To see if I could find the reason for the segfault, I tested cpqarrayd on one of our 385-machines with valgrind. As you can see, there are problems with it. The malloc() in cciss_interrogate_controller() seem to be one byte short. Could this be the source of the segfault? Also, note that -d do not disable the forking into the background as claimed by the usage help text. This is annoying and should be fixed.
hp385# valgrind cpqarrayd -d ==22268== Memcheck, a memory error detector. ==22268== Copyright (C) 2002-2005, and GNU GPL'd, by Julian Seward et al. ==22268== Using LibVEX rev 1575, a library for dynamic binary translation. ==22268== Copyright (C) 2004-2005, and GNU GPL'd, by OpenWorks LLP. ==22268== Using valgrind-3.1.1, a dynamic binary instrumentation framework. ==22268== Copyright (C) 2000-2005, and GNU GPL'd, by Julian Seward et al. ==22268== For more details, rerun with: -v ==22268== Checking for controllers.. DEBUG: Device /dev/ida/c0d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/ida/c1d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/ida/c2d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/ida/c3d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/ida/c4d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/ida/c5d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/ida/c6d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/ida/c7d0 could not be opened DEBUG: reason: No such file or directory ==22268== Syscall param ioctl(generic) points to uninitialised byte(s) ==22268== at 0x304FDBFC19: ioctl (in /lib64/tls/libc-2.3.4.so) ==22268== by 0x402EDB: cciss_get_logical_luns (cciss_functions.c:106) ==22268== by 0x40219E: cciss_interrogate_controller (discover.c:282) ==22268== by 0x401950: discover_controllers (discover.c:98) ==22268== by 0x401332: main (cpqarrayd.c:164) ==22268== Address 0x7FF0005F8 is on thread 1's stack ==22268== ==22268== Invalid write of size 1 ==22268== at 0x490656F: strcpy (mac_replace_strmem.c:269) ==22268== by 0x402287: cciss_interrogate_controller (discover.c:300) ==22268== by 0x401950: discover_controllers (discover.c:98) ==22268== by 0x401332: main (cpqarrayd.c:164) ==22268== Address 0x4A1A137 is 0 bytes after a block of size 15 alloc'd ==22268== at 0x4904A06: malloc (vg_replace_malloc.c:149) ==22268== by 0x402260: cciss_interrogate_controller (discover.c:299) ==22268== by 0x401950: discover_controllers (discover.c:98) ==22268== by 0x401332: main (cpqarrayd.c:164) ==22268== ==22268== Syscall param ioctl(generic) points to uninitialised byte(s) ==22268== at 0x304FDBFC19: ioctl (in /lib64/tls/libc-2.3.4.so) ==22268== by 0x40307D: cciss_get_event (cciss_functions.c:165) ==22268== by 0x4022D6: cciss_interrogate_controller (discover.c:308) ==22268== by 0x401950: discover_controllers (discover.c:98) ==22268== by 0x401332: main (cpqarrayd.c:164) ==22268== Address 0x7FF0005F9 is on thread 1's stack DEBUG: Discarding old event 5/0/0 ==22268== ==22268== Syscall param ioctl(generic) points to uninitialised byte(s) ==22268== at 0x304FDBFC19: ioctl (in /lib64/tls/libc-2.3.4.so) ==22268== by 0x40307D: cciss_get_event (cciss_functions.c:165) ==22268== by 0x40230C: cciss_interrogate_controller (discover.c:312) ==22268== by 0x401950: discover_controllers (discover.c:98) ==22268== by 0x401332: main (cpqarrayd.c:164) ==22268== Address 0x7FF0005F9 is on thread 1's stack DEBUG: Discarding old event 5/2/0 DEBUG: Discarding old event 5/0/0 DEBUG: Discarding old event 5/2/0 DEBUG: /dev/cciss/c0d0 is a existing controller DEBUG: Device /dev/cciss/c1d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/cciss/c2d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/cciss/c3d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/cciss/c4d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/cciss/c5d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/cciss/c6d0 could not be opened DEBUG: reason: No such file or directory DEBUG: Device /dev/cciss/c7d0 could not be opened DEBUG: reason: No such file or directory Done Monitoring list ==22268== ==22268== Invalid read of size 1 ==22268== at 0x4906503: strlen (mac_replace_strmem.c:243) ==22268== by 0x304FD42DC8: vfprintf (in /lib64/tls/libc-2.3.4.so) ==22268== by 0x304FD48617: printf (in /lib64/tls/libc-2.3.4.so) ==22268== by 0x401771: main (cpqarrayd.c:176) ==22268== Address 0x4A1A137 is 0 bytes after a block of size 15 alloc'd ==22268== at 0x4904A06: malloc (vg_replace_malloc.c:149) ==22268== by 0x402260: cciss_interrogate_controller (discover.c:299) ==22268== by 0x401950: discover_controllers (discover.c:98) ==22268== by 0x401332: main (cpqarrayd.c:164) [ 0] Controller type 'CCISS Controller' at /dev/cciss/c0d0 ==22268== Warning: ignored attempt to set SIGKILL handler in sigaction(); ==22268== the SIGKILL signal is uncatchable Pid is 22314 ==22268== ==22268== ERROR SUMMARY: 8 errors from 5 contexts (suppressed: 4 from 1) ==22268== malloc/free: in use at exit: 2,770 bytes in 9 blocks. ==22268== malloc/free: 68 allocs, 59 frees, 10,077 bytes allocated. ==22268== For counts of detected errors, rerun with: -v ==22268== searching for pointers to 9 not-freed blocks. ==22268== checked 63,160 bytes. ==22268== ==22268== LEAK SUMMARY: ==22268== definitely lost: 2,738 bytes in 7 blocks. ==22268== possibly lost: 0 bytes in 0 blocks. ==22268== still reachable: 32 bytes in 2 blocks. ==22268== suppressed: 0 bytes in 0 blocks. ==22268== Use --leak-check=full to see details of leaked memory. hp385# ==22314== ==22314== Syscall param open(filename) points to unaddressable byte(s) ==22314== at 0x304FDB9C22: __open_nocancel (in /lib64/tls/libc-2.3.4.so) ==22314== by 0x402A8F: cciss_status_check (status.c:206) ==22314== by 0x4015EF: main (cpqarrayd.c:249) ==22314== Address 0x4A1A137 is 0 bytes after a block of size 15 alloc'd ==22314== at 0x4904A06: malloc (vg_replace_malloc.c:149) ==22314== by 0x402260: cciss_interrogate_controller (discover.c:299) ==22314== by 0x401950: discover_controllers (discover.c:98) ==22314== by 0x401332: main (cpqarrayd.c:164) ==22314== ==22314== Syscall param ioctl(generic) points to uninitialised byte(s) ==22314== at 0x304FDBFC19: ioctl (in /lib64/tls/libc-2.3.4.so) ==22314== by 0x40307D: cciss_get_event (cciss_functions.c:165) ==22314== by 0x402AAC: cciss_status_check (status.c:209) ==22314== by 0x4015EF: main (cpqarrayd.c:249) ==22314== Address 0x7FEFFFA59 is on thread 1's stack hp385# kill 22314 hp385# ==22314== ==22314== ERROR SUMMARY: 14 errors from 7 contexts (suppressed: 4 from 1) ==22314== malloc/free: in use at exit: 10,450 bytes in 15 blocks. ==22314== malloc/free: 91 allocs, 76 frees, 38,010 bytes allocated. ==22314== For counts of detected errors, rerun with: -v ==22314== searching for pointers to 15 not-freed blocks. ==22314== checked 62,784 bytes. ==22314== ==22314== LEAK SUMMARY: ==22314== definitely lost: 10,418 bytes in 13 blocks. ==22314== possibly lost: 0 bytes in 0 blocks. ==22314== still reachable: 32 bytes in 2 blocks. ==22314== suppressed: 0 bytes in 0 blocks. ==22314== Use --leak-check=full to see details of leaked memory. Happy hacking, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
