10:16 a.m.
Hi !
While tracking down the problem that makes the current XEmacs 21.5 unusable on
my linux box (it's been so for a long long time now), I went into this:
Current directory is /usr/local/src/xemacs/src/
GNU gdb 5.0.90-cvs (MI_OUT)
Copyright 2001 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show warranty" for details.
This GDB was configured as "i386-linux"...
(gdb) break reinit_vars_of_eval
Breakpoint 1 at 0x80e7f76: file /usr/local/src/XEmacs-21.5/src/eval.c, line
6333.
(gdb) run -vanilla
Starting program: /usr/local/src/xemacs/src/xemacs -vanilla
Breakpoint 1, reinit_vars_of_eval ()
at /usr/local/src/XEmacs-21.5/src/eval.c:6333
(gdb) next
(gdb)
(gdb)
(gdb)
(gdb)
(gdb)
(gdb) set always_gc = 1
(gdb) set max_specpdl_size = 10000
(gdb) cont
Continuing.
Program received signal SIGSEGV, Segmentation fault.
0x0808c08e in mark_string (obj={...})
at /usr/local/src/XEmacs-21.5/src/alloc.c:1874
[ weird: it's DECLARE_FIXED_TYPE_ALLOC (string, Lisp_String); ]
(gdb) pobj obj
$3 = (Lisp_String *) 0x405cae34
$4 = {u = {...}, size_ = 6, data_ = 0x406d7d78 "newbuf", plist = {...}}
But then, I don't really know what to do next. Any advice ?
--
Didier Verna, didier(a)lrde.epita.fr, http://www.lrde.epita.fr/~didier
EPITA / LRDE, 14-16 rue Voltaire Tel.+33 (1) 53 14 59 22
94276 Le Kremlin-Bicêtre, France Fax.+33 (1) 53 14 59 22 didier(a)xemacs.org
2:57 p.m.
>>>> "drv" == Didier Verna
<didier(a)xemacs.org> writes:
drv> Program received signal SIGSEGV, Segmentation fault.
drv> 0x0808c08e in mark_string (obj={...})
drv> at /usr/local/src/XEmacs-21.5/src/alloc.c:1874
drv> [ weird: it's DECLARE_FIXED_TYPE_ALLOC (string, Lisp_String); ]
Our dlmalloc code is alleged broken (by Wolfram.Gloger(a)malloc.de),
specifically the use of mmap, and I've seen breakage for glibc >
2.2.5. Try --pdump and/or --with-system-malloc for some relief, but I
don't think they actually are fixes. I don't know why pdump helps,
but --with-system-malloc basically has the effect of bumping the mmap
threshold to 1MB, and XEmacs rarely mallocs that much at once, we
prefer to do it a cons at a time (but they're big conses ;-).
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
5:26 p.m.
Stephen J. Turnbull wrote:
Our dlmalloc code is alleged broken (by Wolfram.Gloger(a)malloc.de),
specifically the use of mmap, and I've seen breakage for glibc >
2.2.5. Try --pdump and/or --with-system-malloc for some relief, but I
don't think they actually are fixes. I don't know why pdump helps,
but --with-system-malloc basically has the effect of bumping the mmap
threshold to 1MB, and XEmacs rarely mallocs that much at once, we
prefer to do it a cons at a time (but they're big conses ;-).
Actually, I've tried several combinations but with no luck. I smell a
GCPRO bug but I'm a bit at a loss as far as tracking it down. For the record,
here's a valgrind output I get. The interesting part is ==16854==.
--
Didier Verna, didier(a)lrde.epita.fr, http://www.lrde.epita.fr/~didier
EPITA / LRDE, 14-16 rue Voltaire Tel.+33 (1) 53 14 59 22
94276 Le Kremlin-Bicêtre, France Fax.+33 (1) 53 14 59 22 didier(a)xemacs.org
7:48 a.m.
>>>> "drv" == Didier Verna
<didier(a)xemacs.org> writes:
drv> Stephen J. Turnbull wrote:
> Our dlmalloc code is alleged broken (by
> Wolfram.Gloger(a)malloc.de), specifically the use of mmap, and
> I've seen breakage for glibc > 2.2.5. Try --pdump and/or
> --with-system-malloc for some relief, but I don't think they
> actually are fixes. I don't know why pdump helps, but
> --with-system-malloc basically has the effect of bumping the
> mmap threshold to 1MB, and XEmacs rarely mallocs that much at
> once, we prefer to do it a cons at a time (but they're big
> conses ;-).
drv> Actually, I've tried several combinations but with no
drv> luck. I smell a GCPRO bug but I'm a bit at a loss as far as
drv> tracking it down.
OK, (1) I see you've got glibc 2.2.4, so you are not a victim of the
2.3 changes. 21.5 always built fine for me with glibc <= 2.2.5.
(2) Have you tried turning mmap off? This is how Gloger did it in
21.4. Cf the thread anchored at
<a
href="http://list-archive.xemacs.org/xemacs-beta/200211/msg00065.htm...
87d6pl5sw9.fsf(a)tleepslib.sk.tsukuba.ac.jp
</a>
subject "XEmacs 21.4.10 crashes with glibc 2.3.1."
--- src/alloc.c.orig Mon Dec 17 06:49:22 2001
+++ src/alloc.c Thu Nov 7 00:27:31 2002
＠＠ -3882,8 +3890,8 ＠＠
#ifdef DOUG_LEA_MALLOC
mallopt (M_TRIM_THRESHOLD, 128*1024); /* trim threshold */
mallopt (M_MMAP_THRESHOLD, 64*1024); /* mmap threshold */
-#if 0 /* Moved to emacs.c */
- mallopt (M_MMAP_MAX, 64); /* max. number of mmap'ed areas */
+#if 1 /* Moved to emacs.c */
+ mallopt (M_MMAP_MAX, 0); /* max. number of mmap'ed areas */
#endif
#endif
init_string_alloc ();
--- src/emacs.c.orig Thu Oct 31 16:07:36 2002
+++ src/emacs.c Thu Nov 7 00:25:47 2002
＠＠ -2713,7 +2713,8 ＠＠
if (!initialized)
{
#ifdef DOUG_LEA_MALLOC
- mallopt (M_MMAP_MAX, 0);
+ if (mallopt (M_MMAP_MAX, 0) != 1)
+ abort();
#endif
run_temacs_argc = 0;
if (! SETJMP (run_temacs_catch))
＠＠ -2770,7 +2771,8 ＠＠
defined(_NO_MALLOC_WARNING_) || \
(defined(__GLIBC__) && __GLIBC_MINOR__ < 1 && !defined(MULE)) ||
\
defined(DEBUG_DOUG_LEA_MALLOC)
- mallopt (M_MMAP_MAX, 64);
+ if(mallopt (M_MMAP_MAX, 0) != 1)
+ abort();
#endif
#ifdef REL_ALLOC
r_alloc_reinit ();
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
7835
days inactive
7837
days old
3 comments
2 participants
participants (2)
-
Didier Verna -
Stephen J. Turnbull