From mboxrd@z Thu Jan 1 00:00:00 1970 From: Danny Cox Subject: Re: [PATCH] kmalloc old_hwif Date: Fri, 21 May 2004 14:55:19 -0400 Sender: linux-ide-owner@vger.kernel.org Message-ID: <1085165718.833.83.camel@vom> References: <1084717146.3576.3.camel@patibmrh9> <200405211746.35015.bzolnier@elka.pw.edu.pl> <1085163796.833.61.camel@vom> <200405212036.08190.bzolnier@elka.pw.edu.pl> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from blackbox.ecweb.com ([199.72.99.40]:30731 "EHLO blackbox.ecweb.com") by vger.kernel.org with ESMTP id S265988AbUEUSwp (ORCPT ); Fri, 21 May 2004 14:52:45 -0400 In-Reply-To: <200405212036.08190.bzolnier@elka.pw.edu.pl> List-Id: linux-ide@vger.kernel.org To: Bartlomiej Zolnierkiewicz Cc: linux-ide@vger.kernel.org Bartlomiej, On Fri, 2004-05-21 at 14:39, Bartlomiej Zolnierkiewicz wrote: > You do 'goto abort' while not holding locks. Yes, you're correct. Should I be holding the locks? I didn't think that you'd want to call kmalloc() with GFP_KERNEL inside a spinlock because it might sleep. So, I moved the kmalloc() to be one of the first things done to avoid being inside the lock(s). The comment at the top of the function is correct: it IS bonkers! > Anyway this is the minor issue, > the major issue is that... you are late 52 hours... ;-) > > Similar patch from Chris Wedgwood has been merged recently. Cool! Like I said, I mainly wanted to get some form of this idea into ide_unregister(). Thanks! -- Daniel S. Cox Electronic Commerce Systems