Date: Tue, 22 Sep 1998 00:21:33 -0600 (MDT) From: "Kenneth D. Merry" <ken@plutotech.com> To: jmz@FreeBSD.ORG (Jean-Marc Zucconi) Cc: freebsd-current@FreeBSD.ORG Subject: Re: [CAM] boot messages, not configured devices Message-ID: <199809220621.AAA12885@panzer.plutotech.com> In-Reply-To: <199809212329.BAA00396@qix> from Jean-Marc Zucconi at "Sep 22, 98 01:29:31 am"
next in thread | previous in thread | raw e-mail | index | archive | help
--ELM906445293-12734-0_
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Jean-Marc Zucconi wrote...
> >>>>> Kenneth D Merry writes:
>
> > Jean-Marc Zucconi wrote...
> >> Thanks to the latest changes in the ncr driver I am able to boot a cam
> >> kernel. It works well, but I get a lot of
> >> 'MSG_MESSAGE_REJECT received (2:8)' messages during boot (see log
> >> below). What do they mean?
>
> > It looks like one of your CDROM drives doesn't like it when we try to scan
> > additional LUNs. I've attached a patch to cam_xpt.c, see if that elminates
> > the messages.
>
> Indeed, my drives had a 'SC_ONE_LU' entry in the old scsiconf.c
> file. However, your patch does not help - I still get the messages.
Hmm. Well, I'm out of ideas to fix that one for now, unless I somehow
botched the quirk entry in the patch....yep, I did. Sorry about that. Can
you try the new patch I attached?
> >> In addition the cd driver refuses to attach my CD-R if there is
> >> no CD inside at boot time (cd2: fatal error: NOT READY), although it is
> >> correctly probed during the scan phase:
> >> pass5 at ncr1 bus 0 target 3 lun 0
> >> pass5: <PHILIPS CDD2600 1.07> Removable CD-ROM SCSI2 device
> >> pass5: 3.300MB/s transfers
>
> > Well, that's the passthrough driver, not the probe pseudo-peripheral. The
> > passthrough driver will attach to anything that survives the probe phase.
> > (i.e., anything that responds to an inquiry)
>
> [...]
> > There are a couple of possible explanations for this. How long is your bus
> > settle delay? It's possible that it takes a while to figure out that
>
> it always was 8 seconds (8000 milliseconds now :-) and the old drivers
> worked. Increasing to 30 seconds gives the same results.
Okay, that means it's most likely not timing dependent.
> > doesn't have a CD in it or something like that. Try increasing SCSI_DELAY
> > to 30000 (30 seconds). Generally, when CDROM/WORM drives don't have any
> > media on board, they return a status of 0x3a,0x00. (medium not present)
> > If that fixes the problem, you can decrease it gradually until you figure
> > out how long it needs.
>
> It seems that the PHILIPS CDD2600 reply 'not ready' rather than
> 'medium not present' to the READ CAPACITY command if the medium is not
> present.
I agree. Well, try the attached patch, it should make your drive attach.
Let me know how it works. i.e. if you can mount CDs, etc. Assuming it
works okay, I'll have to figure out how to handle this problem. (I guess
I'll either have to accept 0x04,* as nonfatal conditions for CDROM drives
on attach, or do a device-specific quirk entry for this drive.)
Ken
--
Kenneth Merry
ken@plutotech.com
--ELM906445293-12734-0_
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: attachment; filename=cam_xpt.sony.diffs2
Content-Description: cam_xpt.sony.diffs2
Content-Transfer-Encoding: 7bit
==== //depot/cam/sys/cam/cam_xpt.c#156 - /usr/home/ken/perforce/cam/sys/cam/cam_xpt.c ====
*** /tmp/tmp.7316.0 Tue Sep 22 00:19:16 1998
--- /usr/home/ken/perforce/cam/sys/cam/cam_xpt.c Tue Sep 22 00:18:50 1998
***************
*** 300,305 ****
--- 300,313 ----
CAM_QUIRK_NOLUNS, /*mintags*/0, /*maxtags*/0
},
{
+ /* This drive doesn't like multiple LUN probing */
+ {
+ T_CDROM, SIP_MEDIA_REMOVABLE, "SONY",
+ "CD-ROM CDU-80*", "*"
+ },
+ CAM_QUIRK_NOLUNS, /*mintags*/0, /*maxtags*/0
+ },
+ {
/* Default tagged queuing parameters for all devices */
{
T_ANY, SIP_MEDIA_REMOVABLE|SIP_MEDIA_FIXED,
--ELM906445293-12734-0_
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: attachment; filename=jmz.cddiffs
Content-Description: jmz.cddiffs
Content-Transfer-Encoding: 7bit
==== //depot/cam/sys/cam/scsi/scsi_cd.c#84 - /usr/home/ken/perforce/cam/sys/cam/scsi/scsi_cd.c ====
*** /tmp/tmp.12821.0 Tue Sep 22 00:12:31 1998
--- /usr/home/ken/perforce/cam/sys/cam/scsi/scsi_cd.c Tue Sep 22 00:12:00 1998
***************
*** 1711,1717 ****
* the error is anything else, though, we
* shouldn't attach.
*/
! if ((have_sense) && (asc == 0x3a)
&& (error_code == SSD_CURRENT_ERROR))
sprintf(announce_buf,
"Attempt to query device "
--- 1711,1718 ----
* the error is anything else, though, we
* shouldn't attach.
*/
! if ((have_sense)
! && ((asc == 0x3a) || (asc == 0x04))
&& (error_code == SSD_CURRENT_ERROR))
sprintf(announce_buf,
"Attempt to query device "
--ELM906445293-12734-0_--
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199809220621.AAA12885>