1.1.14 segmentation fault found

From: Michael Pelletier <mikep@dont-contact.us>
Date: Mon, 28 Jul 1997 15:56:25 -0400 (EDT)

(gdb) run -f /usr/local/squid/etc/squid-test.conf
Starting program: /usr/local/squid/bin/./squid-1.1.14+retry -f
/usr/local/squid/etc/squid-test.conf

Program received signal SIGSEGV, Segmentation fault.
icpDetectClientClose (fd=30, data=0x0) at icp.c:1982
icp.c:1982: No such file or directory.
(gdb) where
#0 icpDetectClientClose (fd=30, data=0x0) at icp.c:1982
#1 0xa19c in comm_select (sec=1) at comm.c:1134
#2 0x1edde in main (argc=3, argv=0xefbfda80) at main.c:702
(gdb)

-----------
   1976 void
   1977 icpDetectClientClose(int fd, void *data)
   1978 {
   1979 icpStateData *icpState = data;
   1980 LOCAL_ARRAY(char, buf, 256);
   1981 int n;
-> 1982 StoreEntry *entry = icpState->entry; <-----------
   1983
   1984 errno = 0;

It's referencing a null pointer here ("data=0x0"), which is causing the
segmentation fault. Under what circumstances would icpDetectClientClose()
be called with NULL data? Is this because I don't have any ICP peers
configured?

I managed to induce this failure by clicking around the msnbc.com site for
about 5-10 minutes. I was the only user of the proxy at the time, since
I'd set it up on a separate port from my production proxy.

        -Mike Pelletier.
Received on Mon Jul 28 1997 - 13:02:05 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:35:51 MST