5.3 ata / atapicam issues (update)

Thu Feb 3 02:51:47 PST 2005

Hello,

I think time has come to fill a problem report.
I did started a initial one, which is attached
to this mail.

Can you read through the attached PR and comment
its contents, please?

Rob schrieb:
>>Hope this helps someone. I'm not at all sure
>>where/what the problem is but
>>I'll be happy to run on the embedded controller for
>>the time being. (while
>>PIO 4 != DMA 133, reliability/functionality is
> 
> king!)   
> 
> Neither have I an idea what going on here, but
> after having upgraded a few PCs from 4.10 to 5.3,
> I empirically deduced following conclusions:
> 
> 1) The problems is since 5.3, not with any 4.X.
> 
> 2) Such problems occur when the media (harddisk in my
>   case) are capable of doing a higher speed than the
>   motherboard. For example, harddisk can do maximum
>   UDMA100, but motherboard can maximum handle UDMA66.
>   In such cases I need  hw.ata.ata_dma="0"  in
>   loader.conf, which slows down the speed to PIO4.
> 
> Without the hw.ata.ata_dma="0", I get various
> results, which can be 'kernel panic at bootup after
> installation' on one PC, to 'odd crashing of X for
> no good reason' on another PC.
> 
> When the maximum speeds of harddisk and motherboard
> match, there's no problem, eventhough they may both
> operate at a lower than maximum speed.
> 
> Something really uggly has entered the code of 5.3,
> with respect to disk speed matching, I believe.
> 
> Regards,
> Rob.
> 
> 
> 		
> __________________________________ 
> Do you Yahoo!? 
> Yahoo! Mail - 250MB free storage. Do more. Manage less. 
> http://info.mail.yahoo.com/mail_250
> _______________________________________________
> freebsd-questions at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> To unsubscribe, send any mail to "freebsd-questions-unsubscribe at freebsd.org"
-------------- next part --------------
SEND-PR: -*- send-pr -*-
SEND-PR: Lines starting with `SEND-PR' will be removed automatically, as
SEND-PR: will all comments (text enclosed in `<' and `>').
SEND-PR:
SEND-PR: Please consult the following URL if you are not sure how to
SEND-PR: fill out a problem report:
SEND-PR: http://www.freebsd.org/doc/en/articles/problem-reports/
SEND-PR:
SEND-PR: Note that the Synopsis field is mandatory.
SEND-PR:
SEND-PR: Please note that (unless you state otherwise) if your report 
SEND-PR: includes a patch then it will be taken under the same license as 
SEND-PR: the one on the file(s) you want to change.
SEND-PR:
SEND-PR: BE ADVISED THAT FREEBSD PROBLEM REPORTS ARE PUBLIC INFORMATION AND
SEND-PR: WILL BE PUBLISHED AS-IS ON THE PROJECT'S MAILING LISTS AND WEB SITES.
SEND-PR: DO NOT SUBMIT ANY INFORMATION YOU DO NOT WANT MADE PUBLIC.
SEND-PR:
SEND-PR: If you wish to submit a problem report confidentially, then contact
SEND-PR: the FreeBSD bugmaster (bugmaster at FreeBSD.org) to arrange for a
SEND-PR: relevant developer to be contacted.
SEND-PR:
SEND-PR: For sensitive security issues, consider contacting the FreeBSD
SEND-PR: security officer team (security-officer at freebsd.org) directly.
SEND-PR:
SEND-PR: Choose from the following categories:
SEND-PR:
SEND-PR: advocacy  alpha     amd64     bin       conf      docs      
SEND-PR: gnu       i386      ia64      java      kern      misc      
SEND-PR: ports     powerpc   sparc64   standards threads   usb       
SEND-PR: www       
SEND-PR:
To: FreeBSD-gnats-submit at freebsd.org
From: Daniel S. Haischt <me at daniel.stefan.haischt.name>
Reply-To: Daniel S. Haischt <me at daniel.stefan.haischt.name>
Cc: 
X-send-pr-version: 3.113
X-GNATS-Notify: 

>Submitter-Id:	current-users
>Originator:	Daniel S. Haischt
>Organization:	Daniel S. Haischt IT Consulting
>Confidential:	no
>Synopsis:	FreeBSD 5.3 ATA unreliability if using ATAPICAM
>Severity:	serious
>Priority:	high
>Category:	kern
>Class:		sw-bug
>Release:	FreeBSD 5.3-STABLE i386
>Environment:
System: FreeBSD abyssone.abyssworld.de 5.3-STABLE FreeBSD 5.3-STABLE #0: Thu Jan 20 13:07:40 CET 2005 root@:/usr/obj/usr/src/sys/ABYSSONE i386

* PI 233 MMX with an additional Promise PDC20269 UDMA133 controller
* ATA 100/133 IDE cable
* IBM disk: 78533MB <IC35L080AVVA07-0/VA4OA52A> [170197/15/63] at ata2-master UDMA100
* Plextor CDROM drives
  + CDRW <PLEXTOR CD-R PREMIUM/1.05> at ata3-master UDMA33
  + DVDROM <PLEXTOR DVD-ROM PX-116A2 0100/1.00> at ata3-slave UDMA66

>Description:
This problem report is a synthesis of several issues posted to the
FreeBSD questions mailing list!

The ATA issues were reported numeros times by people using a Promise
PDC20269 UDMA133 controler, so I will stick with that controller.

If using the just mentioned controller, especially the 2nd IDE channel,
it was observed that during boot time this particular hardware configuration
may cause the following issues:

* Interrupt Storms
* READ_DMA timed out
* Kernel Panics

To trace the issue, the following hardware configurations has been tested:

Configuration #1 (success):
 * harddrive attached to IDE channel #1
 * no device attached to IDE channel #2
Configuration #2 (failure):
 * harddrive attached to IDE channel #1
 * CDROM drive(s) attached to IDE channel #2
Configuration #3 (failure):
 * no device attached to IDE channel #1
 * harddrive attached to IDE channel #2
Configuration #4 (failure):
 * CDROM drive(s) attached tp IDE channel #1
 * harddrive attached to IDE channel #2
Configuration #5 (success):
 * no devices attached to the external
   Promise controller
 * harddrive attached to the motherboard's
   IDE controller
 * CDROM drive(s): ditto

Additionally, to reduce problems, the following
additon were made to loader.conf ...

 -> the hw.ata.ata_dma="0"

>How-To-Repeat:
Attach any device (disk or CDROM) to a 2nd channel of the previously
mentioned Promise controller using a ATA 100/133 IDE cable. Do not
use the motherboard's IDE controller.
>Fix:

This is not a fix! But it seems necessary that the ATA code which
drives for example a Promise PDC20269 controller, has to be tested
against the issue and hardware configurations described by this PR.