From mboxrd@z Thu Jan  1 00:00:00 1970
From: Olaf Hering <olaf@aepfle.de>
Subject: [PATCH 0 of 4] xentrace [v2]: non-contiguous allocation
 of per-cpu buffer
Date: Tue, 10 May 2011 16:32:35 +0200
Message-ID: <patchbomb.1305037955@localhost>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Return-path: <xen-devel-bounces@lists.xensource.com>
List-Unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xensource.com>
List-Help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-Subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
Sender: xen-devel-bounces@lists.xensource.com
Errors-To: xen-devel-bounces@lists.xensource.com
To: xen-devel@lists.xensource.com
Cc: George Dunlap <george.dunlap@citrix.com>
List-Id: xen-devel@lists.xenproject.org

This series implements non-contiguous trace buffers.

There was an overflow in the offset calculation because it was stored in
an int. Also the calculation of the per-cpu buffer did not take the type
of the offset value into account. This was the reason why I ran into the
checks in bogus(). The crash in out_dealloc: was caused by the lack of
check for the offset value.

Regarding the math added to next_record() another array of pointers could be
added to remove the mfn_to_virt() calls. In my testing 9343 trace pages on 8
cpus are used. Such an array would required additional (9343*8*8)/1024=583Kb.
Which is not too much, given the amount of 291MB tracebuffers.
Such a change is not part of this series.

Olaf