lines

As I foreseen to use this disk to other test on the N4k, I grab the list of packages installed (dpkg --get-selection > mnt/chroot/var/tmp/Selection.doc) and chroot to /mnt/chroot to install previously mentioned pkg list. All

works

Quoted 5 lines
Quoted 2 lines

fine (just 3 pkg failled to install because no /proc available in chroot .

Try binding your chroot if you want to expose your process list to the chroot 'mount -o bind /proc /mnt/chroot/proc

That works fine (thanks).

[snip]

Quoted 2 lines

That should fix the mmap failures, please check to see if the other two failures are timeout related.

I now rebuild a 2.6.4-pa1 with this same libc and reboot the system with this kernel (you had right it was still runing an old 2.4.21-pa2 :( ) and here are the make -k check results:
without TIMEOUTFACTOR:
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-float.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-double.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-ifloat.out] Error 1
make[1]: *** [math/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/tst-strtod.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/bug-strtod.out] Error 1
make[1]: *** [stdlib/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/posix/tst-regex] Error 1 make[1]: *** [posix/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex10] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex11] Error 1
make[1]: *** [linuxthreads/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/resolvfail] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-tls13.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-dlmodcount.out] Error 139
make[1]: *** [elf/tests] Error 2
make: *** [check] Error 2

with export TIMEOUTFACTOR=100

make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-float.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-double.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-ifloat.out] Error 1
make[1]: *** [math/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/tst-strtod.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/bug-strtod.out] Error 1
make[1]: *** [stdlib/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/posix/tst-regex] Error 1 make[1]: *** [posix/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex10] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex11] Error 1
make[1]: *** [linuxthreads/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/resolvfail] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-dlmodcount.out] Error 139
make[1]: *** [elf/tests] Error 2
make: *** [check] Error 2

So only [/Develop/parisc-linux/build/glibc/elf/tst-tls13.out] seems to be related to TIMEOUTFACTOR

hth,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

#12 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 5 lines · Joel Soete

with export TIMEOUTFACTOR=100

make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-float.out] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-double.out] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-ifloat.out] Error 1

Failure is OK.

Quoted 3 lines

make[1]: *** [math/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/tst-strtod.out] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/bug-strtod.out] Error 1

Failure is OK.

Quoted 2 lines

make[1]: *** [stdlib/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/posix/tst-regex] Error 1

I haven't seen this one before, what's the problem with the test?

Quoted 3 lines

make[1]: *** [posix/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex10] Error 1 make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex11] Error 1

These are new, usually means the compiler optimized something that it shouldn't have optimized. Can you see what the error is? e.g. Look at the out file.

Quoted 2 lines

make[1]: *** [linuxthreads/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/resolvfail] Error 1

That's bad.

Quoted 1 line

make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-dlmodcount.out] Error 139

This is worse, the program crashes.

Quoted 2 lines

make[1]: *** [elf/tests] Error 2
make: *** [check] Error 2

Yes. tst-tls13 should be the only failure affected by TIMEOUTFACTOR, though I wasn't sure about the system you were running so I suggested it.

c.

author
#13 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Hi C.,

Ok I just mentioned know error to be complete :)

Quoted 2 lines
Quoted 2 lines

make[1]: *** [stdlib/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/posix/tst-regex] Error

1

Quoted 5 lines

I haven't seen this one before, what's the problem with the rest?

Quoted 2 lines

make[1]: *** [posix/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex10] Error

1

Quoted 1 line
Quoted 1 line

make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex11] Error

1

Quoted 7 lines

These are new, usually means the compiler optimized something that it shouldn't have optimized. Can you see what the error is? e.g. Look at the out file.

Quoted 2 lines

make[1]: *** [linuxthreads/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/resolvfail] Error

1

Quoted 3 lines

That's bad.

Hmmm all output files realted to Error1 was empty because of:
hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/posix/tst-regex
-Wl,-dynamic-linker=/lib/ld.so.1
/Develop/parisc-linux/build/glibc/csu/crt1.o /Develop/parisc-linux/build/glibc/csu/crti.o
hppa-linux-gcc --print-file-name=crtbegin.o /Develop/parisc-linux/build/glibc/posix/tst-regex.o
/Develop/parisc-linux/build/glibc/rt/librt.so -Wl,-rpath-link=/Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:
/Develop/parisc-linux/build/glibc/nis:/Develop/parisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads /Develop/parisc-linux/build/glibc/libc.so.6 /Develop/parisc-linux/build/glibc/libc_nonshared.a
-lgcc hppa-linux-gcc --print-file-name=crtend.o /Develop/parisc-linux/build/glibc/csu/crtn.o
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_create'
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `_pthread_cleanup_pop'
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_once'
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `_pthread_cleanup_push'
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_atfork'
collect2: ld returned 1 exit status
make[2]: *** [/Develop/parisc-linux/build/glibc/posix/tst-regex] Error 1

hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/linuxthreads/ex10
-Wl,-dynamic-linker=/lib/ld.so.1
/Develop/parisc-linux/build/glibc/csu/crt1.o /Develop/parisc-linux/build/glibc/csu/crti.o
hppa-linux-gcc --print-file-name=crtbegin.o /Develop/parisc-linux/build/glibc/linuxthreads/ex10.o
/Develop/parisc-linux/build/glibc/linuxthreads/libpthread.so /Develop/parisc-linux/build/glibc/linuxthreads/libpthread_nonshared.a
/Develop/parisc-linux/build/glibc/rt/librt.so -Wl,-rpath-link=/Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:
/Develop/parisc-linux/build/glibc/nis:/Develop/parisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads /Develop/parisc-linux/build/glibc/libc.so.6 /Develop/parisc-linux/build/glibc/libc_nonshared.a
-lgcc hppa-linux-gcc --print-file-name=crtend.o /Develop/parisc-linux/build/glibc/csu/crtn.o
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_atfork'
collect2: ld returned 1 exit status
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex10] Error
1

hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/linuxthreads/ex11
-Wl,-dynamic-linker=/lib/ld.so.1
/Develop/parisc-linux/build/glibc/csu/crt1.o /Develop/parisc-linux/build/glibc/csu/crti.o
hppa-linux-gcc --print-file-name=crtbegin.o /Develop/parisc-linux/build/glibc/linuxthreads/ex11.o
/Develop/parisc-linux/build/glibc/linuxthreads/libpthread.so /Develop/parisc-linux/build/glibc/linuxthreads/libpthread_nonshared.a
/Develop/parisc-linux/build/glibc/rt/librt.so -Wl,-rpath-link=/Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:
/Develop/parisc-linux/build/glibc/nis:/Develop/parisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads /Develop/parisc-linux/build/glibc/libc.so.6 /Develop/parisc-linux/build/glibc/libc_nonshared.a
-lgcc hppa-linux-gcc --print-file-name=crtend.o /Develop/parisc-linux/build/glibc/csu/crtn.o
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_atfork'
collect2: ld returned 1 exit status
make[2]: *** [/Develop/parisc-linux/build/glibc/linuxthreads/ex11] Error
1

hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/elf/resolvfail
-Wl,-dynamic-linker=/lib/ld.so.1 /Develop/parisc-linux/build/glibc/csu/crt1.o
/Develop/parisc-linux/build/glibc/csu/crti.o hppa-linux-gcc --print-file-name=crtbegin.o
/Develop/parisc-linux/build/glibc/elf/resolvfail.o /Develop/parisc-linux/build/glibc/dlfcn/libdl.so.2
/Develop/parisc-linux/build/glibc/linuxthreads/libpthread_nonshared.a /Develop/parisc-linux/build/glibc/linuxthreads/libpthread.so
-Wl,-rpath-link=/Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:/Develop/parisc-linux/build/glibc/nis:/Develop
/parisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads
/Develop/parisc-linux/build/glibc/libc.so.6 /Develop/parisc-linux/build/glibc/libc_nonshared.a
-lgcc hppa-linux-gcc --print-file-name=crtend.o /Develop/parisc-linux/build/glibc/csu/crtn.o
/Develop/parisc-linux/build/glibc/linuxthreads/libpthread.so: undefined reference
to `_h_errno'
/Develop/parisc-linux/build/glibc/linuxthreads/libpthread.so: undefined reference
to `_res'
/Develop/parisc-linux/build/glibc/linuxthreads/libpthread.so: undefined reference
to `_errno'
collect2: ld returned 1 exit status
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/resolvfail] Error 1

Quoted 1 line
Quoted 1 line

make [2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-dlmodcount.out]

Error 139

Quoted 2 lines

This is worse, the program crashes.

Yes:(
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-dlmodcount.out] Error 139
do_page_fault() pid=9891 command='ld.so.1' type=15 address=0x401aca28 vm_start = 0x40177000, vm_end = 0x4017c000

YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001000000000000001111 Not tainted

r00-03  00000000 41029750 40137993 bff01640
r04-07  40173088 00000000 00000010 00000004
r08-11  41029750 40152edc 40152f1c 40152e30
r12-15  00000000 40152edc 40152ee8 40152e30
r16-19  40152e30 40152edc 40152ec0 40173088
r20-23  00000001 00000001 401aca24 00000001
r24-27  00000001 4102a5d8 41029770 00020ed4
r28-31  00000048 00000049 bff01740 40137993
sr0-3   0000039f 0000039f 00000000 0000039f
sr4-7   0000039f 0000039f 0000039f 0000039f

IASQ: 0000039f 0000039f IAOQ: 4013c65b 4013c65f
IIR: 0ec81093 ISR: 0000039f IOR: 401aca28
CPU: 0 CR30: 14584000 CR31: 1039c000
ORIG_R28: 00000000
IAOQ[0]: 0x4013c65b
IAOQ[1]: 0x4013c65f
RP(r2): 0x40137993

And here the only difference between this test and previous is the kernel :(

On Monday (I don't have access to those system during the week-end), I can try to compile a 2.4.25 and relaunch test. What is your opinion?

{ here are the packages upgraded today:
Get:1 http://ftp.de.debian.org unstable/main dpkg 1.10.20 [1123kB] Get:2 http://ftp.de.debian.org unstable/main login 1:4.0.3-21 [269kB] Get:3 http://ftp.de.debian.org unstable/main libncurses5-dev 5.4-2 [1087kB] Get:4 http://ftp.de.debian.org unstable/main libncurses5 5.4-2 [309kB] Get:5 http://ftp.de.debian.org unstable/main ncurses-bin 5.4-2 [236kB] Get:6 http://ftp.de.debian.org unstable/main ncurses-base 5.4-2 [11.1kB] Get:7 http://ftp.de.debian.org unstable/main debconf-i18n 1.4.16 [63.6kB] Get:8 http://ftp.de.debian.org unstable/main debconf 1.4.16 [98.5kB] Get:9 http://ftp.de.debian.org unstable/main dselect 1.10.20 [116kB] Get:10 http://ftp.de.debian.org unstable/main passwd 1:4.0.3-21 [491kB] Get:11 http://ftp.de.debian.org unstable/main bc 1.06-13 [69.7kB] Get:12 http://ftp.de.debian.org unstable/main dc 1.06-13 [60.2kB] Get:13 http://ftp.de.debian.org unstable/main dpkg-dev 1.10.20 [122kB] Get:14 http://ftp.de.debian.org unstable/main libncursesw5 5.4-2 [320kB] Get:15 http://ftp.de.debian.org unstable/main ncurses-term 5.4-2 [279kB] Get:16 http://ftp.de.debian.org unstable/main base-config 2.15 [261kB] Get:17 http://ftp.de.debian.org unstable/main debconf-utils 1.4.16 [32.5kB] Get:18 http://ftp.de.debian.org unstable/main libaudiofile-dev 0.2.6-2 [132kB] Get:19 http://ftp.de.debian.org unstable/main libaudiofile0 0.2.6-2 [85.5kB] Get:20 http://ftp.de.debian.org unstable/main libxft2 2.1.2-6 [54.7kB] Get:21 http://ftp.de.debian.org unstable/main libxml1 1:1.8.17-5 [249kB] Get:22 http://ftp.de.debian.org unstable/main tasksel 1.45 [130kB] }

Quoted 3 lines

Yes. tst-tls13 should be the only failure affected by TIMEOUTFACTOR, though I wasn't sure about the system you were running so I suggested it.

Yes it is a b180 (n4k not anymore available for the moment :( )

hth beter,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

#14 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 8 lines · Joel Soete

Hmmm all output files realted to Error1 was empty because of:
hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/posix/tst-regex
-Wl,-dynamic-linker=/lib/ld.so.1
/Develop/parisc-linux/build/glibc/csu/crt1.o /Develop/parisc-linux/build/glibc/csu/crti.o
hppa-linux-gcc --print-file-name=crtbegin.o /Develop/parisc-linux/build/glibc/posix/tst-regex.o
/Develop/parisc-linux/build/glibc/rt/librt.so -Wl,-rpath-link=/Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:
/Develop/parisc-linux/build/glibc/nis:/Develop/parisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads /Develop/parisc-linux/build/glibc/libc.so.6 /Develop/parisc-linux/build/glibc/libc_nonshared.a

Quoted 8 lines

-lgcc hppa-linux-gcc --print-file-name=crtend.o /Develop/parisc-linux/build/glibc/csu/crtn.o /Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to pthread_create' /Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to _pthread_cleanup_pop' /Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to pthread_once' /Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to _pthread_cleanup_push' /Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_atfork'
collect2: ld returned 1 exit status
make[2]: *** [/Develop/parisc-linux/build/glibc/posix/tst-regex] Error 1

Are you sure you thread library was built correctly? Do you see those symbols in libpthread? All those issues are related to libpthread being broken somehow.

Quoted 3 lines

make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-dlmodcount.out] Error 139 do_page_fault() pid=9891 command='ld.so.1' type=15 address=0x401aca28 vm_start = 0x40177000, vm_end = 0x4017c000

Can you run the loader through gdb, and give an assembly dump around the address where it crashes?

c.

author
#15 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 13 lines · Carlos O'Donell
Quoted 11 lines · Joel Soete

Hmmm all output files realted to Error1 was empty because of:
hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/posix/tst-regex
-Wl,-dynamic-linker=/lib/ld.so.1
/Develop/parisc-linux/build/glibc/csu/crt1.o /Develop/parisc-linux/build/glibc/csu/crti.o
hppa-linux-gcc --print-file-name=crtbegin.o /Develop/parisc-linux/build/glibc/posix/tst-regex.o
/Develop/parisc-linux/build/glibc/rt/librt.so -Wl,-rpath-link=/Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:
/Develop/parisc-linux/build/glibc/nis:/Develop/parisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads /Develop/parisc-linux/build/glibc/libc.so.6 /Develop/parisc-linux/build/glibc/libc_nonshared.a
-lgcc hppa-linux-gcc --print-file-name=crtend.o /Develop/parisc-linux/build/glibc/csu/crtn.o
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_create'
/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `_pthread_cleanup_pop'

Quoted 8 lines
Quoted 5 lines

/Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to pthread_once' /Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to _pthread_cleanup_push' /Develop/parisc-linux/build/glibc/rt/librt.so: undefined reference to `pthread_atfork'
collect2: ld returned 1 exit status
make[2]: *** [/Develop/parisc-linux/build/glibc/posix/tst-regex] Error 1

Are you sure you thread library was built correctly?

oops, I need more check on Monday (sorry) (btw I forget to transmit you my final patch about feupdate

Quoted 5 lines

Do you see those
symbols in libpthread? All those issues are related to libpthread being broken somehow.

Quoted 8 lines
Quoted 3 lines

make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-dlmodcount.out] Error 139 do_page_fault() pid=9891 command='ld.so.1' type=15 address=0x401aca28 vm_start = 0x40177000, vm_end = 0x4017c000

Can you run the loader through gdb, and give an assembly dump around the address where it crashes?

No pb if gdb don't segv (but iirc that only appened with gcc-3.5) and if lucky I could also garb a bt :).

Btw in previous test, I noticed that IAOQ was for instance =0x41007293 when gdb shows me the actual address of the crash at '0x41007293 - 3' and I find well the corresponding code at 7290 in an 'objdump -d'. Do you know why this diff of 3?

Sorry for delay, at Monday,
Joel

#16 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 7 lines

No pb if gdb don't segv (but iirc that only appened with gcc-3.5) and if lucky I could also garb a bt :).

Btw in previous test, I noticed that IAOQ was for instance =0x41007293 when gdb shows me the actual address of the crash at '0x41007293 - 3' and I find well the corresponding code at 7290 in an 'objdump -d'. Do you know why this diff of 3?

Instructions can only start at addresses that are modulo 32-bits, gdb is returning the start address of the insn that executed. While IAOQ is showing that the insn has finished (e.g. pointing at the last byte 0 1 2 3).

c.

#17 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 8 lines · Carlos O'Donell
Quoted 7 lines

No pb if gdb don't segv (but iirc that only appened with gcc-3.5) and if lucky I could also garb a bt :).

Btw in previous test, I noticed that IAOQ was for instance =0x41007293 when gdb shows me the actual address of the crash at '0x41007293 - 3' and I find well the corresponding code at 7290 in an 'objdump -d'. Do you know why this diff of 3?

Quoted 4 lines

Instructions can only start at addresses that are modulo 32-bits, gdb is returning the start address of the insn that executed. While IAOQ is showing that the insn has finished (e.g. pointing at the last byte 0 1 2 3).

Sorry, I'm smoking crack.

Address OR Privilege = IAOQ

Privilege Kernel = 0
Privilege User = 3

c.

author
#18 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

In reply to Joel Soete

Carlos,

I just noticed a border effect when I try to use James method: use prefix=/usr and make install_root=/mnt/chroot install;

Quoted 9 lines · Joel Soete
Quoted 5 lines · Carlos O'Donell
Quoted 3 lines · Joel Soete

Hmmm all output files realted to Error1 was empty because of:
hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/posix/tst-regex -Wl,-dynamic-linker=/lib/ld.so.1

                        ^^^^^^^^^^^^

When I use the previous method (prefix=/mnt/chroot) it become :

-dynamic-linker=/mnt/chroot/lib/ld.so.1

So here I can drive make -k check without pb in basic env (the one in which I build the new glibc)

OTC with James' method I would have to do it in chroot /mnt/chroot env.
And it wasn't what I did; I will so restart build and test with prefix=/mnt/chroot (as I use to do) and advise you of the new results ;)

That's a way to lurn ;)

Joel

#19 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

C

Quoted 7 lines · Joel Soete

arlos,

I just noticed a border effect when I try to use James method: use prefix=/usr and make install_root=/mnt/chroot install;

Quoted 2 lines · Joel Soete

Carlos O'Donell wrote:

Quoted 7 lines
Quoted 6 lines
Quoted 6 lines · Joel Soete
Quoted 5 lines

Hmmm all output files realted to Error1 was empty because of:
hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/posix/tst-regex -Wl,-dynamic-linker=/lib/ld.so.1

                         ^^^^^^^^^^^^

Are you running make check after make install?

Jim

author
#20 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 3 lines

Are you running make check after make install?

Yes, but as mentioned in the original build env and not in the chrooted env as it seems I would had to do.

hth,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

#21 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 7 lines · Joel Soete
Quoted 2 lines

Are you running make check after make install?

Yes, but as mentioned in the original build env and not in the chrooted env as it seems I would had to do.

hth,
Joel

Humm, this seems backwards to me. I wonder if make install mangles some
variables to look at the newly installed versions of ld.so, etc. Anyway, I'll be quiet for a while since I don't have a build on hand to check stuff out.

Jim

author
#22 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

In reply to Joel Soete

Hello Jim,

Sorry for confusion but you have definitely right.

[snip]use prefix=/usr and make install_root=/mnt/chroot install;

This is the right method because if I use prefix=/mnt/chroot and 'make install', when the install was finished and I tried chroot/mnt/chroot: here was the error message:
chroot /mnt/chroot
/bin/bash: error while loading shared libraries: libncurses.so.5: cannot open shared object file: No such file or directory

and strace shows me that it looking for libncurses.so.5 into /mnt/chroot/lib :)

So to launch make -k check in this config, I foreseen to use prefix=/usr and make install_root=/mnt/chroot; then mount -o bind /GlibcSrcFS /mnt/chroot/GlibcSrcFS; chroot /mnt/chroot; and in this env launch the make -k check. Do you know another solution?

Many thanks for advise,
Joel

Quoted 11 lines · Joel Soete
Quoted 1 line · Carlos O'Donell

On Fri, Mar 12, 2004 at 06:33:11PM +0100, Joel Soete wr

te:

Quoted 5 lines
Quoted 5 lines

Hmmm all output files realted to Error1 was empty because of:
hppa-linux-gcc -nostdlib -nostartfiles -o /Develop/parisc-linux/build/glibc/posix/tst-regex -Wl,-dynamic-linker=/lib/ld.so.1

                         ^^^^^^^^^^
Quoted 1 line

^

When I use the previous method (prefix=/mnt/chroot) it become :

-dynamic-linker=/mnt/chroot/lib/ld.so.1

So here I can drive make -k check without pb in basic env (the one in which I build the new glibc)

OTC with James' method I would have

Quoted 1 line

to do it in chroot /mnt/chroot env.

And it wasn't what I did; I will so restart build and test with prefix=/mnt/chroot (as I use to do) and advise you of the new
results ;)

That's a way to lurn ;)

Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

#23 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 13 lines · Joel Soete

Hello Jim,

Sorry for confusion but you have definitely right.

[snip]use prefix=/usr and make install_root=/mnt/chroot install;

This is the right method because if I use prefix=/mnt/chroot and 'make install', when the install was finished and I tried chroot/mnt/chroot: here was the error message:
chroot /mnt/chroot
/bin/bash: error while loading shared libraries: libncurses.so.5: cannot open shared object file: No such file or directory

Quoted 7 lines

and strace shows me that it looking for libncurses.so.5 into /mnt/chroot/lib :)

So to launch make -k check in this config, I foreseen to use prefix=/usr and make install_root=/mnt/chroot; then mount -o bind /GlibcSrcFS /mnt/chroot/GlibcSrcFS; chroot /mnt/chroot; and in this env launch the make -k check. Do you know another solution?

That sounds like a perfect solution! :)

Thanks Joel.

c.

author
#24 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Hello Carlos,

Sorry for delay,...

Quoted 9 lines
Quoted 9 lines · Joel Soete

Hello Jim,

Sorry for confusion but you have definitely right.

[snip]use prefix=/usr and make install_root=/mnt/chroot install;

This is the right method because if I use prefix=/mnt/chroot and 'make

install',

Quoted 1 line
Quoted 1 line

when the install was finished and I tried chroot/mnt/chroot: here was

the

Quoted 8 lines
Quoted 8 lines

error message:
chroot /mnt/chroot
/bin/bash: error while loading shared libraries: libncurses.so.5: cannot open shared object file: No such file or directory

and strace shows me that it looking for libncurses.so.5 into /mnt/chroot/lib :)

Quoted 6 lines
Quoted 4 lines

So to launch make -k check in this config, I foreseen to use prefix=/usr and make install_root=/mnt/chroot; then mount -o bind /GlibcSrcFS /mnt/chroot/GlibcSrcFS; chroot /mnt/chroot; and in this env launch the make -k check. Do you know another solution?

That sounds like a perfect solution! :)

but that way I only get know errors ie:
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-float.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-double.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-ifloat.out] Error 1
make[1]: *** [math/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/tst-strtod.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/bug-strtod.out] Error 1
make[1]: *** [stdlib/tests] Error 2
make[2]: [/Develop/parisc-linux/build/glibc/posix/annexc.out] Error 1 (ignored) make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-tls13.out] Error 1
make[1]: *** [elf/tests] Error 2
make: *** [check] Error 2

(sorry I forget TIMEOUTFACTOR) :)

Very nice job Carlos.

Just still have to check more X,

Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

#25 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 13 lines

but that way I only get know errors ie:
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-float.out] Error make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-double.out] Error make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-ifloat.out] Error make[1]: *** [math/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/tst-strtod.out] Error make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/bug-strtod.out] Error make[1]: *** [stdlib/tests] Error 2
make[2]: [/Develop/parisc-linux/build/glibc/posix/annexc.out] Error 1 (ignored) make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-tls13.out] Error make[1]: *** [elf/tests] Error 2
make: *** [check] Error 2

Quoted 1 line

(sorry I forget TIMEOUTFACTOR) :)

:) No problem!

Quoted 2 lines

Very nice job Carlos.
Just still have to check more X,

No, thank you Joel, your testing and bravery is an enourmous help :)

c.

author
#26 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Hello Carlos,

Quoted 2 lines
Quoted 2 lines

Very nice job Carlos.
Just still have to check more X,

I now boot on the (previously /mnt/chroot) disk with new glibc (to be accurate cvs 2004 03 15 + your patch + my feupdateenv) with kernel 2.6.4-pa1 first.

I try to rebuild 2.6.4-pa2 as well as 2.6.5-rc1-pa0 (but small pb) and so rebuild 2.6.4-pa1 to add kbd and mouse support: rebuild ok, reboot ok.

I also relaunch Xfce4 and do some test with its file manager, its configuration interface and xterm, all seems to works fine (well at least, i don't log any error and any critical trouble in its usage ;)).

Quoted 1 line

No, thank you Joel, your testing and bravery is an enourmous help :)

Welcome (if you foreseen some additional test don't hesitate), thanks for your trust and support.

Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

#27 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 6 lines · Joel Soete

I now boot on the (previously /mnt/chroot) disk with new glibc (to be accurate cvs 2004 03 15 + your patch + my feupdateenv) with kernel 2.6.4-pa1 first.

I try to rebuild 2.6.4-pa2 as well as 2.6.5-rc1-pa0 (but small pb) and so rebuild 2.6.4-pa1 to add kbd and mouse support: rebuild ok, reboot ok.

Quoted 8 lines

I also relaunch Xfce4 and do some test with its file manager, its configuration interface and xterm, all seems to works fine (well at least, i don't log any error and any critical trouble in its usage ;)).

Quoted 1 line

No, thank you Joel, your testing and bravery is an enourmous help :)

Welcome (if you foreseen some additional test don't hesitate), thanks for your trust and support.

Excellent!

c.

author
#28 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

In reply to Joel Soete

Hello Carlos,

Sorry for delay,...

Quoted 2 lines
Quoted 1 line

make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-float.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-double.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-ifloat.out] Error 1
make[1]: *** [math/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/tst-strtod.out] Error 1
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/bug-strtod.out] Error 1
make[1]: *** [stdlib/tests] Error 2
make[2]: [/Develop/parisc-linux/build/glibc/posix/annexc.out] Error 1 (ignored) make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-tls13.out] Error 1
make[1]: *** [elf/tests] Error 2
make: *** [check] Error 2

So same known errors with cvs dated 2004 03 22 :)

hth,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

#29 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 3 lines · Joel Soete

Hello Carlos,

Sorry for delay,...

No problem!

Quoted 1 line

So same known errors with cvs dated 2004 03 22 :)

Perfect. Thanks!

c.

author
#30 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Hello Carlos,

yes more delay ;) (we have to move desk and so on)

Never the less always same known errors with glibc cvs dated 2004 03 29. But with very last gcc-3.3 installed, the feupdateenv patch is required (gcc now refuse 'assignment of read-only member)

hth,
Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

#31 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 3 lines

Never the less always same known errors with glibc cvs dated 2004 03 29. But with very last gcc-3.3 installed, the feupdateenv patch is required (gcc now refuse 'assignment of read-only member)

Noted, I'm making a patch from your patches for debian. Thanks Joel!

c.

#32 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

In reply to Joel Soete
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

2. Compile glibc with --prefix=/mnt/chroot

Shouldn't this be --prefix=/usr then make install_root=/mnt/chroot install?

Jim

#33 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 1 line · James Morrison

Shouldn't this be --prefix=/usr then make install_root=/mnt/chroot install?

Yes, you can go that way, I think just prefix works aswell. Unless you have seen this fail before?

Glibc doesn't do any path mangling so far as I know.

c.

#34 lore raw

Re: [parisc-linux] RE: New glibc patches 2004-03-08, fixes GLRO changes.

Quoted 9 lines · Carlos O'Donell
Quoted 1 line · James Morrison

Shouldn't this be --prefix=/usr then make install_root=/mnt/chroot install?

Yes, you can go that way, I think just prefix works aswell. Unless you have seen this fail before?

Glibc doesn't do any path mangling so far as I know.

c.

Humm, it may be a Hurd thing where libc needs the correct prefix.

Jim