[9fans] bootmessages, part one

[9fans] bootmessages, part one

[Sascha Retzki]

Hello everybody,

I'm installing Plan9 on my AMD at the moment, and booting it gives me:
lba 0 -> 0, count 1 -> 1 (1)
 0x00 0x00 0x00 0x00 0x00 0xE0 0x58
command 20
data 80e82c98 limit 80e82e98 dlen 512 status 0 errror 0

several times. It slows the booting process down, of couse ;). It happens after
the "Install or Boot it"-menue, and before it asks me what mouse, and so on I
use.

I recall that these Problems started when I added my beloved Adaptec ATA Raid
1200 PCI controller to the system.

Anybody a rough idea where that comes from in particular? Any fix?

Re: [9fans] bootmessages, part one

[jmk]

The ATA driver issued a 'read sectors' command and did not receive a
'done' interrupt within 30 seconds.
I would guess either the interrupt is not going where we think it is
(the drive is not 'busy' any more according to the status) or the chip
used in the controller needs special code.

Can you get the output of the /bin/pci command?

--jim

On Wed Apr 20 10:27:38 EDT 2005, sretzki@gmx.de wrote:
> 
> 
> Hello everybody,
> 
> I'm installing Plan9 on my AMD at the moment, and booting it gives me:
> lba 0 -> 0, count 1 -> 1 (1)
>  0x00 0x00 0x00 0x00 0x00 0xE0 0x58
> command 20
> data 80e82c98 limit 80e82e98 dlen 512 status 0 errror 0
> 
> several times. It slows the booting process down, of couse ;). It happens after
> the "Install or Boot it"-menue, and before it asks me what mouse, and so on I
> use.
> 
> I recall that these Problems started when I added my beloved Adaptec ATA Raid
> 1200 PCI controller to the system.
> 
> Anybody a rough idea where that comes from in particular? Any fix?

Re: [9fans] bootmessages, part one

[Sascha Retzki]

On Wed, Apr 20, 2005 at 11:49:24AM -0400, jmk@plan9.bell-labs.com wrote:
> The ATA driver issued a 'read sectors' command and did not receive a
> 'done' interrupt within 30 seconds.
> I would guess either the interrupt is not going where we think it is
> (the drive is not 'busy' any more according to the status) or the chip
> used in the controller needs special code.
> 
> Can you get the output of the /bin/pci command?
> 
> --jim

Hello,

I found out that there is no problem if there is no disk attached to the 
controler. However, the corresponding line from /bin/pci is:
0.10.0: 01.80.00 1103/004 10 0:0000ec01 16 1:0000e801 16 2:0000e401 16 3:0000e001 16 4:0000dc01 256

Another question: Is there some device/command that shows the vendor/device
IDs or similar? It'd be a pleasure to grep for "Adaptec" or "controller" or
alike, and get "0.10.0" back ;) .. I actually booted netbsd2, and hope that
it is right; I can't get the /bin/pci-output to my unix-machine atm, some
newbie problems, so I handpasted the above.

> 
> On Wed Apr 20 10:27:38 EDT 2005, sretzki@gmx.de wrote:
> > 
> > 
> > Hello everybody,
> > 
> > I'm installing Plan9 on my AMD at the moment, and booting it gives me:
> > lba 0 -> 0, count 1 -> 1 (1)
> >  0x00 0x00 0x00 0x00 0x00 0xE0 0x58
> > command 20
> > data 80e82c98 limit 80e82e98 dlen 512 status 0 errror 0
> > 
> > several times. It slows the booting process down, of couse ;). It happens after
> > the "Install or Boot it"-menue, and before it asks me what mouse, and so on I
> > use.
> > 
> > I recall that these Problems started when I added my beloved Adaptec ATA Raid
> > 1200 PCI controller to the system.
> > 
> > Anybody a rough idea where that comes from in particular? Any fix?
> 

Re: [9fans] bootmessages, part one

[Mechiel Lukkien]

On Fri, Apr 22, 2005 at 03:15:32PM +0200, Sascha Retzki wrote:
> On Wed, Apr 20, 2005 at 11:49:24AM -0400, jmk@plan9.bell-labs.com wrote:
> > The ATA driver issued a 'read sectors' command and did not receive a
> > 'done' interrupt within 30 seconds.
> > I would guess either the interrupt is not going where we think it is
> > (the drive is not 'busy' any more according to the status) or the chip
> > used in the controller needs special code.
> > 
> > Can you get the output of the /bin/pci command?
> > 
> > --jim
> 
> Hello,
> 
> I found out that there is no problem if there is no disk attached to the 
> controler. However, the corresponding line from /bin/pci is:
> 0.10.0: 01.80.00 1103/004 10 0:0000ec01 16 1:0000e801 16 2:0000e401 16 3:0000e001 16 4:0000dc01 256
> 
> Another question: Is there some device/command that shows the vendor/device
> IDs or similar? It'd be a pleasure to grep for "Adaptec" or "controller" or
> alike, and get "0.10.0" back ;) .. I actually booted netbsd2, and hope that
> it is right; I can't get the /bin/pci-output to my unix-machine atm, some
> newbie problems, so I handpasted the above.

if you want to print vendor/device strings to go with the id's (and
also the type of device), try /n/sources/contrib/mjl/src/pcidevs.tgz.
also available as: 

	http://www.xs4all.nl/~mechiel/plan9/files/pcidevs.tgz

it uses the openbsd pci devices list (download the latest version
from their cvsweb by `mk fetch').

mechiel

Re: [9fans] bootmessages, part one

[Sascha Retzki]

On Fri, Apr 22, 2005 at 07:59:03PM +0200, Mechiel Lukkien wrote:
> On Fri, Apr 22, 2005 at 03:15:32PM +0200, Sascha Retzki wrote:
> > On Wed, Apr 20, 2005 at 11:49:24AM -0400, jmk@plan9.bell-labs.com wrote:
> > > The ATA driver issued a 'read sectors' command and did not receive a
> > > 'done' interrupt within 30 seconds.
> > > I would guess either the interrupt is not going where we think it is
> > > (the drive is not 'busy' any more according to the status) or the chip
> > > used in the controller needs special code.
> > > 
> > > Can you get the output of the /bin/pci command?
> > > 
> > > --jim
> > 
> > Hello,
> > 
> > I found out that there is no problem if there is no disk attached to the 
> > controler. However, the corresponding line from /bin/pci is:
> > 0.10.0: 01.80.00 1103/004 10 0:0000ec01 16 1:0000e801 16 2:0000e401 16 3:0000e001 16 4:0000dc01 256
> > 
> > Another question: Is there some device/command that shows the vendor/device
> > IDs or similar? It'd be a pleasure to grep for "Adaptec" or "controller" or
> > alike, and get "0.10.0" back ;) .. I actually booted netbsd2, and hope that
> > it is right; I can't get the /bin/pci-output to my unix-machine atm, some
> > newbie problems, so I handpasted the above.
> 
> if you want to print vendor/device strings to go with the id's (and
> also the type of device), try /n/sources/contrib/mjl/src/pcidevs.tgz.
> also available as: 
> 
> 	http://www.xs4all.nl/~mechiel/plan9/files/pcidevs.tgz
> 
> it uses the openbsd pci devices list (download the latest version
> from their cvsweb by `mk fetch').
> 
> mechiel
> 


Thank you:
0.10.0  01.80.00  (Mass storage/miscellaneous)  1103/0004 (HighPoint, HPT36x/37x IDE)

NetBSD calls it hptide, if anybody is interested in that sources. 
http://cvsweb.netbsd.org/bsdweb.cgi/~checkout~/src/sys/dev/pci/hptide.c?rev=1.16.6.1&content-type=text/plain

So, after we found out that all, what happens there? And how can that be fixed?

Re: [9fans] bootmessages, part one

[jmk]

	>So, after we found out that all, what happens there? And how can that be fixed?

Someone has to write and debug code for any peculiarities of this chip. For instance,
in the NetBSD driver you mention:

	if ((sc->sc_pp->ide_product == PCI_PRODUCT_TRIONES_HPT366 &&
	    (revision == HPT370_REV || revision == HPT370A_REV ||
	     revision == HPT372_REV)) ||
	    sc->sc_pp->ide_product == PCI_PRODUCT_TRIONES_HPT302 ||
	    sc->sc_pp->ide_product == PCI_PRODUCT_TRIONES_HPT371 ||
	    sc->sc_pp->ide_product == PCI_PRODUCT_TRIONES_HPT372A ||
	    sc->sc_pp->ide_product == PCI_PRODUCT_TRIONES_HPT374) {
		/*
		 * HPT370_REV and highter has a bit to disable interrupts,
		 * make sure to clear it
		 */
		pciide_pci_write(sc->sc_pc, sc->sc_tag, HPT_CSEL,
		    pciide_pci_read(sc->sc_pc, sc->sc_tag, HPT_CSEL) &
		    ~HPT_CSEL_IRQDIS);
	}

--jim

Re: [9fans] bootmessages, part one

[boyd, rounin]

scary.  hand me a dz-11 ...
--
MGRS 31U DQ 52572 12604

Re: [9fans] bootmessages, part one

[jmk]

On Fri Apr 22 15:58:20 EDT 2005, boyd@insultant.net wrote:
> scary.  hand me a dz-11 ...
> --
> MGRS 31U DQ 52572 12604
> 

why?