* OF default catch on grub2 on 43p-260
@ 2006-03-07 20:29 Chris Abbey
2006-03-08 17:35 ` Hollis Blanchard
0 siblings, 1 reply; 4+ messages in thread
From: Chris Abbey @ 2006-03-07 20:29 UTC (permalink / raw)
To: grub-devel
I just tried out grub2 on my ppc64 box, a 43p-260 using cvs as of 6 March. It
all seems to have built ok, I didn't see any kind of issues there, trying to
boot it however fails with an default catch from OF:
0 > printenv boot-device
-------------- Partition: common -------- Signature: 0x70 ---------------
boot-device floppy:,grubof.mod
ok
0 > setenv debug all ok
0 > 0 > 0 > printenv debug
debug all
ok
0 > boot Failed to claim heap at 0x4000, len 0x1fb000
DEFAULT CATCH!, code=fff00700 at %SRR0: 00206090 %SRR1: 00023000
ok
0 > setenv boot-device /pci/scsi/sd@8 ok
0 > reset-all
That mess mean anything to anyone? Suggestions for further debug?
--
Chris Abbey - IBM Linux Technology Center
The value of Open Standards is only evident to Open Minds.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: OF default catch on grub2 on 43p-260
2006-03-07 20:29 OF default catch on grub2 on 43p-260 Chris Abbey
@ 2006-03-08 17:35 ` Hollis Blanchard
2006-03-08 21:28 ` Chris Abbey
0 siblings, 1 reply; 4+ messages in thread
From: Hollis Blanchard @ 2006-03-08 17:35 UTC (permalink / raw)
To: grub-devel
On Tuesday 07 March 2006 14:29, Chris Abbey wrote:
> I just tried out grub2 on my ppc64 box, a 43p-260 using cvs as of 6 March.
It
> 0 > setenv debug all ok
> 0 > 0 > 0 > printenv debug
> debug all
> ok
This part here won't work. GRUB doesn't look for an OF variable named
"debug" (technically such a thing does not exist, according to IEEE1275), but
rather a commandline argument. So you would do this:
0 > boot debug=all
or even
0 > boot floppy debug=all
... but in this particular case, the crash is very very early, and also we
already got our helpful debugging message.
> 0 > boot Failed to claim heap at 0x4000, len 0x1fb000
This is telling us that GRUB attempted to claim some memory for its heap (at
that address and that length), but some of it was already in use, so the
request failed.
From the OF prompt, before you boot GRUB, could you do the following and send
the output:
0 > dev /memory@0
0 > .properties
In particular, the "available" property is the one I'm interested in. That
will tell us what areas of memory are considered free...
-Hollis
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: OF default catch on grub2 on 43p-260
2006-03-08 17:35 ` Hollis Blanchard
@ 2006-03-08 21:28 ` Chris Abbey
2006-03-10 4:11 ` Hollis Blanchard
0 siblings, 1 reply; 4+ messages in thread
From: Chris Abbey @ 2006-03-08 21:28 UTC (permalink / raw)
To: grub-devel
On Wednesday 08 March 2006 11:35 am, Hollis Blanchard wrote:
> This part here won't work. GRUB doesn't look for an OF variable named
> "debug" (technically such a thing does not exist, according to IEEE1275),
> but rather a commandline argument. So you would do this:
oops, guess I missread the conversation I was reading, thanks for clearing
that up.
> From the OF prompt, before you boot GRUB, could you do the following and
> send the output:
> 0 > dev /memory@0
> 0 > .properties
> In particular, the "available" property is the one I'm interested in. That
> will tell us what areas of memory are considered free...
0 > dev /memory@0
ok
0 > .properties
name memory
device_type memory
reg 00000000 40000000
available 00004000 00bfc000 01400000 3ec00000
#address-cells 00000001
#size-cells 00000000
0 > boot floppy:,grubof.mod debug=all
Failed to claim heap at 0x4000, len 0x1fb000
DEFAULT CATCH!, code=fff00700 at %SRR0: 00206090 %SRR1: 00023000
ok
0 >
So If I'm reading that correctly, there is 0x00bfc000 bytes availabel at an
offest of 0x00004000, since grub is asking for 0x001fb000, and that is less
than 0x00bfc000... I don't get why it's failing.
(And just for reference: no changes to the setup in nvram and my crusty old
yaboot loads just fine, boots the kernel and I'm off and running... the
machine has been beaten on pretty hard on a regular basis, so hardware wise
things seem to be in good shape.)
--
6OB/7T -=- LTC Embedded Linux -=- Base Lead Nerd and Cat Herder
Cube: 025-3/B207 -=- TieLine: 553.5061 -=- Extern: 507.253.5061
Notes: Christopher Abbey/Rochester/IBM -=- Net: cabbey@us.ibm.com
The value of Open Standards is only evident to Open Minds.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: OF default catch on grub2 on 43p-260
2006-03-08 21:28 ` Chris Abbey
@ 2006-03-10 4:11 ` Hollis Blanchard
0 siblings, 0 replies; 4+ messages in thread
From: Hollis Blanchard @ 2006-03-10 4:11 UTC (permalink / raw)
To: The development of GRUB 2
On Mar 8, 2006, at 3:28 PM, Chris Abbey wrote:
> On Wednesday 08 March 2006 11:35 am, Hollis Blanchard wrote:
>
>> From the OF prompt, before you boot GRUB, could you do the following
>> and
>> send the output:
>> 0 > dev /memory@0
>> 0 > .properties
>> In particular, the "available" property is the one I'm interested in.
>> That
>> will tell us what areas of memory are considered free...
> 0 > dev /memory@0
> ok
> 0 > .properties
> name memory
> device_type memory
> reg 00000000 40000000
> available 00004000 00bfc000 01400000 3ec00000
> #address-cells 00000001
> #size-cells 00000000
> 0 > boot floppy:,grubof.mod debug=all
> Failed to claim heap at 0x4000, len 0x1fb000
> DEFAULT CATCH!, code=fff00700 at %SRR0: 00206090 %SRR1: 00023000
> ok
> 0 >
>
> So If I'm reading that correctly, there is 0x00bfc000 bytes availabel
> at an
> offest of 0x00004000, since grub is asking for 0x001fb000, and that is
> less
> than 0x00bfc000... I don't get why it's failing.
Me neither. In the source, could you try adjusting the 0x4000 (it's
hardcoded in kern/powerpc/ieee1275/init.c) to 0x5000 or 0x8000 or
something?
-Hollis
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2006-03-10 4:11 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-03-07 20:29 OF default catch on grub2 on 43p-260 Chris Abbey
2006-03-08 17:35 ` Hollis Blanchard
2006-03-08 21:28 ` Chris Abbey
2006-03-10 4:11 ` Hollis Blanchard
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.