From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alex Elder Subject: [PATCH 0/5, v2] libceph clean up con_work() Date: Fri, 22 Feb 2013 11:21:57 -0600 Message-ID: <5127A935.8020605@inktank.com> References: <5127A85D.1070000@inktank.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from mail-ie0-f169.google.com ([209.85.223.169]:37571 "EHLO mail-ie0-f169.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757524Ab3BVRWA (ORCPT ); Fri, 22 Feb 2013 12:22:00 -0500 Received: by mail-ie0-f169.google.com with SMTP id 13so998660iea.14 for ; Fri, 22 Feb 2013 09:22:00 -0800 (PST) In-Reply-To: <5127A85D.1070000@inktank.com> Sender: ceph-devel-owner@vger.kernel.org List-ID: To: "ceph-devel@vger.kernel.org >> ceph-devel" (Re-posting because these changes have been rebased.) This series cleans up con_work() a bit. The original motivation was to get rid of a warning issued by the sparse utility, but addressing that required a little rework and it was fairly straightforward once that was done to make that function fairly simple. The problem sparse reported was really due to sparse not being able to follow the logic between multiple functions that together implement locking. The result of these changes makes both acquiring and releasing the connection mutex occur in con_work(). -Alex [PATCH 1/5] libceph: encapsulate connection backoff [PATCH 2/5] libceph: separate non-locked fault handling [PATCH 3/5] libceph: use a flag to indicate a fault has occurred [PATCH 4/5] libceph: use a do..while loop in con_work() [PATCH 5/5] libceph: indent properly