On Tue, 2013-05-07 at 12:46 -0700, Kamaljit Singh wrote: > James, Jack, > > Thanks for your responses. > > > I don't understand the question ... it should just work; is it > broken? > sas_create_task() only copies 16 bytes to the sas_ssp_task struct’s > cdb field, which itself is a 16 byte wide array. So it looks like > sas_queuecommand() doesn’t support a 32-byte cdb. Hm, it does it as a bare 16 instead of SCSI_MAX_CDB which is probably how it got missed in the long command conversion. > > It's easy to add that, define cdb 32 to ssp_task and modify > sas_queuecommand, and lldd to pass it to fw > Yes, that’s exactly what I’ve designed it as, i.e. in my lldd’s > queuecommand function. At the moment I only need this function for > 32-byte cdb support. But if there are plans to add that to libsas > then I’d prefer to use sas_queuecommand rather than having to do > manual merges from future libsas fixes/updates. I wouldn't do it that way. Since everything has to form a command IU anyway by copying, I'd just make sas_ssp->cdb a pointer to the actual command instead of an array of the command; if it points to the scsi_cmnd, then it has access to cmd_len as well. That way no copying and no problem when anyone wants longer commands. There should probably also be a macro to populate the command IU since the format of CDB > 16 is different. James -- 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
