On Mon, Mar 1, 2010 at 11:24 PM, Srinivas Naga Venkatasatya Pasagadugula - ERS, HCL Tech <satyasrinivasp@xxxxxx> wrote: > Hi, > > Did you tried with latest mvsas patch submitted by me? If not try with attached patch and let me know if you have any issues. > Note: Apply this patch on 2.6.32 kernels. Thanks for the suggestion. Graham Reed also suggested I try your patch. I glanced at the patch -- it seems fairly similar to patch 6/7 from Andy Yan from Nov 2009, although with some differences. I will try both your patch and also Andy Yan's patch series (separately) and see how it goes. > > --Srini. > > -----Original Message----- > From: linux-scsi-owner@xxxxxxxxxxxxxxx [mailto:linux-scsi-owner@xxxxxxxxxxxxxxx] On Behalf Of Ilia Mirkin > Sent: Monday, March 01, 2010 8:58 AM > To: linux-scsi@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx > Subject: mvsas panic > > Hi, > > I had earlier e-mailed about an apparent NCQ vs non-NCQ command issue > on this machine in mvsas/libsas that didn't get any attention > (http://kerneltrap.org/mailarchive/linux-scsi/2010/1/24/6725113), but > now I've run into a crash/panic issue as well... perhaps related which > is why I mention it. > > Unfortunately I don't have a full backtrace, but I'm going to try to > reproduce it tonight with netconsole set up. > > Attached is a screenshot of the crash; the visible part of the trace > goes something like > > mvs_interrupt > mvs_64xx_isr > mvs_int_full > mvs_int_rx > scsi_io_completion > mvs_slot_compete > sas_ata_task_done > ata_qc_complete > __ata_qc_complete > ata_sg_clean > swiotlb_unmap_sg_attrs > ... and then it cuts off. > > After this the machine was completely unresponsive, didn't even ping. > Please let me know if you have any ideas for figuring out what's wrong > here, if there's something I should test, etc. I noticed that there > were a bunch of patches a few months ago for mvsas, but they were > never merged... should I try those? I checked the commits that went > into 2.6.33, that didn't seem to have any relevant changes to mvsas, > libsas, or libata core features, but I can try that too if you think > it's likely to help things out. > > The workload that triggered this was just dd reading and writing > (simultaneously) to 11 of the disks (/dev/zero -> disk, disk -> > /dev/null, bs=1M). > > Some information about the machine: > > Supermicro X8DTE, 2x L5520 cpu, 24G ram (12G per cpu). Areca 1300ix-16 > controller, with 12 drives attached. > > Various virtualization features enabled in bios, like VT-d, SR-IOV, > etc (none of which I'm using, but wanted to play with at some point). > Boot log below. > ... logs trimmed ... -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html