Kernel booting issues on a MPC8270 board

Kernel booting issues on a MPC8270 board

[annamaya]

I am trying to boot a mvista Linux Kernel ver. 2.4.20
on  a MPC8270 board. I am using the embed_config() in
zImage to pass the board info since the bootloader is
not ppcboot. After the message, "Now booting the
kernel", the board just reboots, as if someone yanked
the reset line. I have BDI hooked up but it doesn't
seem to catch the reset point in the code. And I am
unable to trace this to anything in the code. Any
suggestions? Thanks in advance.


		
__________________________________ 
Do you Yahoo!? 
Yahoo! Mail - now with 250MB free storage. Learn more.
http://info.mail.yahoo.com/mail_250

Re: Kernel booting issues on a MPC8270 board

[annamaya]

I debugged this problem a little more and realized
that the reset was happenning right after the
execution of the first kernel instruction at
0x0000000c. The instruction at this location was
"mfpvr r28". I traced this back to a snippet of code
that was dealing with pipeline depth issues on the
MPC8260. Once I commented this code out, more
instructions were issued but the reboot happens
elsewhere. I am not able to pin-point the location of
the debug since my BDI fails to catch the reset when
it happens. Can someone point me to the cause of this
situation? Thanks in advance.

--- annamaya <annamaya@yahoo.com> wrote:

> I am trying to boot a mvista Linux Kernel ver.
> 2.4.20
> on  a MPC8270 board. I am using the embed_config()
> in
> zImage to pass the board info since the bootloader
> is
> not ppcboot. After the message, "Now booting the
> kernel", the board just reboots, as if someone
> yanked
> the reset line. I have BDI hooked up but it doesn't
> seem to catch the reset point in the code. And I am
> unable to trace this to anything in the code. Any
> suggestions? Thanks in advance.
> 
> 
> 		
> __________________________________ 
> Do you Yahoo!? 
> Yahoo! Mail - now with 250MB free storage. Learn
> more.
> http://info.mail.yahoo.com/mail_250
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
>
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> 



		
__________________________________ 
Do you Yahoo!? 
Read only the mail you want - Yahoo! Mail SpamGuard. 
http://promotions.yahoo.com/new_mail 

Re: Kernel booting issues on a MPC8270 board

[Randy Vinson]

annamaya wrote:
> I debugged this problem a little more and realized
> that the reset was happenning right after the
> execution of the first kernel instruction at
> 0x0000000c. The instruction at this location was
> "mfpvr r28". I traced this back to a snippet of code
> that was dealing with pipeline depth issues on the
> MPC8260. Once I commented this code out, more
> instructions were issued but the reboot happens
> elsewhere. I am not able to pin-point the location of
> the debug since my BDI fails to catch the reset when
> it happens. Can someone point me to the cause of this
> situation? Thanks in advance.

I'm not familiar with the 8260, but this sounds very much like a 
hardware watchdog is expiring somewhere. Check to see if the hardware 
has an on-board watchdog (most likely in the timer section of the 8260) 
and see if the firmware has armed it. If so, you'll need to add a 
snippet of code to disable the watchdog to allow boot-up to complete.

		Randy Vinson

Re: Kernel booting issues on a MPC8270 board

[annamaya]

Thanks for the input Randy. I thought of this too
early on and verified that the Software Watchdog timer
was disabled. And as I pointed it, it first happened
on the very first instruction even when I was single
stepping on the debugger.

--- Randy Vinson <rvinson@mvista.com> wrote:

> annamaya wrote:
> > I debugged this problem a little more and realized
> > that the reset was happenning right after the
> > execution of the first kernel instruction at
> > 0x0000000c. The instruction at this location was
> > "mfpvr r28". I traced this back to a snippet of
> code
> > that was dealing with pipeline depth issues on the
> > MPC8260. Once I commented this code out, more
> > instructions were issued but the reboot happens
> > elsewhere. I am not able to pin-point the location
> of
> > the debug since my BDI fails to catch the reset
> when
> > it happens. Can someone point me to the cause of
> this
> > situation? Thanks in advance.
> 
> I'm not familiar with the 8260, but this sounds very
> much like a 
> hardware watchdog is expiring somewhere. Check to
> see if the hardware 
> has an on-board watchdog (most likely in the timer
> section of the 8260) 
> and see if the firmware has armed it. If so, you'll
> need to add a 
> snippet of code to disable the watchdog to allow
> boot-up to complete.
> 
> 		Randy Vinson
> 



		
__________________________________ 
Do you Yahoo!? 
The all-new My Yahoo! - Get yours free! 
http://my.yahoo.com 
 

Re: Kernel booting issues on a MPC8270 board

[annamaya]

I am sorry but I dont believe the "mfpvr" instruction
has anything to do with it. I thought removing it was
getting rid of the problem but I was wrong. The board
still gets reset when the very first instruction is
executed. I am stumped.

--- annamaya <annamaya@yahoo.com> wrote:

> I debugged this problem a little more and realized
> that the reset was happenning right after the
> execution of the first kernel instruction at
> 0x0000000c. The instruction at this location was
> "mfpvr r28". I traced this back to a snippet of code
> that was dealing with pipeline depth issues on the
> MPC8260. Once I commented this code out, more
> instructions were issued but the reboot happens
> elsewhere. I am not able to pin-point the location
> of
> the debug since my BDI fails to catch the reset when
> it happens. Can someone point me to the cause of
> this
> situation? Thanks in advance.
> 
> --- annamaya <annamaya@yahoo.com> wrote:
> 
> > I am trying to boot a mvista Linux Kernel ver.
> > 2.4.20
> > on  a MPC8270 board. I am using the embed_config()
> > in
> > zImage to pass the board info since the bootloader
> > is
> > not ppcboot. After the message, "Now booting the
> > kernel", the board just reboots, as if someone
> > yanked
> > the reset line. I have BDI hooked up but it
> doesn't
> > seem to catch the reset point in the code. And I
> am
> > unable to trace this to anything in the code. Any
> > suggestions? Thanks in advance.
> > 
> > 
> > 		
> > __________________________________ 
> > Do you Yahoo!? 
> > Yahoo! Mail - now with 250MB free storage. Learn
> > more.
> > http://info.mail.yahoo.com/mail_250
> > _______________________________________________
> > Linuxppc-embedded mailing list
> > Linuxppc-embedded@ozlabs.org
> >
>
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> > 
> 
> 
> 
> 		
> __________________________________ 
> Do you Yahoo!? 
> Read only the mail you want - Yahoo! Mail SpamGuard.
> 
> http://promotions.yahoo.com/new_mail 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: Kernel booting issues on a MPC8270 board

[annamaya]

As someone who has worked on the 8xx processors
before, I went looking for the DER for the 8270 but
was unable to find such a register. If someone can
point me to such a register on the 8270, I would
really appreciate it.

It looks like single-stepping doesn't work till later
in the Kernel. Anyways, I found out after much
debugging that it wasn't the first instruction that
the break was occuring but the code made it all the
way to the "cacheable_memzero" routine and fails in
the "dcbz" instruction. The cpu gets reset as soon as
I try to single-step this instruction. Any ideas as to
why this is happening?

--- Mark Chambers <mchambers@microfirst.com> wrote:

> I'm not familiar with the 8270 per se, but on the
> 8xx you've got this Debug
> Enable Register (DER) where you set different bits
> for things you want the
> processor to jump into debug mode when they happen. 
> You might need to play
> with that to get the BDI to catch whatever is
> happening.  It's probably set
> in the BDI config file somewhere.
> 
> ----- Original Message ----- 
> From: "annamaya" <annamaya@yahoo.com>
> To: "annamaya" <annamaya@yahoo.com>;
> <linuxppc-embedded@ozlabs.org>
> Sent: Friday, January 21, 2005 4:20 PM
> Subject: Re: Kernel booting issues on a MPC8270
> board
> 
> 
> > I am sorry but I dont believe the "mfpvr"
> instruction
> > has anything to do with it. I thought removing it
> was
> > getting rid of the problem but I was wrong. The
> board
> > still gets reset when the very first instruction
> is
> > executed. I am stumped.
> >
> > --- annamaya <annamaya@yahoo.com> wrote:
> >
> > > I debugged this problem a little more and
> realized
> > > that the reset was happenning right after the
> > > execution of the first kernel instruction at
> > > 0x0000000c. The instruction at this location was
> > > "mfpvr r28". I traced this back to a snippet of
> code
> > > that was dealing with pipeline depth issues on
> the
> > > MPC8260. Once I commented this code out, more
> > > instructions were issued but the reboot happens
> > > elsewhere. I am not able to pin-point the
> location
> > > of
> > > the debug since my BDI fails to catch the reset
> when
> > > it happens. Can someone point me to the cause of
> > > this
> > > situation? Thanks in advance.
> > >
> > > --- annamaya <annamaya@yahoo.com> wrote:
> > >
> > > > I am trying to boot a mvista Linux Kernel ver.
> > > > 2.4.20
> > > > on  a MPC8270 board. I am using the
> embed_config()
> > > > in
> > > > zImage to pass the board info since the
> bootloader
> > > > is
> > > > not ppcboot. After the message, "Now booting
> the
> > > > kernel", the board just reboots, as if someone
> > > > yanked
> > > > the reset line. I have BDI hooked up but it
> > > doesn't
> > > > seem to catch the reset point in the code. And
> I
> > > am
> > > > unable to trace this to anything in the code.
> Any
> > > > suggestions? Thanks in advance.
> > > >
> > > >
> > > >
> > > > __________________________________
> > > > Do you Yahoo!?
> > > > Yahoo! Mail - now with 250MB free storage.
> Learn
> > > > more.
> > > > http://info.mail.yahoo.com/mail_250
> > > >
> _______________________________________________
> > > > Linuxppc-embedded mailing list
> > > > Linuxppc-embedded@ozlabs.org
> > > >
> > >
> >
>
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> > > >
> > >
> > >
> > >
> > >
> > > __________________________________
> > > Do you Yahoo!?
> > > Read only the mail you want - Yahoo! Mail
> SpamGuard.
> > >
> > > http://promotions.yahoo.com/new_mail
> > >
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Tired of spam?  Yahoo! Mail has the best spam
> protection around
> > http://mail.yahoo.com
> > _______________________________________________
> > Linuxppc-embedded mailing list
> > Linuxppc-embedded@ozlabs.org
> >
>
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: Kernel booting issues on a MPC8270 board

[annamaya]

I finally figured out my problem had to do with the
fact that my MPC8270 processor was not supported in
the cpu table. I was able to find the definition for
82xx, HiP7 603e entry for CPU table in the 2.4.24 from
denx and copied that table into the cpu table of the
MVista kernel. Otherwise, the cpu_spec[] structure was
defaulting to some entry that set the
CPU_FTR_HPTE_TABLE option in the cpu_spec which
resulted in execution of some code in
cacheable_memzero() routine that resulted in resetting
my CPU and my board. 

Sorry for all the noise but I thought I should share
my debug experience here since someone else might go
through this trying to run MVista kernel on a HiP7
82xx processor.

And for the record, the kernel from denx works great.
I was just trying to figure out why the MVista one
wasn't working.

Thanks for everyone's input on this.

--- annamaya <annamaya@yahoo.com> wrote:

> As someone who has worked on the 8xx processors
> before, I went looking for the DER for the 8270 but
> was unable to find such a register. If someone can
> point me to such a register on the 8270, I would
> really appreciate it.
> 
> It looks like single-stepping doesn't work till
> later
> in the Kernel. Anyways, I found out after much
> debugging that it wasn't the first instruction that
> the break was occuring but the code made it all the
> way to the "cacheable_memzero" routine and fails in
> the "dcbz" instruction. The cpu gets reset as soon
> as
> I try to single-step this instruction. Any ideas as
> to
> why this is happening?
> 
> --- Mark Chambers <mchambers@microfirst.com> wrote:
> 
> > I'm not familiar with the 8270 per se, but on the
> > 8xx you've got this Debug
> > Enable Register (DER) where you set different bits
> > for things you want the
> > processor to jump into debug mode when they
> happen. 
> > You might need to play
> > with that to get the BDI to catch whatever is
> > happening.  It's probably set
> > in the BDI config file somewhere.
> > 
> > ----- Original Message ----- 
> > From: "annamaya" <annamaya@yahoo.com>
> > To: "annamaya" <annamaya@yahoo.com>;
> > <linuxppc-embedded@ozlabs.org>
> > Sent: Friday, January 21, 2005 4:20 PM
> > Subject: Re: Kernel booting issues on a MPC8270
> > board
> > 
> > 
> > > I am sorry but I dont believe the "mfpvr"
> > instruction
> > > has anything to do with it. I thought removing
> it
> > was
> > > getting rid of the problem but I was wrong. The
> > board
> > > still gets reset when the very first instruction
> > is
> > > executed. I am stumped.
> > >
> > > --- annamaya <annamaya@yahoo.com> wrote:
> > >
> > > > I debugged this problem a little more and
> > realized
> > > > that the reset was happenning right after the
> > > > execution of the first kernel instruction at
> > > > 0x0000000c. The instruction at this location
> was
> > > > "mfpvr r28". I traced this back to a snippet
> of
> > code
> > > > that was dealing with pipeline depth issues on
> > the
> > > > MPC8260. Once I commented this code out, more
> > > > instructions were issued but the reboot
> happens
> > > > elsewhere. I am not able to pin-point the
> > location
> > > > of
> > > > the debug since my BDI fails to catch the
> reset
> > when
> > > > it happens. Can someone point me to the cause
> of
> > > > this
> > > > situation? Thanks in advance.
> > > >
> > > > --- annamaya <annamaya@yahoo.com> wrote:
> > > >
> > > > > I am trying to boot a mvista Linux Kernel
> ver.
> > > > > 2.4.20
> > > > > on  a MPC8270 board. I am using the
> > embed_config()
> > > > > in
> > > > > zImage to pass the board info since the
> > bootloader
> > > > > is
> > > > > not ppcboot. After the message, "Now booting
> > the
> > > > > kernel", the board just reboots, as if
> someone
> > > > > yanked
> > > > > the reset line. I have BDI hooked up but it
> > > > doesn't
> > > > > seem to catch the reset point in the code.
> And
> > I
> > > > am
> > > > > unable to trace this to anything in the
> code.
> > Any
> > > > > suggestions? Thanks in advance.
> > > > >
> > > > >
> > > > >
> > > > > __________________________________
> > > > > Do you Yahoo!?
> > > > > Yahoo! Mail - now with 250MB free storage.
> > Learn
> > > > > more.
> > > > > http://info.mail.yahoo.com/mail_250
> > > > >
> > _______________________________________________
> > > > > Linuxppc-embedded mailing list
> > > > > Linuxppc-embedded@ozlabs.org
> > > > >
> > > >
> > >
> >
>
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> > > > >
> > > >
> > > >
> > > >
> > > >
> > > > __________________________________
> > > > Do you Yahoo!?
> > > > Read only the mail you want - Yahoo! Mail
> > SpamGuard.
> > > >
> > > > http://promotions.yahoo.com/new_mail
> > > >
> > >
> > >
> > >
> __________________________________________________
> > > Do You Yahoo!?
> > > Tired of spam?  Yahoo! Mail has the best spam
> > protection around
> > > http://mail.yahoo.com
> > > _______________________________________________
> > > Linuxppc-embedded mailing list
> > > Linuxppc-embedded@ozlabs.org
> > >
> >
>
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> > 
> > 
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam
> protection around 
> http://mail.yahoo.com 
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
>
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com