Re: Reboot Command Makes kernel to hang (MPC8560)

["Ansari" <ma.harounansari@gdatech.co.in>]

Hi Koller & Kumar

Thanks for ur reply.

Is there a way to reset the full chip (MPC8560)  whenever core reset occu=
rs=20
(using hardware or software) ??

Regards
Haroun Ansari M A





----- Original Message -----=20
From: "Kumar Gala" <galak@kernel.crashing.org>
To: "Clemens Koller" <clemens.koller@anagramm.de>
Cc: "Ansari" <ma.harounansari@gdatech.co.in>; <linuxppc-embedded@ozlabs.o=
rg>
Sent: Tuesday, July 31, 2007 7:41 PM
Subject: Re: Reboot Command Makes kernel to hang (MPC8560)



On Jul 31, 2007, at 6:44 AM, Clemens Koller wrote:

> Hi, Ansari!
>
> Ansari schrieb:
>> Hi Kumar,
>> First of all thanks for ur reply .
>> Even i went through the linux source . And i have observe that the=20
>> reboot command used to hard reset the core . I have few doubts can  u=20
>> please clarify me.
>> 1. Is there any way to reset the full chip with out using any  externa=
l=20
>> signal (MPC8560) ? (like any register that can be used  for reseting t=
he=20
>> processor)
>
> I RTFM:
> It should be the bits RST[1:0] in the Debug Control Register 0  (DBCR0)=
.

This only resets the core on the 8560.

> I didn't find details how the external signals are affected:  HRESET_RE=
Q#=20
> and friends.
> The HRESET_REQ# is usually fed back to the CPU's HRESET#.
> So if the HRESET_REQ# gets asserted by writing to above registers  it=20
> should really bring
> down the CPU, it's internal as well as it's external components,  which=
=20
> are usually
> connected to a replication of that signal.

This is roughly correct.  The only way on 8560 to generate
HRESET_REQ# is to cause a core watchdog timeout.

> However the existence of cpm2_reset() and a qe_reset() (QuiccEngine?)
> in the code tells me that the above expectations could be wrong.
>
> Would be nice to have that verified by some hardware guys from=20
> freescale...

cpm2_reset/qe_reset are more related to SW than any HW reset.

>
>> 2. Even same reboot command works fine for MPC8540 Processor ?.
>
> ...because it doesn't have a cpm ?

That's more luck than anything else.

>> 3. what are the factors that makes ramdisk hangs . When its=20
>> uncompressing ?
>
> Well, side effects ?
>
> Regards,
> --=20
> Clemens Koller
> __________________________________
> R&D Imaging Devices
> Anagramm GmbH
> Rupert-Mayer-Stra=DFe 45/1
> Linhof Werksgel=E4nde
> D-81379 M=FCnchen
> Tel.089-741518-50
> Fax 089-741518-19
> http://www.anagramm-technology.com