From mboxrd@z Thu Jan 1 00:00:00 1970 From: Michael Schmitz Subject: Re: [PATCH 23/29] atari_scsi: Convert to platform device Date: Sun, 26 Oct 2014 20:37:05 +1300 Message-ID: <544CA4A1.6050802@gmail.com> References: <20141002065628.256592712@telegraphics.com.au> <20141002065633.766165263@telegraphics.com.au> <5444BAB7.5000908@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-m68k-owner@vger.kernel.org To: Finn Thain Cc: "James E.J. Bottomley" , Sam Creasey , linux-scsi@vger.kernel.org, linux-m68k@vger.kernel.org, Geert Uytterhoeven List-Id: linux-scsi@vger.kernel.org Finn, > On Mon, 20 Oct 2014, Michael Schmitz wrote: > > >> Hi Finn, >> >> not certain it is related to this exact patch - the driver crashes >> pretty much on the spot when selecting the first target on the bus. >> > > If it isn't that exact patch then it will probably be one of the other > atari_scsi patches. Most of the other relevant stuff was covered by my own > testing. > > Anyway, thanks for testing and for sending the backtrace. I'll figure out > what went wrong and fix it before I send a new patch set. > > Turns out I missed a reject in patch 29 of the series (hunk #4) which caused a local_irq_save(flags) to go AWOL. Both commenting out the resulting lone local_irq_restore(flags), as well as fixing the reject properly makes the driver behave as it should. No locking races, and no more warnings about bytes stuck in the DMA fifo. We might be on to a winner here :-) Cheers, Michael