AIC7902 w/ seagate U320 drive issue on releng-4 (and current)

[ict technician]

On Wednesday 08 October 2003 7:11 am, Andre Albsmeier wrote:
> On Tue, 07-Oct-2003 at 16:37:35 +0100, ict technician wrote:
> > On Tuesday 07 October 2003 1:25 pm, Andre Albsmeier wrote:
> > > On Tue, 07-Oct-2003 at 12:54:08 +0100, ict technician wrote:
> > > > > On Sat, 26 Jul 2003, Don Bowman wrote:
> > > > > > Jul 26 19:27:20  /kernel: <<<<<<<<<<<<<<<< Dump Card State Ends
> > > > >
> > > > >FYI, i've solved this problem for me by moving to
> > > > >firmware version 5 on the ST318453LW (U320 15KRPM 18GB)
> > > > >seagate drive.
> > > > >
> > > > >If you are seeing the problem I mentioned with U320
> > > > >seagate drive and AIC7902 controller I highly recommend
> > > > >contacting seagate and requesting the latest rom.
> > > > >Sadly the actual changes are 'proprietary' so I cannot
> > > > >be sure this was the root cause of the problem. I'm
> > > > >guessing there is an issue with large # of tags outstanding,
> > > > >but its only a hunch.
> > > > >
> > > > >I have code to program this ROM under FreeBSD i
> > > > >will share with anyone who asks. I obtained this
> > > > >from andre __AT__ albsmeier __DOT__ net, who had
> > > > >posted information to this list in the past [thank
> > > > >you!].
> > > > >
> > > > >I cannot be 100% sure this fixes the problem, but it was
> > > > >pretty easy to reproduce with 'dd if=/dev/da0 of=/dev/null bs=20k',
> > > > >and now I cannot reproduce on the 15 systems I am trying.
> > > > >
> > > > >--don
> > > >
> > > > I'm still seeing occasional problems. I can't find my previous posts
> > > > either (on stable probably).
> > > >
> > > > Anyway, I have:
> > > > Adaptec 29320
> > > > ST336607LW P/N 9V4005-002 firmware r=0006
> > > > (4x Vinum RAID10)
> > > >
> > > > tags depth set to 31.
> > > >
> > > > Seagate have released a new firmware revision. Can you send me your
> > > > update
> > >
> > > Here is the software from the one who wrote it :-). Please read the
> > > first 70 lines and tell me about success or failure.
> > >
> > > 	-Andre
> >
> > Update seemed to go okay.firewall# ./fwprog -s /dev/da3
> > ST336607LW_S_007.SEA
> >
> > Updating SEAGATE ST336607LW      00063JA2403Q
> >
> > NOTE: Skipping 80 byte header.
> > -------------------------------------------------
> > PktNo.  PktSize        BytesRemaining   LastPkt
> > -------------------------------------------------
> >   0   32768 (0x08000)    504832 (0x07B400)   0
> >   1   32768 (0x08000)    472064 (0x073400)   0
> >   2   32768 (0x08000)    439296 (0x06B400)   0
> >   3   32768 (0x08000)    406528 (0x063400)   0
> >   4   32768 (0x08000)    373760 (0x05B400)   0
> >   5   32768 (0x08000)    340992 (0x053400)   0
> >   6   32768 (0x08000)    308224 (0x04B400)   0
> >   7   32768 (0x08000)    275456 (0x043400)   0
> >   8   32768 (0x08000)    242688 (0x03B400)   0
> >   9   32768 (0x08000)    209920 (0x033400)   0
> >  10   32768 (0x08000)    177152 (0x02B400)   0
> >  11   32768 (0x08000)    144384 (0x023400)   0
> >  12   32768 (0x08000)    111616 (0x01B400)   0
> >  13   32768 (0x08000)     78848 (0x013400)   0
> >  14   32768 (0x08000)     46080 (0x00B400)   0
> >  15   32768 (0x08000)     13312 (0x003400)   0
> >  16   13312 (0x03400)         0 (0x000000)   1
> >
> > Download successful
> >
> > The ID string didn't change tho'
> >
> > da0 at ahd1 bus 0 target 0 lun 0
> > da0: <SEAGATE ST336607LW 0006> Fixed Direct Access SCSI-3 device
> > da0: 320.000MB/s transfers (160.000MHz, offset 63, 16bit), Tagged
> > Queueing Enabled
> >
> > ...
>
> Oops, didn't notice you sent this mail to the list as well... I should
> have made a group reply when answering you yesterday :-).
>
> Just for reference: The -s flag doesn't download the fw to the drive. It
> is just showing what would have been done. You might want to try it again
> without the -s flag.
>
> I have now added some printf's to the code which clarify that the -s flag
> had been used and no download was actually done :-).
>
> 	-Andre

Doh!

I did say I was having one of those weeks. In my defence the "download 
successful" message is somewhat counter-intuitive :))

Spookily no card dumps have happened (kiss of death). Lunch time is our peak 
load with kids on Internet games. I'll drop the upgrade in just before.

Thanks

-- 
i j hart

ICT Technician
Cardinal Newman Catholic School & Community College