Solaris 2.6 / Asynchronous IO

Is anyone running with Solaris 2.6 and squid 1.2.b24 (including first patch
set). We've configured with

--enable-icmp
--enable-cache-digests
--enable-async-io
--enable-snmp
--enable-forw-via-db
--enable-delay-pools
--enable-dlmalloc

and it dumps core immediately it starts. It's dying in the dlmalloc.c "free"
routine, a backtrace of the process shows

[snip]
#7 <signal handler called>
#8 free (mem=0x25dcd0) at dlmalloc.c:2352
#9 0x80fb8 in xfree (s=0x25dcd0) at util.c:451
#10 0x3a724 in diskHandleWriteComplete (fd=29, data=0x3a7398, len=146, errcode=0) at disk.c:337
#11 0x26e74 in aioCheckCallbacks () at async_io.c:390
#12 0x38028 in comm_poll (msec=0) at comm_select.c:273
#13 0x54990 in main (argc=4, argv=0xeffffad4) at main.c:580

Not using dlmalloc has some equally unpleasant results (namely, that the process
bus errors at seemingly random intervals, and has to be kill -9'd to
restart it.).

This with gcc 2.8.1. Anyone running a similar combination with success under
Solaris 2.6?

Cheers,

Chris

-- 
Chris Tilbury, UNIX Systems Administrator, IT Services, University of Warwick
EMAIL: cudch+s@csv.warwick.ac.uk PHONE: +44 1203 523365(V)/+44 1203 523267(F)
                            URL: http://www.warwick.ac.uk/staff/Chris.Tilbury