Init hangs on Xilinx4

Init hangs on Xilinx4

[schardt]

Hi all,

i still have some problems booting linux from flash and don't know how
to debug.

i have a xilinx4-ppc with 64MB Ram and 4MB Flash, U-Boot works fine,
kernel is booting, but my little hello world program runs as init, died
after printing "He" (intead of "Hello World"). i tried initramfs with
build in cpio-image and jffs2 rootfs

Is there a log-buffer to look at ?
Should i try another compiler version ? (btw. booting from systemace
works with the version i use).

---------------------------
famous last words :

[   50.476641] VFS: Mounted root (jffs2 filesystem).
[   50.533581] Freeing unused kernel memory: 68k init
He
---------------------------


regards
Georg



-------------------------------------------------------------------
-------------------------------------------------------------------
Forschungszentrum Jülich GmbH
52425 Jülich

Sitz der Gesellschaft: Jülich
Eingetragen im Handelsregister des Amtsgerichts Düren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDirig'in Bärbel Brumme-Bothe
Geschäftsführung: Prof. Dr. Achim Bachem (Vorsitzender),
Dr. Ulrich Krafft (stellv. Vorsitzender), Dr. Sebastian M. Schmidt
-------------------------------------------------------------------
-------------------------------------------------------------------

Re: Init hangs on Xilinx4

[David H. Lynch Jr.]

I had a similar problem. In my instance it turned out that I was
inadvertently using the physical address of my uart as the output address.
This worked during boot becasue there was a 1:1 physical-logical entry
stuffed into the MMU. But since Linux did nto know about it, as soon as
the VM system decided to use that TLB entry IO died - right in the
middle of a string.

While it is unlikely you made exactly the same mistake. One area you can
check is memory.

What is your output device ?

schardt wrote:
> Hi all,
>
> i still have some problems booting linux from flash and don't know how
> to debug.
>
> i have a xilinx4-ppc with 64MB Ram and 4MB Flash, U-Boot works fine,
> kernel is booting, but my little hello world program runs as init, died
> after printing "He" (intead of "Hello World"). i tried initramfs with
> build in cpio-image and jffs2 rootfs
>
> Is there a log-buffer to look at ?
> Should i try another compiler version ? (btw. booting from systemace
> works with the version i use).
>
> ---------------------------
> famous last words :
>
> [   50.476641] VFS: Mounted root (jffs2 filesystem).
> [   50.533581] Freeing unused kernel memory: 68k init
> He
> ---------------------------
>
>
> regards
> Georg
>
>
>
> -------------------------------------------------------------------
> -------------------------------------------------------------------
> Forschungszentrum Jülich GmbH
> 52425 Jülich
>
> Sitz der Gesellschaft: Jülich
> Eingetragen im Handelsregister des Amtsgerichts Düren Nr. HR B 3498
> Vorsitzende des Aufsichtsrats: MinDirig'in Bärbel Brumme-Bothe
> Geschäftsführung: Prof. Dr. Achim Bachem (Vorsitzender),
> Dr. Ulrich Krafft (stellv. Vorsitzender), Dr. Sebastian M. Schmidt
> -------------------------------------------------------------------
> -------------------------------------------------------------------
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>   


-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein

Re: Init hangs on Xilinx4

[schardt]

David H. Lynch Jr. wrote:
> I had a similar problem. In my instance it turned out that I was
> inadvertently using the physical address of my uart as the output address.
> This worked during boot becasue there was a 1:1 physical-logical entry
> stuffed into the MMU. But since Linux did nto know about it, as soon as
> the VM system decided to use that TLB entry IO died - right in the
> middle of a string.
>
> While it is unlikely you made exactly the same mistake. One area you can
> check is memory.
>
> What is your output device ?
>
>   
My Output Device is a Xilinx UartLite at 0x40600000 and the kernel
detects this
as /dev/ttyUL0.

Georg







-------------------------------------------------------------------
-------------------------------------------------------------------
Forschungszentrum Jülich GmbH
52425 Jülich

Sitz der Gesellschaft: Jülich
Eingetragen im Handelsregister des Amtsgerichts Düren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDirig'in Bärbel Brumme-Bothe
Geschäftsführung: Prof. Dr. Achim Bachem (Vorsitzender),
Dr. Ulrich Krafft (stellv. Vorsitzender), Dr. Sebastian M. Schmidt
-------------------------------------------------------------------
-------------------------------------------------------------------

Re: Init hangs on Xilinx4

[David H. Lynch Jr.]

schardt wrote:
> Hi David,
>
> thanks for your help.
>   
>>>   
>>>     
>>>       
>>     I am presuming you are using PK's UartLite driver.
>>   
>>     
> I use the Kernel from Grant's git server,  and let me look... yes its
> from Peter Korsgaard :)
>
>   
>>     Try printing out membase some time possibly in ulite_console_setup.
>>     If the value of membase is 0x406000000 that is your problem.
>>     It should be something like 0xfdxxxxxxxx I think.
>>   
>>     
> Mmmh, my EDK Project maps the uartlite to 0x40600000, so i think its
> okay. but i can try it at a higher adress
>
> it is very strange. i use the same hardware setup and boot from
> system-ace without any problems.
    Another posibility that I thought of since you are using Peter K's
driver.
    If you have any interrupt problems I think you could get your
current behavior.
   
    During boot console I/O is not interrupt driven. When the OS sends a
string the whole string gets sent and then the console write exits,
    but very close to running init, the console switches to using the
full driver. PK's driver is interrupt driven. If the UartLite TX FIFO
empty interrupt
    is broke in anyway, it will send until the first time the FIFO fills
and then stop.  The interrupt could be broken in firmware - not properly
connected to the PIC, or it could be broken in software - however your
driver is receiving its interrupt configuration, the driver may be
receiving the incorrect intterrupt #.

    At one point I tried to add timer driven polling to PK's driver
similar to that in the 8250 and others - we have clients that run
without any PIC,
    but trying to figure out what was going on proved sufficiently
difficult I just went back to my own driver.    


>  
>
>
> Regards
> Georg
>
>
> -------------------------------------------------------------------
> -------------------------------------------------------------------
> Forschungszentrum Jülich GmbH
> 52425 Jülich
>
> Sitz der Gesellschaft: Jülich
> Eingetragen im Handelsregister des Amtsgerichts Düren Nr. HR B 3498
> Vorsitzende des Aufsichtsrats: MinDirig'in Bärbel Brumme-Bothe
> Geschäftsführung: Prof. Dr. Achim Bachem (Vorsitzender),
> Dr. Ulrich Krafft (stellv. Vorsitzender), Dr. Sebastian M. Schmidt
> -------------------------------------------------------------------
> -------------------------------------------------------------------
>   


-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein