segfault at startup

It looks like it is blowing the stack.
Ideas?

Morten


gdb loader/wine-pthread 
...]


(gdb) where
#0  0x4010afdc in _int_malloc () from /lib/libc.so.6
#1  0x4010a18c in malloc () from /lib/libc.so.6
#2  0x404adb72 in _xmalloc (size=25) at ../../misc/registry.c:95
#3  0x404aeb24 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
    rgkn=0x405c0020, dke=0x406444d4, f=0x77f04f60, level=0)
    at ../../misc/registry.c:717
#4  0x404aeba7 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
    rgkn=0x405c0020, dke=0x4064449c, f=0x77f04f60, level=0)
    at ../../misc/registry.c:725
#5  0x404aeba7 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
    rgkn=0x405c0020, dke=0x40644464, f=0x77f04f60, level=0)
    at ../../misc/registry.c:725
#6  0x404aeba7 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
    rgkn=0x405c0020, dke=0x4064442c, f=0x77f04f60, level=0)
    at ../../misc/registry.c:725
[...]



(gdb) p *dkh
$1 = {nextkeyoff = 86, nrLS = 188, nrMS = 75, bytesused = 86, keynamelen = 23, 
  values = 1, xx1 = 0, name = "I"}
(gdb) p *dke
$2 = {x1 = 0, hash = 0, x3 = 0, prevlvl = 0, nextsub = 1, next = 0, nrLS = 0, 
  nrMS = 0}

Note dke->next==0.

[Michael Stefaniuc]

Morten Welinder wrote:
> It looks like it is blowing the stack.
> Ideas?
> Let me guess: this is on FC2?

bye
	michael
> gdb loader/wine-pthread 
> ...]
>
> (gdb) where
> #0  0x4010afdc in _int_malloc () from /lib/libc.so.6
> #1  0x4010a18c in malloc () from /lib/libc.so.6
> #2  0x404adb72 in _xmalloc (size=25) at ../../misc/registry.c:95
> #3  0x404aeb24 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
>     rgkn=0x405c0020, dke=0x406444d4, f=0x77f04f60, level=0)
>     at ../../misc/registry.c:717
> #4  0x404aeba7 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
>     rgkn=0x405c0020, dke=0x4064449c, f=0x77f04f60, level=0)
>     at ../../misc/registry.c:725
> #5  0x404aeba7 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
>     rgkn=0x405c0020, dke=0x40644464, f=0x77f04f60, level=0)
>     at ../../misc/registry.c:725
> #6  0x404aeba7 in _w95_dump_dke (key_name=0x4055470c "", creg=0x405c0000, 
>     rgkn=0x405c0020, dke=0x4064442c, f=0x77f04f60, level=0)
>     at ../../misc/registry.c:725
> [...]
>
> (gdb) p *dkh
> $1 = {nextkeyoff = 86, nrLS = 188, nrMS = 75, bytesused = 86, keynamelen = 23, 
>   values = 1, xx1 = 0, name = "I"}
> (gdb) p *dke
> $2 = {x1 = 0, hash = 0, x3 = 0, prevlvl = 0, nextsub = 1, next = 0, nrLS = 0, 
>   nrMS = 0}
>
> Note dke->next==0.

-- 
Michael Stefaniuc               Tel.: [phone]
System Administration           Fax.: [phone]
Red Hat GmbH                    Email: mstefani(a)redhat.com
Hauptstaetterstr. 58            http://www.redhat.de/
D-70178 Stuttgart

From: Michael Stefaniuc <mstefani(a)redhat.com>

   Morten Welinder wrote:
> It looks like it is blowing the stack.
> Ideas?
>    Let me guess: this is on FC2?

No, as mentioned it's SuSE 9.0.

One more thing: about 1 in 5 runs it turns out that wine does start.
Never under strace, though.

/me smells a race condition.

Morten

[Duane Clark]

Morten Welinder wrote:
> ...
> One more thing: about 1 in 5 runs it turns out that wine does start.
> Never under strace, though.
>
> /me smells a race condition.

I'll just add a "me too" to that one. I have not tried putting numbers 
to it, so I don't know whether it is 1 in 5. But I have already gotten 
accustomed to just rerunning a program that doesn't start up. This on RH9.