>>>>> "Brian" == Brian King <brking@xxxxxxxxxxxxxxxxxx> writes: Brian, Brian> As defined in the transfer limit table in the block limits VPD Brian> page section of SBC4, this value is only supposed to apply to a Brian> small subset of SCSI opcodes. However, we are using that value Brian> for all SCSI commands. Well, I guess the real issue here is that we use the same constraint for FS and BLOCK_PC commands. It's not really the kernel's business to enforce target device preferences on sg/bsg callers. The main headache here is that max_sectors_kb in sysfs is widely used by users that want to tweak their read/write I/O size. So we can't really change the meaning of that knob, nor do I think it's necessarily a good idea to introduce a max_rw_sectors_kb (which I agree would have been a better name for the sysfs knob in the first place). So I think my preference would be to have an internal max_rw_sectors queue limit that sd can set based on the VPD. And use that as the limit for REQ_TYPE_FS requests. I.e. max_sectors and max_sectors_kb need to be constrained by min_not_zero(max_rw_sectors, max_hw_sectors). max_hw_sectors would then only indicate the HBA hardware request size limit which is what BLOCK_PC callers should be constrained by exclusively. That way you don't get into having to parse the CDB and other evils. -- Martin K. Petersen Oracle Linux Engineering -- 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
