This looks very odd to me. None of this code looks like it belongs to
LAM/MPI.
Are you sure that this corefile came from lamboot?
On Dec 13, 2004, at 1:31 PM, martin wood wrote:
> I have rebuilt the LAM with -g, as requested, however, the dbg info
> does not look too helpful to me,
>
> dbx version 7.3.4 (89708_Mar27 patchSG0004885) Mar 27 2003 14:33:10
> Unable to correlate regions with rld object list:( debugger internal
> status code 82). Allowing some minimal use of the core file, but
> debugger will work poorly (if at all)
> Executable /usr/local/lam-7.0.4/bin/lamboot
> (dbx) run -v lamhosts.origin
> Process 9930909 (lamboot) started
> Process 9930909 (lamboot) stopped on signal SIGSEGV: Segmentation
> violation (default) at [obj_init:138 ,0xfb62a20]
> Source (of /j10/mtibuild/v741m/workarea/v7.4.1m/rld/obj.c) not
> available for Process 9930909
> (dbx) where
>> 0 obj_init(0xfbe1c28, 0xffffffffffffffff, 0x8040, 0xfbe1d9e, 0x0,
>> 0x6f, 0x2e, 0x31)
>> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/obj.c":138, 0xfb62a20]
> 1 map_object_into_mem_and_init_object_info(0xfbe1c28, 0xfbe1418,
> 0x8040, 0xfbe1418, 0x0, 0x6f, 0x2e, 0x31)
> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/rld_load_object.c":1251,
> 0xfb6b264]
> 2 build_obj_list_by_liblist_index(0xfbe1418, 0x0, 0xffffff5c60,
> 0xfbe1d9e, 0x0, 0x6f, 0x2e, 0x31)
> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/rld_startup.c":2212,
> 0xfb731f0]
> 3 build_obj_list(0xfbe1418, 0xffffffffffffffff, 0x8040, 0xfbe1d9e,
> 0x0, 0x6f, 0x2e, 0x31)
> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/rld_startup.c":2060,
> 0xfb72fd8]
> 4 build_entire_object_list(0xfbe1418, 0x2, 0xffffff5c60, 0xfbe1d9e,
> 0x0, 0x6f, 0x2e, 0x31)
> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/rld_startup.c":521,
> 0xfb714dc]
> 5 read_aux_vec_and_build_objlist(0x2, 0xfbe1418, 0x8040, 0xfbe1d9e,
> 0x0, 0x6f, 0x2e, 0x31)
> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/rld_startup.c":1505,
> 0xfb72680]
> 6 sgi_main(0x0, 0xffffff5e18, 0xffffff5e18, 0xfbe1d9e, 0x0, 0x6f,
> 0x2e, 0x31)
> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/rld_startup.c":1645,
> 0xfb728a0]
> 7 newmain()
> ["/j10/mtibuild/v741m/workarea/v7.4.1m/rld/rld_newmain.s":68,
> 0xfb601b4]
> (dbx) quit
>
>
>
>
>
>
>
>
>
>
>
> Original Message and Reply :
>
> On Nov 15, 2004, at 8:46 AM, martin wood wrote:
>
>
> I can achieve, what appears to be a successful build and successful
> install on an SGI Origin, but I cannot run.
>
> I am using MIPSpro Compilers: Version 7.41 on OS level 6.5 6.5.21m.
> I have tried building a 32 bit version and a 64 bit version, but all
> behave the same.
>
> I am using the SGI , f77 ; cc and CC
>
> After the build I can get “recon” to work fine.
>
> When I get to lamboot I just get a Segmentation fault (core dumped).
>
> Has anyone built a successful 64 bit version 7.0.4 on an SGI
>
> Sorry about the slow reply - You caught us right after SC'04 and we're
> still trying to dig out :).
>
> We no longer have access to SGI machines, so I can't test anything
> right now. But we did have access to machines at the time of 7.0.4's
> release and would have run our regression tests on the platform at the
> time of release. Can you build LAM with CFLAGS set to -g and get a
> stack trace of where the segmentation fault is occurring?
>
> Thanks!
>
> Brian
>
> --
> Brian Barrett
> LAM/MPI developer and all around nice guy
> Have an LAM/MPI day: http://www.lam-mpi.org/
>
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
