On 01/26/2017 07:37 PM, Dan Zach wrote:
> On Thursday, 26 January 2017 12:33:45 UTC+2, Ralf Ramsauer  wrote:
>> On 01/26/2017 11:10 AM, Dan Zach wrote:
>>> On Thursday, 26 January 2017 11:04:37 UTC+2, Ralf Ramsauer  wrote:
>>>> Hi Dan,
>>>>
>>>> On 01/26/2017 09:29 AM, Dan Zach wrote:
>>>>> I am in process of starting the non-root linux on the inmate.
>>>>> The inmate RAM starts at 0xe8000000 (phys and virt).
>>>>> The memory is carved out Ok from the root cell by using mem=1650M.
>>>>>
>>>>> I am using:
>>>>>         $ LOAD_ADDR=0xe9000000    
>>>>>   $ DTB_ADDR= 0xea000000
>>>>>   $
>>>>>   $ jailhouse cell create jetson-tk1-linux-demo.cell
>>>>>   $ jailhouse cell load --name jetson-tk1-linux-inmate-demo \
>>>>>                         linux-loader.bin -a 0x0 \
>>>>>                         Image -a $LOAD_ADDR \
>>>>>                         configs/dts/inmate-jetson-tk1.dtb -a $DTB_ADDR \
>>>>>                         -s "kernel=$LOAD_ADDR dtb=$DTB_ADDR" -a 0x1000
>>>> you have to load command line arguments to 0x100 for ARM architecture.
>>>> 0x1000 is for ARMv8.
>>>>
>>>> Why don't you use jailhouse cell linux?
>>>>
>>>>   Ralf
>>>>>
>>>>> The Image file is the one created by vanilla 4.8.2 build with
>> DGBCONSOLE patch
>>>>> First I saw that command line parsing returns 0 for both entry and dtb,
>>>>> so I saw linux-loader jumps to itself in endless loop.
>>>>>
>>>>> I forced the entry to be at 0xe9000000 and dtb at 0xea000000, just
>> gets stuck.
>>>>> So what do I do wrong here?
>>>>>
>>>>> Thanks
>>>>> Dan
>>>>>
>>>>
>>>>
>>>> --
>>>> Ralf Ramsauer
>>>> GPG: 0x8F10049B
>>>
>>> I followed the non-linux-arm64 doc, assuming the other one is for x86...
>> Nope, x86 uses 0xf0000 :-)
>>
>> Just have a look at the linker files in inmates/lib/*/*.lds, .cmdline
>> section.
>>>
>>> So I should I go with this:
>>>
>>>  jailhouse cell linux /path/to/linux.cell /path/to/bzImage \
>>>         -i /path/to/initrd -c "console=ttyS0,1152000"
>>>
>>> - And the script takes care of laying out the image inside the cell?
>> Yes, cell linux automatically searches for a proper free memory area and
>> patches the device tree binary. It prepares everything for you.
>>
>> Ah, don't forget to provide the device tree binary with --dtb.
>>> - Would it boot without initrd for starters?
>> Yes, and I would recommend to start without providing an initrd.  We had
>> a recent discussion on that [1].  It might occur, that your unpacked
>> kernel overlaps with the initrd but still, you should see your kernel
>> booting.
>>   Ralf
>>
>> [1]
>> https://groups.google.com/forum/#!msg/jailhouse-dev/r2ON07be3TI/tETOqmIJBgAJ
>>>
>>> Thanks
>>> Dan
>>>
> 
> Its alive! Thanks Ralf.
> [    0.871096] VFS: Cannot open root device "(null)" or unknown-block(0,0): 
> error -6
> No it reaches till :
> 
> 
> [    0.878614] Please append a correct "root=" boot option; here are the 
> available partitions:
> [    0.886999] Kernel panic - not syncing: VFS: Unable to mount root fs on 
> unknown-block(0,0)
> 
> 
> I guess its because there is no initrd?
Yes, that's a "cannot mount root, don't know what to do"-panic, which is
absolutely legitimate in this case. Try to provide your initrd.

  Ralf

PS: Don't know if you know about it, but that's what I like to use for
building initrds: https://buildroot.org/

-- 
Ralf Ramsauer
GPG: 0x8F10049B

-- 
You received this message because you are subscribed to the Google Groups 
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to