quoted
quoted
quoted
quoted
"Tom" == Tom Yan [off-list ref] writes:
Tom,
Tom> Sounds like libata-scsi is doing something wrong then. It should
Tom> not set max_hw_sectors to dev->max_sectors that is set by
Tom> libata-core:
Tom> but instead it should report it as Maximum Transfer Length and let
Tom> sd set it as max_dev_sectors:
Well, it's mostly academic since there is a 1:1 mapping. It is important
in SCSI where you have many devices with different properties hooked up
to the same controller.
SCSI MTL is a recent addition. There wasn't much point in making a
distinction between hw and dev outside of sd.c since it is the only
device type that has a way to communicate a transfer limit.
Tom> Although that means setting dev->max_sectors to a value larger than
Tom> 1024 will probably be a no-op, if that's really an issue, we should
Tom> have the host templates in libata updated.
I am still not sure exactly which problem it is you are trying to fix?
--
Martin K. Petersen Oracle Linux Engineering