From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: [Patch] Fix Suspend I/O block/unblock patch Date: 16 Oct 2004 11:39:27 -0500 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <1097944773.2283.123.camel@mulgrave> References: <0B1E13B586976742A7599D71A6AC733C02F20B@xbl3.ma.emulex.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from stat16.steeleye.com ([209.192.50.48]:52940 "EHLO hancock.sc.steeleye.com") by vger.kernel.org with ESMTP id S268779AbUJPQjh (ORCPT ); Sat, 16 Oct 2004 12:39:37 -0400 In-Reply-To: <0B1E13B586976742A7599D71A6AC733C02F20B@xbl3.ma.emulex.com> List-Id: linux-scsi@vger.kernel.org To: James.Smart@Emulex.Com Cc: SCSI Mailing List On Thu, 2004-10-14 at 09:31, James.Smart@Emulex.Com wrote: > In further testing of the fc block/unblock patch, the recommendation to use device_for_each_child() instead of shost_for_each_device() in the target functions was not a good one. The calling sequences on the unblock path, which may eventually block, are not good for device_for_each_child(), which takes and holds a semaphore. Could you give an example of this problem? If there's a demonstrable issue with device_for_each_child(), we can probably persuade Greg to alter it. James