ski fails to complete boot with debian root fs

ski fails to complete boot with debian root fs

[Peter Chubb]

Hi,
	Using the current kernel (2.6.13-rc3), and sim_defconfig, and
a debian root filesystem created with debootstrap, the kernel boots,
then when attempting to log in as root, displays /etc/issue then hangs.

At this point, it is constantly calling
do_page_fault->handle_mm_fault->pte_fault to try to populate address
0x6007fffffff8000 (Is that the right number of Fs ?) apparently
succeeding, then retrying the fault on return to user space (the value
in question is held in the bspst register)

Rebuilding with CONFIG_ITANIUM works around the problem.


Is there any way to tell Ski which processor to simulate?  It seems to
return Itanium 2 in the cpuid register.

I'm using Glibc 2.3.2.ds1-22 (debian unstable), bash 3.0-15, and ski
version v0.9.81-1 on IA64.

--
Dr Peter Chubb     http://www.gelato.unsw.edu.au  peterc@gelato.unsw.edu.au
You are lost in a maze of BitKeeper repositories,   all slightly different.

Re: [lia64-sim] ski fails to complete boot with debian root fs

[Christian Hildner]

Peter Chubb schrieb:

>Hi,
>	Using the current kernel (2.6.13-rc3), and sim_defconfig, and
>a debian root filesystem created with debootstrap, the kernel boots,
>then when attempting to log in as root, displays /etc/issue then hangs.
>
>At this point, it is constantly calling
>do_page_fault->handle_mm_fault->pte_fault to try to populate address
>0x6007fffffff8000 (Is that the right number of Fs ?) apparently
>succeeding, then retrying the fault on return to user space (the value
>in question is held in the bspst register)
>
>Rebuilding with CONFIG_ITANIUM works around the problem.
>
>
>Is there any way to tell Ski which processor to simulate?  It seems to
>return Itanium 2 in the cpuid register.
>
What Ski simulates is quite close to an Itanium 2. Can you catch the bug 
with the tracing facility?

>I'm using Glibc 2.3.2.ds1-22 (debian unstable), bash 3.0-15, and ski
>version v0.9.81-1 on IA64.
>
The current version is 0.9.81-l6. But I don't know the current 
maintainer at HP. Maybe someone from HP want to comment on this.

Christian