Re: [9fans] Hang at boot.

Re: [9fans] Hang at boot.

[Jean-Baptiste Campesato]

>On Fri Apr  1 14:37:00 EST 2005, steve@quintile.net wrote:
>>some left over debug I reckon
>>
>>/sys/src/boot/pc/etherigbe.c:print("status0 %8.8uX\n", csr32r(ctlr, 
>>Status));
>>
>>-Steve
>
>hmm. could be related to this in detach() which is called soon after that 
>print:
>
>	/* apparently needed on multi-GHz processors to avoid infinite loops */
>	delay(1);
>	while(csr32r(ctlr, Ctrl) & Devrst)
>		;
>
>we should probably time out the loops in there and return a failure value
>instead of void.

Hello,
I'm not sure to understand (I'm French).
I must to calm down the Ghz from bios ? Or search a little Computer ?
Thanks.

[9fans] Hang at boot.

[Jean-Baptiste Campesato]

Hello,
I installed Plan9 for three times and tested to boot with a floppy or Lilo.
And i always got a bug:
On startup I hung with this error : "status0 00000343" (Photo
http://www.a2lf.org/plan9/plan9.jpg) And the LED of the floppy stayed
lighted (Photo http://www.a2lf.org/plan9/floppylect.jpg).
Thanks for your help

Re: [9fans] Hang at boot.

[Steve Simon]

The problem appears to be your Intel Gigabit ether card.
It may be as simple as adding support for the device id of your
card - looking at the source some cards are supported, some not
and yours is not listed; somone else on the list may be able to help more
or you could just experiment.

To get yourself going I suggest you swap the ether card for
an more boring 100Mb one, the wiki lists quite a few supported cards:

http://plan9.bell-labs.com/wiki/plan9/Supported_PC_hardware/index.html

-Steve

Re: [9fans] Hang at boot.

[jmk]

i'm not sure. the unexpected cls message should be benign and
i can't find anywhere in the kernel source where the following
'status0 00000343' could be coming from. but i just did some
greps, i might have missed something.

On Fri Apr  1 13:32:18 EST 2005, steve@quintile.net wrote:
> The problem appears to be your Intel Gigabit ether card.
> It may be as simple as adding support for the device id of your
> card - looking at the source some cards are supported, some not
> and yours is not listed; somone else on the list may be able to help more
> or you could just experiment.
> 
> To get yourself going I suggest you swap the ether card for
> an more boring 100Mb one, the wiki lists quite a few supported cards:
> 
> http://plan9.bell-labs.com/wiki/plan9/Supported_PC_hardware/index.html
> 
> -Steve

Re: [9fans] Hang at boot.

[Steve Simon]

some left over debug I reckon

/sys/src/boot/pc/etherigbe.c:print("status0 %8.8uX\n", csr32r(ctlr, Status));

-Steve

Re: [9fans] Hang at boot.

[jmk]

On Fri Apr  1 14:37:00 EST 2005, steve@quintile.net wrote:
> some left over debug I reckon
> 
> /sys/src/boot/pc/etherigbe.c:print("status0 %8.8uX\n", csr32r(ctlr, Status));
> 
> -Steve

hmm. could be related to this in detach() which is called soon after that print:

	/* apparently needed on multi-GHz processors to avoid infinite loops */
	delay(1);
	while(csr32r(ctlr, Ctrl) & Devrst)
		;

we should probably time out the loops in there and return a failure value
instead of void.

Re: [9fans] Hang at boot.

[lucio]

> some left over debug I reckon
> 
> /sys/src/boot/pc/etherigbe.c:print("status0 %8.8uX\n", csr32r(ctlr, Status));

This reminds me, I have a small change to the NCR driver to address
exactly this type of irritant, best I submit them to patch.

++L

PS: I note the patch/email command is not documented in patch(1).

Re: [9fans] Hang at boot.

[Charles Forsyth]

>>i can't find anywhere in the kernel source where the following
>>'status0 00000343' could be coming from. but i just did some
>>greps, i might have missed something.

it's /sys/src/boot/pc, not /sys/src/9/pc
hence the `loop 84740' after the cpu speed

status0 is always printed by the igbe boot driver on initialisation,
but it might not be at fault.

i'd suspect interrupts were getting lost and since 8086/24db is ATA,
it's mentioned in a pci routing message, and that's often fussy.
*nopcirouting won't work because boot ignores it.

Re: [9fans] Hang at boot.

[Charles Forsyth]

>>i'd suspect interrupts were getting lost and since 8086/24db is ATA,

you can probably ignore all that since i made the mistake of reading
my mail in time order and had not read the subsequent replies to the orginal question.