Re: Problem with ffi/forwardInvocation and exception handling on OpenSolaris

[Saso Kiselkov]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi David,

I just tried libffi-3.0.9, but to no avail - I'm still getting the same
SIGABRT. Interestingly, gdb prints a somewhat different stack trace than
pstack does:

(gdb) bt
#0  0xfffffd7fff29e09a in _lwp_kill () from /lib/64/libc.so.1
#1  0xfffffd7fff2925ed in thr_kill () from /lib/64/libc.so.1
#2  0xfffffd7fff23f5f1 in raise () from /lib/64/libc.so.1
#3  0xfffffd7fff2152a1 in abort () from /lib/64/libc.so.1
#4  0xfffffd7ffdc58994 in objc_exception_throw () from
/usr/lib/64/libobjc.so.2
#5  0xfffffd7ff6473cc2 in -[NSException raise] (self=0x4972d0,
_cmd=0xfffffd7ff665f380) at NSException.m:343
#6  0xfffffd7ff6473814 in +[NSException raise:format:arguments:]
(self=0xfffffd7ff665ef20, _cmd=0xfffffd7ff665f350, name=0xfffffd7ff665e960,
    format=0x411bc0, argList=0xfffffd7fffdff280) at NSException.m:275
#7  0xfffffd7ff6473767 in +[NSException raise:format:]
(self=0xfffffd7ff665ef20, _cmd=0x411de0, name=0xfffffd7ff665e960,
format=0x411bc0)
    at NSException.m:261
#8  0x0000000000401448 in do_throw () at main.m:23
#9  0x00000000004013f9 in -[TestClass forwardInvocation:]
(self=0x42b570, _cmd=0x442d50, inv=0x448080) at main.m:15
#10 0xfffffd7ff65694b8 in GSFFIInvocationCallback (cif=0x41b070,
retp=0xfffffd7fffdff640, args=0xfffffd7fffdff4a0, user=0x41b070)
    at GSFFIInvocation.m:641
#11 0xfffffd7ffe08b54e in ffi_closure_unix64_inner
(closure=0xfffffd7ffec20000, rvalue=0xfffffd7fffdff640,
reg_args=0xfffffd7fffdff590,
    argp=0xfffffd7fffdff660 "") at src/x86/ffi64.c:620
#12 0xfffffd7ffe08b708 in ffi_closure_unix64 () from /usr/lib/64/libffi.so.5
#13 0x00000000004014de in main () at main.m:37

How can I make sure that sysv.S is being used to generate the necessary
DWARF data?

BR,
- --
Saso

On 06/28/2010 11:24 AM, David Chisnall wrote:
> Hi Saso!
> 
> I've not encountered this problem, but the cause of it is this line:
> 
> On 28 Jun 2010, at 10:11, Saso Kiselkov wrote:
> 
>> fffffd7ffdc58994 ???????? ()
> 
> 
> The unwind library is looking for the DWARF EH data to try to find how to 
> unwind the stack, but it can't find it for this frame.  Without it, it is 
> unable to to find out where the stack frame above it is.  
> 
> Upgrading to a newer FFI might help.  There is a sysv.S file in libffi that 
> produces the unwind table for an FFI frame, but there is a chance that it is 
> not being used in the Solaris build.
> 
> David
> 
> -- Send from my Jacquard Loom
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkwob3cACgkQRO8UcfzpOHAHjgCfduiYqG/5rb37aKbNoRBy0OAQ
JbEAnihOIZl5c+puP8DaA5OV3vKsU+8r
=OtGK
-----END PGP SIGNATURE-----