On Wed, 2023-07-12 at 00:34 +0000, Quinn Tran wrote: > Hello Nilesh and Marvell > > Any chance to get comments/eyes on this please. > Given its causing system crashes we need to decide how best to deal > with it. > > QT: Laurence, > In understanding the severity, Does end customer uses sg_write_same > as the mechanism to move data? > Other than the sg_write_same utility, how common is end customer uses > 32Byte CDB? It seems like upper layer doesn't have support for > 32Bytes CDB at this time. > > The code path you're modifying is for the T10-PI disk. This disk is > "non-T10-PI" where it may create some confusion for next reader n > Martin on why we've wander down this code path. > > Will queue up a patch that plug this hole. > > > OK, Thank you In this case the customer was specifically using sg_write_same. I am not sure if it was part of a script or some other use case. They were of the opinion it was severe enough of an issue to warrant fixing so they logged a case with us. Thanks for looking into this. Regards Laurence
