On Jan 19, 2007, at 8:21 AM, Scott Atchley wrote:
On Jan 18, 2007, at 4:33 PM, Murali Vilayannur wrote:
Hi Scott,
Just a shot into the wild...
Can you do
% mv /lib/tls /lib/tls.old
and then rerun the server with valgrind..
I fixed a malloc, where I missed the trailing NULL. Valgrind now only
shows:
% valgrind --leak-check=full ./sbin/pvfs2-server -d ./etc/fs.conf ./
etc/server.conf-fog33
==15372== Memcheck, a memory error detector for x86-linux.
==15372== Copyright (C) 2002-2005, and GNU GPL'd, by Julian Seward et
al.
==15372== Using valgrind-2.4.0, a program supervision framework for
x86-linux.
==15372== Copyright (C) 2000-2005, and GNU GPL'd, by Julian Seward et
al.
==15372== For more details, rerun with: -v
==15372==
[D 08:25:51.404385] PVFS2 Server version 2.6.1pre1-2007-01-17-144247
starting.
[D 01/19 08:25] bmi_mx: entering BMI_mx_method_addr_lookup.
[D 01/19 08:25] bmi_mx: leaving BMI_mx_method_addr_lookup.
==15372== Warning: noted but unhandled ioctl 0x4D4F (type=0x4D,
nr=0x4F, size=0) with no size/direction hints.
==15372== ioctl was on file "/dev/mx0"
==15372== This could cause spurious value errors to appear.
==15372== See README_MISSING_SYSCALL_OR_IOCTL for guidance on
writing a proper wrapper.
==15372== Warning: noted but unhandled ioctl 0x4D06 (type=0x4D,
nr=0x6, size=0) with no size/direction hints.
==15372== ioctl was on file "/dev/mx0"
==15372== This could cause spurious value errors to appear.
==15372== See README_MISSING_SYSCALL_OR_IOCTL for guidance on
writing a proper wrapper.
==15372== Warning: noted but unhandled ioctl 0x4D13 (type=0x4D,
nr=0x13, size=0) with no size/direction hints.
==15372== ioctl was on file "/dev/mx0"
==15372== This could cause spurious value errors to appear.
==15372== See README_MISSING_SYSCALL_OR_IOCTL for guidance on
writing a proper wrapper.
[D 01/19 08:25] bmi_mx: memory used at end of initialization 17200540.
[D 01/19 08:25] bmi_mx: entering BMI_mx_set_info.
[D 01/19 08:25] bmi_mx: leaving BMI_mx_set_info.
[D 01/19 08:25] bmi_mx: entering BMI_mx_set_info.
[D 01/19 08:25] bmi_mx: leaving BMI_mx_set_info.
[E 01/19 08:25] TROVE:DBPF:Berkeley DB: read: 0x1f9fc7f0, 4096: Bad
address
[E 01/19 08:25] TROVE:DBPF:Berkeley DB: //scratch/atchley/4e901077/
collection_attributes.db: Bad address
[E 01/19 08:25] src/server/pvfs2-server.c line 1077: Error
initializing filesystem pvfs2-fs
[E 01/19 08:25] [bt] ./sbin/pvfs2-server [0x8054de6]
[E 01/19 08:25] [bt] ./sbin/pvfs2-server [0x8054691]
[E 01/19 08:25] [bt] ./sbin/pvfs2-server(main+0x4b0) [0x8054078]
[E 01/19 08:25] [bt] /lib/tls/libc.so.6(__libc_start_main
+0xf4) [0x1ba42974]
[E 01/19 08:25] [bt] ./sbin/pvfs2-server(aio_cancel64+0x61)
[0x80539a1]
[E 01/19 08:25] Error: Could not initialize server interfaces; aborting.
[E 01/19 08:25] Error: Could not initialize server; aborting.
[D 01/19 08:25] bmi_mx: entering BMI_mx_finalize.
Dumping MX statistics for endpoint 3 on board 0:
isend tiny ( 0- 32): 0
isend small ( 33- 128): 0
isend medium ( 129- 32768): 0
isend large ( > 32768): 0
isend total : 0
issend total : 0
irecv : 0
expected : 0
unexpected : 0 (including 0 claimed by the handler)
early fragments : 0
overlap: 42376 test before completion out of 42376
overlap: 0 non-blocking wait out of 0
overlap: 0 completion overlapped out of 0
resent_slow/resent: 0/0
delayed_acks/total_acks: 0/0
rcache: hits 0 (0 kbytes) miss 0 (0 kbytes)
mcp handles: got 0 missed 0
[D 01/19 08:25] bmi_mx: 0 misses.
[D 01/19 08:25] bmi_mx: memory leaked at shutdown 0.
[D 01/19 08:25] bmi_mx: leaving BMI_mx_finalize.
==15372==
==15372== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 23
from 1)
==15372== malloc/free: in use at exit: 1903 bytes in 33 blocks.
==15372== malloc/free: 802 allocs, 769 frees, 84365483 bytes allocated.
==15372== For counts of detected errors, rerun with: -v
==15372== searching for pointers to 33 not-freed blocks.
==15372== checked 33716004 bytes.
==15372==
==15372== 57 (36 direct, 21 indirect) bytes in 1 blocks are
definitely lost in loss record 15 of 22
==15372== at 0x1B90459D: malloc (vg_replace_malloc.c:130)
==15372== by 0x80B20A9: alloc_method_addr (bmi-method-support.c:99)
==15372== by 0x80AFD8F: bmx_alloc_method_addr (mx.c:1915)
==15372== by 0x80B1ABB: BMI_mx_method_addr_lookup (mx.c:2569)
==15372== by 0x805A267: activate_method (bmi.c:1903)
==15372== by 0x80573F9: BMI_initialize (bmi.c:173)
==15372== by 0x8054A66: server_initialize_subsystems (pvfs2-
server.c:961)
==15372== by 0x8054690: server_initialize (pvfs2-server.c:801)
==15372== by 0x8054077: main (pvfs2-server.c:551)
==15372==
==15372==
==15372== 272 bytes in 4 blocks are possibly lost in loss record 20
of 22
==15372== at 0x1B904F75: calloc (vg_replace_malloc.c:175)
==15372== by 0x1B8F2678: (within /lib/ld-2.3.2.so)
==15372== by 0x1B8F294B: _dl_allocate_tls (in /lib/ld-2.3.2.so)
==15372== by 0x1B9EE24A: allocate_stack (in /lib/tls/
libpthread-0.60.so)
==15372== by 0x1B9EDC54: pthread_create@@GLIBC_2.1 (in /lib/tls/
libpthread-0.60.so)
==15372== by 0x1BA087B5: mx_open_endpoint (mx_open_endpoint.c:453)
==15372== by 0x80ACD0F: bmx_open_endpoint (mx.c:788)
==15372== by 0x80ACF87: BMI_mx_initialize (mx.c:895)
==15372== by 0x805A2C6: activate_method (bmi.c:1914)
==15372== by 0x80573F9: BMI_initialize (bmi.c:173)
==15372== by 0x8054A66: server_initialize_subsystems (pvfs2-
server.c:961)
==15372== by 0x8054690: server_initialize (pvfs2-server.c:801)
==15372==
==15372== LEAK SUMMARY:
==15372== definitely lost: 36 bytes in 1 blocks.
==15372== indirectly lost: 21 bytes in 2 blocks.
==15372== possibly lost: 272 bytes in 4 blocks.
==15372== still reachable: 1574 bytes in 26 blocks.
==15372== suppressed: 0 bytes in 0 blocks.
==15372== Reachable blocks (those to which a pointer was found) are
not shown.
==15372== To see them, rerun with: --show-reachable=yes
_______________________________________________
Pvfs2-developers mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-developers
