more..

icp.c:icpErrorComplete() calls comm_close(fd). Problem is that it's
called from error.c:errorSendComplete with also calls comm_close(fd)
(on the same FD).

My solution was to remove the call to comm_close() in
icp.c:icpErrorComplete as the function is only ever called from
errorSendComplete.

After fixing this, I've actually got a running ASYNC_IO squid! :)

Still got a SEGV somewhere on shutReceived: (from list@localhost)
        by nlanr.net (8.8.6/8.8.6) id XAA06758
        for rrg@ny.ubs.com; Fri, 24 Oct 1997 23:26:37 -0700 (PDT)
Resent-Date: Fri, 24 Oct 1997 23:26:37 -0700 (PDT)
To: squid-dev@nlanr.net
Subject: more..
From: "Michael O'Reilly" <michael@metal.iinet.net.au>
Date: 25 Oct 1997 14:26:17 +0800
Message-ID: <x7zpnyl5o6.fsf@metal.iinet.net.au>
Lines: 15
X-Mailer: Gnus v5.4.37/XEmacs 19.15
Resent-Message-ID: <"2AT9NHhJ5l3.0.tT4.G4PKq"@oceana.nlanr.net>
Resent-From: squid-dev@nlanr.net
X-Mailing-List: <squid-dev@nlanr.net> archive/latest/688
X-Loop: squid-dev@nlanr.net
Precedence: list
Resent-Sender: squid-dev-request@nlanr.net

icp.c:icpErrorComplete() calls comm_close(fd). Problem is that it's
called from error.c:errorSendComplete with also calls comm_close(fd)
(on the same FD).

My solution was to remove the call to comm_close() in
icp.c:icpErrorComplete as the function is only ever called from
errorSendComplete.

After fixing this, I've actually got a running ASYNC_IO squid! :)

Still got a SEGV somewhere on shutdown, but it's giving TCP_HITs et
al.

Michael.
Received on Tue Jul 29 2003 - 13:15:44 MDT