[ECOS] Booting Linux from arm

[ECOS] Booting Linux from arm

[Gilles Bulthé]

Hello,

I have difficulties to boot Linux Kernel from my xm27 based platform, and am
wondering whether there is link with RedBoot or not.
My customized RedBoot starts as following :

----------- StartOfQuote-------------------------------------------

... Read from 0x03ee0000-0x03f00000 at 0xa0060000: .
... Read from 0x03ed3000-0x03ed4000 at 0xa007f000: .
Turning on PMIC regulators: 1,2,3,4,5
Unrecognized chip: 0xf8!!!
hardware reset by WDOG
             
Clock input is 24 MHz
Booting from [NOR flash]

RedBoot(tm) bootstrap and debug environment [ROMRAM] Non-certified release,
version FSL 200904 - built 16:22:29, May 22 2009


Platform: Freescale (i.MX27 )  PASS 1.0 [x32 DDR] Copyright (C) 2000, 2001,
2002, 2003, 2004 Red Hat, Inc.
Copyright (C) 2003, 2004, 2005, 2006 eCosCentric Limited

 
RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available
FLASH: 0xa0000000 - 0xa4000000, 512 blocks of 0x00020000 bytes each.
RedBoot>
RedBoot> load -r -m xmodem -b 0x100000
CRaw file loaded 0x00100000-0x0028940f, assumed entry at 0x00100000 xyzModem -
CRC mode, 12585(SOH)/0(STX)/0(CAN) packets, 1 retries
RedBoot> exec
entry=0x80008000, target=0x80008000
Using base address 0x00100000 and length 0x00189410 Uncompressing
Linux.........................................................................................................
done, booting the.

------------EndOfQuote-----------------------------------------------------------
and then, it stops.


I see "RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available"
               -> why is not all the RAM available ?
And : "entry=0x80008000, target=0x80008000"
               -> this entry point 0x8008000 is then out of available range,
isn't it ? Is it possible to change it ?

Regarding my ARM Linux sources and debugging with LED, I have notice that
Linux bugs from the moment when it "__turn_mmu_on". There may be a link.
I am now thinking that the MMU and the memory related lines quoted above may
have link, but do not really know how.

If you have any idea about what's going on, it would be very helpfull !
Thank you in advance for reading me.
Gilles.

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Booting Linux from arm

[Gary Thomas]

Gilles Bulthé wrote:
> Hello,
> 
> I have difficulties to boot Linux Kernel from my xm27 based platform, and am
> wondering whether there is link with RedBoot or not.
> My customized RedBoot starts as following :
> 
> ----------- StartOfQuote-------------------------------------------
> 
> ... Read from 0x03ee0000-0x03f00000 at 0xa0060000: .
> ... Read from 0x03ed3000-0x03ed4000 at 0xa007f000: .
> Turning on PMIC regulators: 1,2,3,4,5
> Unrecognized chip: 0xf8!!!
> hardware reset by WDOG
>              
> Clock input is 24 MHz
> Booting from [NOR flash]
> 
> RedBoot(tm) bootstrap and debug environment [ROMRAM] Non-certified release,
> version FSL 200904 - built 16:22:29, May 22 2009
> 
> 
> Platform: Freescale (i.MX27 )  PASS 1.0 [x32 DDR] Copyright (C) 2000, 2001,
> 2002, 2003, 2004 Red Hat, Inc.
> Copyright (C) 2003, 2004, 2005, 2006 eCosCentric Limited
> 
>  
> RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available
> FLASH: 0xa0000000 - 0xa4000000, 512 blocks of 0x00020000 bytes each.
> RedBoot>
> RedBoot> load -r -m xmodem -b 0x100000
> CRaw file loaded 0x00100000-0x0028940f, assumed entry at 0x00100000 xyzModem -
> CRC mode, 12585(SOH)/0(STX)/0(CAN) packets, 1 retries
> RedBoot> exec
> entry=0x80008000, target=0x80008000
> Using base address 0x00100000 and length 0x00189410 Uncompressing
> Linux.........................................................................................................
> done, booting the.
> 
> ------------EndOfQuote-----------------------------------------------------------
> and then, it stops.
> 
> 
> I see "RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available"
>                -> why is not all the RAM available ?

RedBoot uses the RAM which is listed as "not available"

> And : "entry=0x80008000, target=0x80008000"
>                -> this entry point 0x8008000 is then out of available range,
> isn't it ? Is it possible to change it ?

There are options to 'exec' for this.  Try 'help exec' :-)

Also, the address '0x80008000' is a PHYSICAL address.  The addresses printed
by RedBoot (available range, etc) are LOGICAL addresses - not the same.

> 
> Regarding my ARM Linux sources and debugging with LED, I have notice that
> Linux bugs from the moment when it "__turn_mmu_on". There may be a link.
> I am now thinking that the MMU and the memory related lines quoted above may
> have link, but do not really know how.
> 
> If you have any idea about what's going on, it would be very helpfull !
> Thank you in advance for reading me.
> Gilles.
> 

You really should ask these questions on the Linux-ARM lists, not here,
as they have basically nothing to do with eCos/RedBoot.

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] Booting Linux from arm

[Shah, Amit]

Hi
I had similar issues when I was trying to boot up linux on a imx31 board. Redboot sets up a board id which linux reads and compares with what it is expecting. If I remember correctly there is a cdl file called hal_arm_board.cdl where the board id is setup. I think you are getting "Unrecognized chip: 0xf8!!" because the board id does not match with what linux expects. Check if the decimal value of 0xf8 appears in this file.  

Another question. What kind of file system are you using? From your messages looks like your kernel is booting up but it fails to find or initialize the file system. 
Amit Shah


-----Original Message-----
From: ecos-discuss-owner@ecos.sourceware.org [mailto:ecos-discuss-owner@ecos.sourceware.org] On Behalf Of Gilles Bulthé
Sent: Tuesday, June 09, 2009 9:30 AM
To: ecos-discuss@ecos.sourceware.org
Subject: [ECOS] Booting Linux from arm

Hello,

I have difficulties to boot Linux Kernel from my xm27 based platform, and am
wondering whether there is link with RedBoot or not.
My customized RedBoot starts as following :

----------- StartOfQuote-------------------------------------------

... Read from 0x03ee0000-0x03f00000 at 0xa0060000: .
... Read from 0x03ed3000-0x03ed4000 at 0xa007f000: .
Turning on PMIC regulators: 1,2,3,4,5
Unrecognized chip: 0xf8!!!
hardware reset by WDOG
             
Clock input is 24 MHz
Booting from [NOR flash]

RedBoot(tm) bootstrap and debug environment [ROMRAM] Non-certified release,
version FSL 200904 - built 16:22:29, May 22 2009


Platform: Freescale (i.MX27 )  PASS 1.0 [x32 DDR] Copyright (C) 2000, 2001,
2002, 2003, 2004 Red Hat, Inc.
Copyright (C) 2003, 2004, 2005, 2006 eCosCentric Limited

 
RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available
FLASH: 0xa0000000 - 0xa4000000, 512 blocks of 0x00020000 bytes each.
RedBoot>
RedBoot> load -r -m xmodem -b 0x100000
CRaw file loaded 0x00100000-0x0028940f, assumed entry at 0x00100000 xyzModem -
CRC mode, 12585(SOH)/0(STX)/0(CAN) packets, 1 retries
RedBoot> exec
entry=0x80008000, target=0x80008000
Using base address 0x00100000 and length 0x00189410 Uncompressing
Linux.........................................................................................................
done, booting the.

------------EndOfQuote-----------------------------------------------------------
and then, it stops.


I see "RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available"
               -> why is not all the RAM available ?
And : "entry=0x80008000, target=0x80008000"
               -> this entry point 0x8008000 is then out of available range,
isn't it ? Is it possible to change it ?

Regarding my ARM Linux sources and debugging with LED, I have notice that
Linux bugs from the moment when it "__turn_mmu_on". There may be a link.
I am now thinking that the MMU and the memory related lines quoted above may
have link, but do not really know how.

If you have any idea about what's going on, it would be very helpfull !
Thank you in advance for reading me.
Gilles.

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss



--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] Booting Linux from arm

[Gilles Bulthé]

Thank you for your answers.

I did not find the value 0xF8 in the hal_arm_board.cdl file, neither in
decimal nor in hexa. But I am not sure that the board number is false.
Indeed, the kernel passed the tests "__lookup_processor_type" and
"__lookup_machine_type". Moreover, the "HAL_PLATFORM_MACHINE_TYPE" is well
setup in this file, and corresponds to the value that Linux compares.

I do not use any file system yet. In fact, kernel is not going on after
turning MMU on. The FS is not request in this part, is it ?



Thomas, thanks also for your answer. I could change the "entry=0x80008000,
target=0x80008000".
Actually, the 0x80000000 is the Physical address of the
0x00000000 Logical address, sorry, I did not precise.
Thanks to your answer, I also understood that 0x80008000 (or 0x00008000)
refers to a memory space RESERVED by RedBoot ! I tried to change it with the
options of "exec" command, but no uncompressing even.

But all seems to accept (or force ?) "entry=0x80008000,
target=0x80008000" in both RedBoot and Linux sources ; I do not understand why
RedBoot reserves this address...

Regards,
Gilles.


-----Original Message-----

Hi
I had similar issues when I was trying to boot up linux on a imx31 board.
Redboot sets up a board id which linux reads and compares with what it is
expecting. If I remember correctly there is a cdl file called
hal_arm_board.cdl where the board id is setup. I think you are getting
"Unrecognized chip: 0xf8!!" because the board id does not match with what
linux expects. Check if the decimal value of 0xf8 appears in this file.  

Another question. What kind of file system are you using? From your messages
looks like your kernel is booting up but it fails to find or initialize the
file system.
Amit Shah

-----Original Message-----
From: ecos-discuss-owner@ecos.sourceware.org
[mailto:ecos-discuss-owner@ecos.sourceware.org] On Behalf Of Gilles Bulthé
Sent: Tuesday, June 09, 2009 9:30 AM
To: ecos-discuss@ecos.sourceware.org
Subject: [ECOS] Booting Linux from arm

Hello,

I have difficulties to boot Linux Kernel from my xm27 based platform, and am
wondering whether there is link with RedBoot or not.
My customized RedBoot starts as following :

----------- StartOfQuote-------------------------------------------

.... Read from 0x03ee0000-0x03f00000 at 0xa0060000: .
.... Read from 0x03ed3000-0x03ed4000 at 0xa007f000: .
Turning on PMIC regulators: 1,2,3,4,5
Unrecognized chip: 0xf8!!!
hardware reset by WDOG
           
Clock input is 24 MHz
Booting from [NOR flash]

RedBoot(tm) bootstrap and debug environment [ROMRAM] Non-certified release,
version FSL 200904 - built 16:22:29, May 22 2009

Platform: Freescale (i.MX27 )  PASS 1.0 [x32 DDR] Copyright (C) 2000, 2001,
2002, 2003, 2004 Red Hat, Inc.
Copyright (C) 2003, 2004, 2005, 2006 eCosCentric Limited


RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available
FLASH: 0xa0000000 - 0xa4000000, 512 blocks of 0x00020000 bytes each.
RedBoot>
RedBoot> load -r -m xmodem -b 0x100000
CRaw file loaded 0x00100000-0x0028940f, assumed entry at 0x00100000 xyzModem -
CRC mode, 12585(SOH)/0(STX)/0(CAN) packets, 1 retries
RedBoot> exec
entry=0x80008000, target=0x80008000
Using base address 0x00100000 and length 0x00189410 Uncompressing
Linux..........................................................................................................
done, booting the.

------------EndOfQuote-----------------------------------------------------------
and then, it stops.

I see "RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available"
              -> why is not all the RAM available ?
And : "entry=0x80008000, target=0x80008000"
              -> this entry point 0x8008000 is then out of available range,
isn't it ? Is it possible to change it ?

Regarding my ARM Linux sources and debugging with LED, I have notice that
Linux bugs from the moment when it "__turn_mmu_on". There may be a link.
I am now thinking that the MMU and the memory related lines quoted above may
have link, but do not really know how.

If you have any idea about what's going on, it would be very helpfull !
Thank you in advance for reading me.
Gilles.

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Booting Linux from arm

[Gilles Bulthé]

Hello,

considering the following 2 lines :
>> RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available
>> entry=0x80008000, target=0x80008000
and the reply of Thomas,

I thought that the 0x80008000 (mapped to 0x00008000 with the MMU) address
refers to a non-available address (0x00008000 < 0x0000eb78).
Then, I tried to execute with "exec -t 0x81000000 0x81000000"... But nothing
more happened. Same pb.

Is it true that the "[0x0000eb78-0x03ed1000] available" means that outside
this range, we cannont access ?
What exactly correspond the "entry=0x80008000, target=0x80008000", and is it
possible to change it once for all ?

Thanks and regards,
Gilles.


---------- Original Message -----------
From: Gary Thomas <gary@mlbassoc.com>
To: gilles.bulthe@phelma.grenoble-inp.fr
Cc: ecos-discuss@ecos.sourceware.org
Sent: Tue, 09 Jun 2009 07:44:00 -0600
Subject: Re: [ECOS] Booting Linux from arm

> Gilles Bulthé wrote:
> > Hello,
> >
> > I have difficulties to boot Linux Kernel from my xm27 based platform, and am
> > wondering whether there is link with RedBoot or not.
> > My customized RedBoot starts as following :
> >
> > ----------- StartOfQuote-------------------------------------------
> >
> > ... Read from 0x03ee0000-0x03f00000 at 0xa0060000: .
> > ... Read from 0x03ed3000-0x03ed4000 at 0xa007f000: .
> > Turning on PMIC regulators: 1,2,3,4,5
> > Unrecognized chip: 0xf8!!!
> > hardware reset by WDOG
> >              
> > Clock input is 24 MHz
> > Booting from [NOR flash]
> >
> > RedBoot(tm) bootstrap and debug environment [ROMRAM] Non-certified release,
> > version FSL 200904 - built 16:22:29, May 22 2009
> >
> >
> > Platform: Freescale (i.MX27 )  PASS 1.0 [x32 DDR] Copyright (C) 2000, 2001,
> > 2002, 2003, 2004 Red Hat, Inc.
> > Copyright (C) 2003, 2004, 2005, 2006 eCosCentric Limited
> >
> >  
> > RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available
> > FLASH: 0xa0000000 - 0xa4000000, 512 blocks of 0x00020000 bytes each.
> > RedBoot>
> > RedBoot> load -r -m xmodem -b 0x100000
> > CRaw file loaded 0x00100000-0x0028940f, assumed entry at 0x00100000 xyzModem -
> > CRC mode, 12585(SOH)/0(STX)/0(CAN) packets, 1 retries
> > RedBoot> exec
> > entry=0x80008000, target=0x80008000
> > Using base address 0x00100000 and length 0x00189410 Uncompressing
> >
Linux.........................................................................................................
> > done, booting the.
> >
> >
------------EndOfQuote-----------------------------------------------------------
> > and then, it stops.
> >
> >
> > I see "RAM: 0x00000000-0x03f00000, [0x0000eb78-0x03ed1000] available"
> >                -> why is not all the RAM available ?
>
> RedBoot uses the RAM which is listed as "not available"
>
> > And : "entry=0x80008000, target=0x80008000"
> >                -> this entry point 0x8008000 is then out of available range,
> > isn't it ? Is it possible to change it ?
>
> There are options to 'exec' for this.  Try 'help exec' :-)
>
> Also, the address '0x80008000' is a PHYSICAL address.  The addresses printed
> by RedBoot (available range, etc) are LOGICAL addresses - not the same.
>
> >
> > Regarding my ARM Linux sources and debugging with LED, I have notice that
> > Linux bugs from the moment when it "__turn_mmu_on". There may be a link.
> > I am now thinking that the MMU and the memory related lines quoted above may
> > have link, but do not really know how.
> >
> > If you have any idea about what's going on, it would be very helpfull !
> > Thank you in advance for reading me.
> > Gilles.
> >
>
> You really should ask these questions on the Linux-ARM lists, not here,
> as they have basically nothing to do with eCos/RedBoot.
>
> --
> ------------------------------------------------------------
> Gary Thomas                 |   Consulting for the
> MLB Associates              |     Embedded world
> ------------------------------------------------------------
------- End of Original Message -------


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss