From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([140.186.70.92]:50821)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <nick@bytemark.co.uk>) id 1R6KTF-00015q-Gi
	for qemu-devel@nongnu.org; Wed, 21 Sep 2011 06:56:06 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <nick@bytemark.co.uk>) id 1R6KTE-0003yK-R7
	for qemu-devel@nongnu.org; Wed, 21 Sep 2011 06:56:05 -0400
From: nick@bytemark.co.uk
Date: Wed, 21 Sep 2011 11:55:48 +0100
Message-Id: <1316602550-13597-1-git-send-email-nick@bytemark.co.uk>
Subject: [Qemu-devel] block/curl: Remedy a crashing bug completing AIOCBs
	from cache
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: nick@bytemark.co.uk, qemu-devel@nongnu.org
Cc: qemu-trivial@nongnu.org

In QEMU master, attempting to read a cached block from a HTTP (or otherwise)
mounted ISO causes an assert to be triggered, killing the entire QEMU process.
It looks like this:

hw/ide/pci.c:314: bmdma_cmd_writeb: Assertion `bm->bus->dma->aiocb == 
((void *)0)' failed.

The following two patches add flush capability to the curl backend, and avoid
triggering the assert by finishing the AIOCB in a QEMUBH callback, rather than
directly in curl_aio_readv().