[parisc-linux] Re: [parisc-linux-cvs] linux-2.6 jejb

[Joel Soete]

Hello James,

> 
> On Thu, 2004-07-29 at 11:38, James Bottomley wrote:
> > CVSROOT:	/var/cvs
> > Module name:	linux-2.6
> > Changes by:	jejb	04/07/29 09:38:07
> > 
> > Modified files:
> > 	arch/parisc/kernel: head.S head64.S 
> > 
> > Log message:
> > Correct boot problems of slave CPUs
> 
> ios was HPMCing when starting the secondary CPUs because of missing
> phys<->virt conversions.
> 
> The 32 bit SMP path still didn't seem to be doing the thread struct
> conversion, so I corrected that as well.
> 
> I've tested this on 64 bit, but don't have access to a 32 bit SMP
> machine.
> 
> James
> 
> Index: head.S
> ===================================================================
> RCS file: /var/cvs/linux-2.6/arch/parisc/kernel/head.S,v
> retrieving revision 1.7
> diff -u -r1.7 head.S
> --- head.S	9 Jul 2004 19:04:20 -0000	1.7
> +++ head.S	29 Jul 2004 15:32:34 -0000
> @@ -267,6 +267,8 @@
>  	ldil		L%PA(smp_init_current_idle_task),%sp
>  	ldo		R%PA(smp_init_current_idle_task)(%sp),%sp
>  	ldw		0(%sp),%sp	/* load task address */
> +	tophys_r1	%sp
> +	ldw		TASK_THREAD_INFO(%sp), %sp
>  	mtctl           %sp,%cr30       /* store in cr30 */
>  	addil		L%THREAD_SZ_ALGN,%sp	/* stack is above task */
>  	ldo		R%THREAD_SZ_ALGN(%r1),%sp
> Index: head64.S
> ===================================================================
> RCS file: /var/cvs/linux-2.6/arch/parisc/kernel/head64.S,v
> retrieving revision 1.10
> diff -u -r1.10 head64.S
> --- head64.S	9 Jul 2004 19:04:20 -0000	1.10
> +++ head64.S	29 Jul 2004 15:32:34 -0000
> @@ -301,6 +301,7 @@
>  	/*  Initialize the SP - monarch sets up smp_init_current_idle_task */
>  	load32		PA(smp_init_current_idle_task),%sp
>  	ldd		0(%sp),%sp	/* load task address */
> +	tophys_r1	%sp
>  	ldd		TASK_THREAD_INFO(%sp), %sp
>  	mtctl           %sp,%cr30       /* store in cr30 */
>  	ldo             THREAD_SZ_ALGN(%sp),%sp
> 
I couldn't resist to test it on the n4k but unforunately it failled to boot
:(
the very last message is not very clear (incomplet):
[...]
Mount-cahe hash ....
        -------------------
        Your Sy

Do you think that DEBUG_SPINLOCK would help or define EARLY_BOOTUP_DEBUG?
what would be better?

Thanks again,
    Joel

PS1: Even with 2.4 I never reach to boot this in smp because of Stretch
(iirc) mmu pb which I never reach to fix (Grant or Matthew could certainly
tell you more about this)

PS2: Sorry I don't have any 32bit smp server available to help more

---------------------------------------------------------------------------
Tiscali ADSL LIGHT, 19,95 EUR/mois pendant 6 mois, c'est le moment de faire
le pas!
http://reg.tiscali.be/default.asp?lg=fr