Re: SIGILL in __libc_csu_fini
On Fri, Apr 02, 2004 at 12:33:26AM +0100, Ian Lynagh wrote: > I'm having some trouble with ghc6 on ia64 (this is not using the > packages, which also give SIGILL, but trying to crossport from x86 > again. What versions of things are you using? I just installed both ghc6 and ghc-cvs packages on an I2, and apart from creating a 8MB executable file from my simple 10 line example both work fine. libc6.1 2.3.2.ds1-11 GNU C Library: Shared libraries and Timezone data ghc-cvs 20031220-3 GHC - the Glasgow Haskell Compilation system ghc66.2-3GHC - the Glasgow Haskell Compilation system gcc 3.3.3-2 The GNU C compiler (might want to leave glibc list off, they have enough to worry about until we can prove it's a glibc bug) -i [EMAIL PROTECTED] http://www.gelato.unsw.edu.au signature.asc Description: Digital signature
SIGILL in __libc_csu_fini
[Note crosspost. I'm not subscribed to either list.] Hi all, I'm having some trouble with ghc6 on ia64 (this is not using the packages, which also give SIGILL, but trying to crossport from x86 again. gdb gives more useful information than for the packages). The gdb output, from merulo's unstable dchroot a day or two ago, is below. I'm afraid I have no idea what to do from here. Any help much appreciated! $ gdb /home/igloo/ghc/x1/ghc/compiler/stage1/ghc-6.2.1 [...] (gdb) b main Breakpoint 1 at 0x426f2a42 (gdb) r Starting program: /home/igloo/ghc/x1/ghc/compiler/stage1/ghc-6.2.1 Program received signal SIGILL, Illegal instruction. 0x42737901 in __libc_csu_fini () (gdb) bt #0 0x42737901 in __libc_csu_fini () #1 0x42737790 in __libc_csu_init () #2 0x202a1220 in __libc_start_main () from /lib/libc.so.6.1 #3 0x40004260 in _start () (gdb) disassemble Dump of assembler code for function __libc_csu_fini: 0x42737830 <__libc_csu_fini+0>: [MII] alloc r36=ar.pfs,6,6,0 0x42737831 <__libc_csu_fini+1>: addl r14=58216,r1 0x42737832 <__libc_csu_fini+2>: addl r16=20416,r1 0x42737840 <__libc_csu_fini+16>:[MII] mov r37=r1 0x42737841 <__libc_csu_fini+17>:mov r35=b0;; 0x42737842 <__libc_csu_fini+18>:nop.i 0x0 0x42737850 <__libc_csu_fini+32>:[MMI] ld8 r15=[r14] 0x42737851 <__libc_csu_fini+33>:ld8 r14=[r16] 0x42737852 <__libc_csu_fini+34>:nop.i 0x0;; 0x42737860 <__libc_csu_fini+48>:[MMI] sub r14=r14,r15;; 0x42737861 <__libc_csu_fini+49>:nop.m 0x0 0x42737862 <__libc_csu_fini+50>:shr r33=r14,3;; 0x42737870 <__libc_csu_fini+64>:[MII] mov r14=r33 0x42737871 <__libc_csu_fini+65>:adds r33=-1,r33;; 0x42737872 <__libc_csu_fini+66>:cmp.eq p7,p6=0,r14 0x42737880 <__libc_csu_fini+80>:[MIB] shladd r34=r33,3,r15 0x42737881 <__libc_csu_fini+81>:nop.i 0x0 0x42737882 <__libc_csu_fini+82>: (p07) br.cond.dpnt.few 0x427378d0 <__libc_csu_fini+160>;; 0x42737890 <__libc_csu_fini+96>:[MII] ld8 r14=[r34],-8 0x42737891 <__libc_csu_fini+97>:mov r32=r33 ---Type to continue, or q to quit--- 0x42737892 <__libc_csu_fini+98>:adds r33=-1,r33;; 0x427378a0 <__libc_csu_fini+112>: [MMI] ld8 r15=[r14],8;; 0x427378a1 <__libc_csu_fini+113>: ld8 r1=[r14] 0x427378a2 <__libc_csu_fini+114>: mov b6=r15 0x427378b0 <__libc_csu_fini+128>: [MIB] nop.m 0x0 0x427378b1 <__libc_csu_fini+129>: nop.i 0x0 0x427378b2 <__libc_csu_fini+130>: br.call.sptk.many b0=b6;; 0x427378c0 <__libc_csu_fini+144>: [MIB] cmp.eq p7,p6=0,r32 0x427378c1 <__libc_csu_fini+145>: mov r1=r37 0x427378c2 <__libc_csu_fini+146>: (p06) br.cond.dptk.few 0x42737890 <__libc_csu_fini+96> 0x427378d0 <__libc_csu_fini+160>: [MIB] nop.m 0x0 0x427378d1 <__libc_csu_fini+161>: nop.i 0x0 0x427378d2 <__libc_csu_fini+162>: br.call.sptk.many b0=0x427379a0 <_fini>;; 0x427378e0 <__libc_csu_fini+176>: [MII] mov r1=r37 0x427378e1 <__libc_csu_fini+177>: mov.i ar.pfs=r36 0x427378e2 <__libc_csu_fini+178>: mov b0=r35 0x427378f0 <__libc_csu_fini+192>: [MIB] nop.m 0x0 0x427378f1 <__libc_csu_fini+193>: nop.i 0x0 0x427378f2 <__libc_csu_fini+194>: br.ret.sptk.many b0;; ---Type to continue, or q to quit--- 0x42737900 <__libc_csu_fini+208>: [MLX] nop.m 0x0 0x42737901 <__libc_csu_fini+209>: brl.few 0x400027e0 <_init>;; End of assembler dump. (gdb) disassemble 0x400027e0 Dump of assembler code for function _init: 0x400027e0 <_init+0>: [MII] alloc r34=ar.pfs,3,3,0 0x400027e1 <_init+1>: mov r32=r12 0x400027e2 <_init+2>: mov r33=b0 0x400027f0 <_init+16>: [MFB] adds r12=-16,r12 0x400027f1 <_init+17>: nop.f 0x0 0x400027f2 <_init+18>: nop.b 0x0;; 0x40002800 <_init+32>: [MII] mov r12=r32 0x40002801 <_init+33>: mov.i ar.pfs=r34 0x40002802 <_init+34>: mov b0=r33 0x40002810 <_init+48>: [MFB] nop.m 0x0 0x40002811 <_init+49>: nop.f 0x0 0x40002812 <_init+50>:
Processed: Re: Bug#231538: *sigh*
Processing commands for [EMAIL PROTECTED]: > clone 231538 -1 Bug#231538: libc6: illegal instruction on a 386 Bug 231538 cloned as bug 241497. > reassign -1 ftp.debian.org Bug#241497: libc6: illegal instruction on a 386 Bug reassigned from package `libc6' to `ftp.debian.org'. > retitle -1 sarge needs upgrade-i386 directory which gives us upgrade-path for real > i386 architecture Bug#241497: libc6: illegal instruction on a 386 Changed Bug title. > thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#231538: *sigh*
GOTO Masanori wrote: > clone 231538 -1 > reassign -1 ftp.debian.org > retitle -1 sarge needs upgrade-i386 directory which gives us upgrade-path for real > i386 architecture > thanks > > I'm sorry posting this mail to so many people... > > Summary: > > - Original bug #231538 will be fixed by debian-glibc maintainer team >at the next glibc 2.3.2.ds1-12 which adds the code to prevent >installing sarge's glibc on real 80386 arch with kernel < 2.4.24. > > - Debian decided to make dists/sarge/main/upgrade-i386 for providing >upgrade path for real 80386 architecture. > > - ftp.debian.org maintainer needs to make upgrade-i386 directory with >appropriate documents/binaries/libraries. > > - Release document/management team need to write additional notes >about upgrade-i386 issue into sarge release announce documents for >i386 architecture. > > - Kernel maintainer may need to put kernel 2.4.24 or newer into >upgrade-i386 directory. The newer kernel has 486 emulation code >which can install sarge's libraries even on real 80386 >architecture. FYI: I've added this to http://people.debian.org/~joey/pr/3.1/i386.html Regards, Joey -- It's practically impossible to look at a penguin and feel angry. Please always Cc to me when replying to me on the lists. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#231538: *sigh*
clone 231538 -1
reassign -1 ftp.debian.org
retitle -1 sarge needs upgrade-i386 directory which gives us upgrade-path for real
i386 architecture
thanks
I'm sorry posting this mail to so many people...
Summary:
- Original bug #231538 will be fixed by debian-glibc maintainer team
at the next glibc 2.3.2.ds1-12 which adds the code to prevent
installing sarge's glibc on real 80386 arch with kernel < 2.4.24.
- Debian decided to make dists/sarge/main/upgrade-i386 for providing
upgrade path for real 80386 architecture.
- ftp.debian.org maintainer needs to make upgrade-i386 directory with
appropriate documents/binaries/libraries.
- Release document/management team need to write additional notes
about upgrade-i386 issue into sarge release announce documents for
i386 architecture.
- Kernel maintainer may need to put kernel 2.4.24 or newer into
upgrade-i386 directory. The newer kernel has 486 emulation code
which can install sarge's libraries even on real 80386
architecture.
At Thu, 01 Apr 2004 03:38:24 -0500,
Nathanael Nerode wrote:
> | So... I don't add "use woody-proposed-updates" message into
> | libc6.preinst, I keep them as the current one. I think Bug#231538 can
> | be closed when 2.3.2.ds1-12 is released. But the current thread is
> | useful for keeping discussing to make dists/sarge/main/upgrade-i386
> | for the near future sarge release. I would clone this bug and
> | reassign it to ftp.debian.org or ftp-master. Nathanael, OK?
> Yes; sounds great. :-) I have no idea whose job it is to create the
> upgrade-i386 directory (ftp.debian.org, I guess), but please clone and
> assign to the appropriate person.
OK, I cloned it and I assigned it to ftp.debian.org. Martin seems
to prepare making upgrade-i386 directory, though.
> Herbert's job, then, is to upload the appropriate packages to that
> directory; the RM's job is to make sure it's mentioned prominently in
> the release notes. And someone has to create the README, but that
> shouldn't be much of a job. ("You must upgrade to these packages, and
> reboot with the new kernel, before installing sarge".)
Yup. Before releasing sarge, we need to work for upgrade-i386 directory.
Thanks to all people who comments to this bug!
Regards,
-- gotom
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Moved on
I left UZL and moved on to a new job. From now on I can be reached at [EMAIL PROTECTED] Your mail regarding "[VIRUS] Re: Your document" will be forwarded to my new account. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Custom compiled glibc
Hi there, I posted about this before but messed the post up. How would I go about compiling a glibc optimised for Pentium-M and SSE2? Thanks, -- Anders Karlsson <[EMAIL PROTECTED]> Trudheim Technology Limited signature.asc Description: This is a digitally signed message part
Bug#231538: *sigh*
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
GOTO Masanori wrote:
| At Tue, 30 Mar 2004 10:46:25 +0200,
| Martin Schulze wrote:
|
|>I have another solution to propose: Provide an upgrade-for-real-i386
|>directory, including a README and kernel packages to install before
|>upgrading to sarge. We've done a similar thing from bo->potato or
|>something.
|>
|>This will require coordination with the release manager and the
|>ftpmasters to provide dists/sarge/main/upgrade-i386 which should
|>contain the kernel packages and a readme file.
| So... I don't add "use woody-proposed-updates" message into
| libc6.preinst, I keep them as the current one. I think Bug#231538 can
| be closed when 2.3.2.ds1-12 is released. But the current thread is
| useful for keeping discussing to make dists/sarge/main/upgrade-i386
| for the near future sarge release. I would clone this bug and
| reassign it to ftp.debian.org or ftp-master. Nathanael, OK?
Yes; sounds great. :-) I have no idea whose job it is to create the
upgrade-i386 directory (ftp.debian.org, I guess), but please clone and
assign to the appropriate person.
Herbert's job, then, is to upload the appropriate packages to that
directory; the RM's job is to make sure it's mentioned prominently in
the release notes. And someone has to create the README, but that
shouldn't be much of a job. ("You must upgrade to these packages, and
reboot with the new kernel, before installing sarge".)
Remember to tell Karolina, who opened the bug in the first place, when
it's all done. :-)
| Regards,
| -- gotom
Glad this is getting dealt with!
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFAa9UARGZ0aC4lkIIRAlEWAJ95b9TP31PT93mC2HUTX+h5ZezHcQCdETNk
v4mTmUkJWMdvW4O4PqFK+ek=
=V8eO
-END PGP SIGNATURE-
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#235759: Comentar on which replacement for German quotes
On Wed, Mar 31, 2004 at 01:00:20AM +0200, Adrian Bunk wrote: > On Mon, Mar 29, 2004 at 11:02:11PM +0200, Denis Barbier wrote: > > On Mon, Mar 29, 2004 at 02:47:42AM +0200, Adrian Bunk wrote: > > [...] > > > I am a bit surprised that this discussion and all patches sent only > > > cover de_DE, although there are altogether three common de_ locales > > > plus two others the locales package supports. > > > If you fix this issue, it should be properly fixed for all de_ locales. > > > > All de_* locales include (either directly or indirectly) de_DE in their > > LC_CTYPE section, so this change propagates to all de_* locales. > > Thanks for the explanation. > > This means only cases like the different quote in the Swiss locale need > to be handled. According to http://people.debian.org/~barbier/intl/l10n/po/ there is currently no package in Debian with de_CH messages. As French quotes are quite common in German catalogs (e.g. they are used in all GNOME packages), this means that de_CH users are currently seeing lots of 'reversed' quotes. And of course ,," transliteration ;) I read your message as: "the proposed patch is an improvement for all German speaking people, but may be even improved for de_CH folks", but note that glibc maintainers seem reluctant to apply the requested changes and may have a different reading of your message. Denis -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
