[PVFS2-developers] Re: noncontig (ROMIO) and segfault

[Robert Latham]

On Wed, Mar 17, 2004 at 03:21:07PM -0500, Pete Wyckoff wrote:
> Can you test now and see if you get similar error messages on your side?
> The server no longer core dumps here.

my hindexed testcase still segfaults.  the backtrace looks the
same:  

(gdb) where
#0  0x4207aecc in chunk_free () from /lib/i686/libc.so.6
#1  0x4207ac24 in free () from /lib/i686/libc.so.6
#2  0x08057ae9 in lebf_encode_req (req=0x80a93a4, target_msg=0x80a93fc) at ../pvfs2-source/src/proto/PINT-le-bytefield.c:284
#3  0x080571ef in PINT_encode (input_buffer=0x80a93a4, input_type=PINT_ENCODE_REQ, target_msg=0x80a93fc, target_addr=134925328, enc_type=ENCODING_LE_BFIELD) at ../pvfs2-source/src/proto/PINT-reqproto-encode.c:97
#4  0x080546df in io_datafile_post_msgpairs (sm_p=0x80aea60, js_p=0x808b700) at ../pvfs2-source/src/client/sysint/sys-io.sm:507
#5  0x0806dca8 in PINT_client_state_machine_test () at ../pvfs2-source/src/common/misc/state-machine-fns.h:150
#6  0x08053e86 in PVFS_sys_io (pinode_ref={handle = 1610612737, fs_id = 9}, file_req=0x80a6d48, file_req_offset=0, buffer=0x0, mem_req=0x80a5948, credentials={uid = 2334, gid = 300}, resp_p=0xbfffeed8, 
    io_type=PVFS_IO_WRITE) at ../pvfs2-source/src/client/sysint/sys-io.sm:223
#7  0x08049e97 in main (argc=2, argv=0xbfffef84) at hindex-test.c:189
#8  0x42017499 in __libc_start_main () from /lib/i686/libc.so.6

valgrind says CASE(PVFS_SERV_IO, io)
(src/proto/PINT-le-bytefield.c:284 in my working copy) is overrrunning
someone's memory.

Anything in particular you'd like to see from my end that can help
track this down?  

==rob

-- 
Rob Latham
Mathematics and Computer Science Division    A215 0178 EA2D B059 8CDF
Argonne National Labs, IL USA                B29D F333 664A 4280 315B