From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tony Lindgren Subject: Re: Detected aborted journal error in Kernel 3.12 Date: Fri, 14 Oct 2016 04:12:00 -0700 Message-ID: <20161014111200.vjpnwlvtdfjfrgee@atomide.com> References: <20160917154123.mfkkhptwvrzavs7y@atomide.com> <20160919181541.45ahl6ktojawwtui@atomide.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from muru.com ([72.249.23.125]:44199 "EHLO muru.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751582AbcJNLME (ORCPT ); Fri, 14 Oct 2016 07:12:04 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Matthijs van Duin Cc: Ankur Tank , linux-omap@vger.kernel.org, linux-mmc@vger.kernel.org * Matthijs van Duin [161013 13:25]: > It would indeed be very useful to have a reproducible case on an > uptodate kernel, but it is definitely not specific to his hardware. > Someone on the beagleboard forum also hit this issue, but sporadically: > > https://groups.google.com/d/msg/beagleboard/5ZwTUQwGSx4/r15mENdfAwAJ > > I'll quote my reply to it: > > On 4 October 2016 at 21:00, Matthijs van Duin wrote: > > I've also run into this once during an apt-get upgrade, leaving the system > > in a pretty hosed state. I managed to recover it but it required a lot of > > effort. Although evidently rare, this is clearly a very serious issue. > > > > The direct cause is edma_prep_slave_sg() failing to allocate memory for the > > struct edma_desc. I don't know whether the kernel is genuinely out of memory > > or if it simply cannot free it up immediately (the allocation is done with > > GPF_ATOMIC), and whether this is because it is filled with a backlog of > > writes or whether a leak of some sort is going on. > > > > Instead of deferring until memory is available, or even just proceeding > > without DMA, the omap_hsmmc driver immediately fails the request with an > > error, thus pretty much guaranteeing loss of data or even filesystem > > corruption. I personally think this is completely unacceptable behaviour of > > a block driver. Yes I can see that being a problem. The fact that omap_hsmmc does not fail back to PIO mode is just strange. We've had the omap1 drivers/mmc/host/omap.c work with and without DMA no problem, no idea why this would not be required also for omap_hsmmc too. > > I've been meaning to persue this matter on the linux-mmc and/or linux-omap > > lists, but since it was a single isolated incident and I have lots of other > > stuff to do I haven't been able to find the time and motivation yet. > > If Ankur has a reproducible case then I think it's time to get > motivated. Ankur, can you please try 4.7 or 4.8 ? Yes please. And also let's set up a new bug in bugzilla.kernel.org for this. Regards, Tony