Hi Scott, On Fri, 2015-09-11 at 21:24 -0400, Scott L. Lykens wrote: > > On Sep 11, 2015, at 5:29 PM, Nicholas A. Bellinger <nab@xxxxxxxxxxxxxxx> wrote: > > > > > > There where a number of RCU changes that went in during the v4.2 time > > frame that effect PR code, so I wouldn't be surprised if a regression is > > lurking.. > > > > It would be very helpful to get your target side log while running MSFT > > validate cluster on both v4.2 and v3.19 kernel code with following > > enabled: > > > > echo 'file target_core_pr.c +p' > /sys/kernel/debug/dynamic_debug/control > > > > Note this requires the kernel to be built with CONFIG_DYNAMIC_DEBUG=y, > > and debugfs is mounted at /sys/kernel/debug. > > > > With that information, it should be pretty easy to troubleshoot. > > Thank you, Nicholas. I’ve run the requested logs and placed them on dropbox as follows: > > 3.19 - https://www.dropbox.com/s/alh1nxnr8uaz3mk/pr-319.log?dl=0 > > 4.2 - https://www.dropbox.com/s/51u7h18bee6q0ga/pr-42.log?dl=0 > > Please note that I’ve filtered out the "Allowing implicit CDB: 0x28 > for Write Exclusive Access, Registrants Only reservation on > unregistered nexus” entries for both of these files as under 3.19 it > was generating probably over 100 per second. There were not nearly as > many under 4.2 but there are bursts here and there. Please let me know > if this is a problem. > > Thanks for your prompt response and for any help you can provide. > Thanks for the logs. A patch to address this v4.2+ specific PR regression was just sent out to the list: target: Fix PR registration + APTPL RCU conversion regression http://marc.info/?l=linux-scsi&m=144220263805235&w=2 I've verified that it works as expected on v4.3-rc1 target code using sg_persist, and should get you unblocked on v4.2 code. Note there are a couple obscure PR cases that still need to be verified, but AFAICT what Hyper-V cluster logic needs should be working as expected now. Please let me know once you can verify on your v4.2 qla2xxx setup. Thank you, --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
