bug#18392: Bug#759400: emacs24-nox: segfault when saving emacs lisp file

bug#18392: Bug#759400: emacs24-nox: segfault when saving emacs lisp file

[Rob Browning]

[If possible, please preserve the -forwarded address in any replies.]

It appears that Emacs 24.3 is crashing on Debian mipsel, though it's not
clear whether it's an Emacs problem, a toolchain issue, or something
else.

For now, I've been able to reproduce the crash by following the
instructions below on one of our porter boxes.

If helpful, I'd be happy to run further tests, or gather any additional
information.

On Ma, 26 aug 14, 17:16:32, David Bremner wrote:

> To reproduce, 
> 
> 0. log into a mipsel host
> 
> 1. run "emacs -Q foo.el", where foo.el is the attached file 
>    (from the notmuch source)
> 
> 2. Press space or modify the file in some other way
> 
> 3. save the file using Control-X Control-S
> 
> For me this prints 
> 
> "Fatal error 6: aborted"
> 
> And then it seems like emacs hangs, and never returns to the shell.
> 
> I originally noticed a possibly related problem with the same file
> causing notmuch to FTBFS on notmuch. Here byte compilation leads to a
> more detailed output.
> 
>  % emacs --quick -batch  --directory .  -f batch-byte-compile notmuch-show.el
> emacs: malloc.c:2372: sysmalloc: Assertion `(old_top == (((mbinptr) (((char *) &((av)->bins[((1) -
> 1) * 2])) - __builtin_offsetof (struct malloc_chunk, fd)))) && old_size == 0) || ((unsigned long)
> (old_size) >= (unsigned long)((((__builtin_offsetof (struct malloc_chunk, fd_nextsize))+((2
> *(sizeof(size_t))) - 1)) & ~((2 *(sizeof(size_t))) - 1))) && ((old_top)->size & 0x1) && ((unsigned
> long) old_end & pagemask) == 0)' failed.

The file mentioned (notmuch-show.el) is avaiable as an attachment to the
full bug report here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=759400

Also, when I tested, I saw:

  emacs24-nox: malloc.c:2372: sysmalloc: Assertion `(old_top == (((mbinptr) (((char *) &((av)->bins[((1) - 1) * 2])) - __builtin_offsetof (struct malloc_chunk, fd)))) && old_size == 0) || ((unsigned long) (old_size) >= (unsigned long)((((__builtin_offsetof (struct malloc_chunk, fd_nextsize))+((2 *(sizeof(size_t))) - 1)) & ~((2 *(sizeof(size_t))) - 1))) && ((old_top)->size & 0x1) && ((unsigned long) old_end & pagemask) == 0)' failed.

  Program received signal SIGABRT, Aborted.
  0x77943ee8 in raise () from /lib/mipsel-linux-gnu/libc.so.6
  (gdb) Quit
  (gdb) where
  #0  0x77943ee8 in raise () from /lib/mipsel-linux-gnu/libc.so.6
  #1  0x77945948 in abort () from /lib/mipsel-linux-gnu/libc.so.6
  #2  0x7798fad0 in ?? () from /lib/mipsel-linux-gnu/libc.so.6

  Lisp Backtrace:
  "find-coding-systems-region-internal" (0x7fff23a0)
  "find-coding-systems-region" (0x7fff2510)
  "select-safe-coding-system" (0x7fff2694)
  "write-region" (0x7fff2ab8)
  "basic-save-buffer-2" (0x7fff2c30)
  "basic-save-buffer-1" (0x7fff2d98)
  "basic-save-buffer" (0x7fff2f10)
  "save-buffer" (0x7fff30ac)
  "call-interactively" (0x7fff31d4)

Thanks
-- 
Rob Browning
rlb @defaultvalue.org and @debian.org
GPG as of 2011-07-10 E6A9 DA3C C9FD 1FF8 C676 D2C4 C0F0 39E9 ED1B 597A
GPG as of 2002-11-03 14DD 432F AE39 534D B592 F9A0 25C8 D377 8C7E 73A4

bug#18392: Bug#759400: emacs24-nox: segfault when saving emacs lisp file

[Stefan Monnier]

> It appears that Emacs 24.3 is crashing on Debian mipsel, though it's not
> clear whether it's an Emacs problem, a toolchain issue, or something
> else.

Could you check whether the same problem shows up for the 24.4 pretest
(i.e. 24.3.93)?  Also, worthwhile would be to test to see if building
with different optimization settings changes the result.

I do have a mispel machine here, but its 64MB of RAM make it rather
painful to build Emacs (and I don't have a Debian install on it any
more, only OpenWRT).


        Stefan

bug#18392: Bug#759400: emacs24-nox: segfault when saving emacs lisp file

[David Bremner]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

>
> I do have a mispel machine here, but its 64MB of RAM make it rather
> painful to build Emacs (and I don't have a Debian install on it any
> more, only OpenWRT).


Hi Stefan;

If it would help at some point, I'd be happy to sponsor you for a guest
account on the debian mipsel porter box.  The procedure isn't too bad,
as far as these things go:

       https://dsa.debian.org/doc/guest-account/

d

bug#18392: Bug#759400: emacs24-nox: segfault when saving emacs lisp file

[Rob Browning]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

> Could you check whether the same problem shows up for the 24.4 pretest
> (i.e. 24.3.93)?  Also, worthwhile would be to test to see if building
> with different optimization settings changes the result.

Of course, I should have done this initially, but I just tried a clean
emacs-24.4.tar.xz build, and assuming I didn't make any mistakes, that
worked fine, so I need to investigate further on the Debian side.

Thanks for the help.
-- 
Rob Browning
rlb @defaultvalue.org and @debian.org
GPG as of 2011-07-10 E6A9 DA3C C9FD 1FF8 C676 D2C4 C0F0 39E9 ED1B 597A
GPG as of 2002-11-03 14DD 432F AE39 534D B592 F9A0 25C8 D377 8C7E 73A4

bug#18392: Bug#759400: emacs24-nox: segfault when saving emacs lisp file

[Lars Ingebrigtsen]

Rob Browning <rlb@defaultvalue.org> writes:

> Stefan Monnier <monnier@IRO.UMontreal.CA> writes:
>
>> Could you check whether the same problem shows up for the 24.4 pretest
>> (i.e. 24.3.93)?  Also, worthwhile would be to test to see if building
>> with different optimization settings changes the result.
>
> Of course, I should have done this initially, but I just tried a clean
> emacs-24.4.tar.xz build, and assuming I didn't make any mistakes, that
> worked fine, so I need to investigate further on the Debian side.
>
> Thanks for the help.

Ok; closing this report in the Emacs bug tracker.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no