On Thu, 2017-02-16 at 09:54 +0900, Damien Le Moal wrote: > On 2/16/17 01:42, Bart Van Assche wrote: > > An additional concern: what if the size of the Data-Out buffer is not a > > multiple of the logical block size? Shouldn't we round down (good_bytes - > > resid) instead of rounding up resid? > > The only REQ_TYPE_FS request that may not operate on LBA size aligned > payloads is REQ_OP_ZONE_REPORT which is handles in a different case of > the switch statement. So I think it is safe.
Hello Damien, Are you aware that it is the software that submits a request that controls the buffer length and not the device that processes a request? Submitting Data-Out buffers with odd sizes is one of the tests in the libiscsi test suite. See e.g. the source file test-tool/test_write10_residuals.c in that test suite. The request submitted by the libiscsi test suite reach the kernel of the target system either through SG IO or through the iSCSI target driver. When using iSCSI, both the LIO and SCST SCSI target frameworks translate these requests into REQ_TYPE_FS requests. I think we should aim not to affect the outcome of libiscsi tests when the underlying device is a SCSI disk. Thanks, Bart.
