Hi Boaz, > Hi Christoph > > Many years ago I have sent these exact patches but no-one cared Including > me I guess. I didn't remember your older patches, sorry. > I think my: > scsi_lib: Remove that __scsi_release_buffers contraption > Is more elegant, is layered better and is smaller code. (please > consider my version) I very much disagree - the bidi code uses a separate request for it's payload, uses separate functions to set it up at the low-level so mirroring it with a separate teardown makes sense. This also avoids having to do any bidi check at all in the fast path. > Also the first patch is some more cleanup you'd like. Doesn't look bad, but not that importan either. > The main patch of collapsing scsi_end_request is basically the same. I like the goto version better beause it avoids additional duplication from inside the switch and the bidi path, but it should be functionally equivalent. > Please note the 4th patch which is a real BUG, titled: > scsi_lib: Can't RETRY scsi_cmnd if some bytes were completed That fix seems very hard to read due to the arithmetic comparism on the enum value. The way I try to understand it is that you never want to retry if ((an error happened) && (bytes were completed)) but the explanation should be expanded. > [Your patches have been tested within my MQ testing right?] Yes. -- 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
