Messages in this thread | | | Date | Wed, 19 Mar 2008 19:40:52 +0100 | From | matthieu castet <> | Subject | Re: [PATCH] mass storage : emulation of sat scsi_pass_thru with ATACB |
| |
Matthew Dharm wrote: > Matt -- > > I just noticed that it's been about a week since I sent this... > > Any progress? Sorry I was a bit busy.
Matt > > On Thu, Mar 13, 2008 at 05:51:39PM -0700, Matthew Dharm wrote: >> On Thu, Mar 13, 2008 at 10:22:09PM +0100, matthieu castet wrote: >>> Matthew Dharm wrote: >>>> On Sun, Mar 09, 2008 at 10:42:50PM +0100, matthieu castet wrote: >>>>> Hi >>>>> >>>>> here an update version of the patch >>>> We're getting very close here. >>> + unsigned char sk, asc, ascq; >>> + sk = RECOVERED_ERROR; >>> + asc = 0; /* ATA PASS THROUGH INFORMATION AVAILABLE */ >>> + ascq = 0x1D; >>> + /* XXX we should generate sk, asc, ascq from status and error >>> + * regs >>> + * (see 11.1 Error translation ? ATA device error to SCSI error map) >>> + * and ata_to_sense_error from libata. >>> + */ >>> + sb[1] = sk; >>> + sb[2] = asc; >>> + sb[3] = ascq; >> I may have missed this in an earlier patch, but why to you declare and use >> these 3 variables like this? Seems odd and confusing to me.... I did that because in previous patch I was saving value from the first sense buffer if the first atacb command failed. But after some test, I saw the returned values were always 0 and I always set the same value. I think I can remove them now.
Patch with sk, asc, ascq removed attached.
Matthieu
| |