Re: [Qemu-devel] [Consult] microblaze: About running upstream main branch qemu

[Chen Gang]

After a little try, I succeed run microblaze under upstream qemu 2.1.2
and upstream Linux kernel 3.17-rc7, with the xilinux own ramdisk:
"microblaze_complete.cpio.gz". :-)

Compiler is "gcc version 4.9.1 20140717 (Red Hat Cross 4.9.1-1) (GCC)"
for microblaze, and "gcc version 4.8.3 20140624 (Red Hat 4.8.3-1) (GCC)"
for x86_64 qemu.

Next I shall try to use virtio file to share the local data to virtual
guest (I guess, it should be OK).


Thanks again!



On 10/27/2014 09:47 AM, Chen Gang wrote:
> On 10/26/14 21:28, Chen Gang wrote:
>> Hello Guenter Roeck:
>>
>> I still shall try qemu and kernel distribution, next, since it can do
>> it successfully (whether 'sim' can be success or not).
>>
> 
> I got confirmation from microblaze members, at present, microblaze sim
> can not full simulate entire system, so I shall try qemu and kernel
> distribution for it (which can do full simulation for microblaze).
> 
> I shall try my best to finish it within this month.
> 
> Thanks.
> 
>> But excuse me, maybe I can not finish microblaze qemu test within this
>> month, hope I can finish within next month (2014-11-30).
>>
>> Thank your information about qemu, again.
>>
>>
>> Thanks.
>>
>> On 10/22/2014 08:28 AM, Chen Gang wrote:
>>> OK, thanks, there are multiple branches in upstream qemu. Sorry
>>> for my original said, and I can only say: for upstream  develop
>>> master branch, it does not support microblaze.
>>>
>>> Excuse me, I does not mainly focus on only using qemu, and I
>>> also found another easier way for test microblaze toolchain (use sim of 
>>> binutils). So next, I will mainly focus on sim.
>>>
>>> I also met some issues for sim, but I guess, I can analyze it,
>>> maybe also have chances to make patch for it (that is the main
>>> reason why I want to focus on it).
>>>
>>> Thanks.
>>>
>>> Send from Lenovo A788t.
>>>
>>>
>>>
>>>
>>>
>>> Guenter Roeck <address@hidden> wrote:
>>>
>>>
>>>
>>> On Tue, Oct 21, 2014 at 06:41:14PM +0800, Chen Gang wrote:
>>>> On 10/21/14 12:37, Guenter Roeck wrote:
>>>>> On 10/20/2014 08:23 AM, Chen Gang wrote:
>>>>>> On 10/19/2014 10:58 PM, Guenter Roeck wrote:
>>>>>>>
>>>>>>> This doesn't use devicetree, but the configurations are known to be 
>>>>>>> working
>>>>>>> with kernel releases all the way back to kernel version 3.10.
>>>>>>>
>>>>>>
>>>>>> Yeah, really it is !
>>>>>>
>>>>>> After try upstream qemu and upstream kernel, for me, neither them can
>>>>>> work well. For xilinx qemu with xilinux kernel, it should be well,
>>>>>> although I still met issue below, is what I have done incorrect?
>>>>>>
>>>>>
>>>>> Difficult to say.
>>>>>
>>>>> I use gcc 4.8.0 from kernel.org and qemu 2.1.2 with its default 
>>>>> configuration
>>>>> for the microblaze big endian target. No idea if that makes a difference,
>>>>> but I would suggest to start with those.
>>>>>
>>>>> For microblazeel I use gcc 4.9.1 built with buildroot 2014.08 for both
>>>>> kernel and initramfs. Initramfs for that target is built from busybox.
>>>>>
>>>>
>>>> OK, thank you very much for your information.
>>>>
>>>> But sorry, I have to stop trying, because that is far from upstream qemu
>>>> and upstream kernel (at present, I have no enough time resource on it).
>>>>
>>>> My original goal is "use upstream microblaze qemu to test upstream gcc,
>>>> binutils and glibc of microblaze". And now I find another ways: "use sim
>>>> which is in upstream binutils to perform the related test".
>>>>
>>> You lost me. qemu 2.1.2 is the upstream qemu, and I do use upstream binutils
>>> (2.24 if I recall corectly) as well as upstream gcc and upstream kernel.
>>> Did I indicate otherwise in anything I said ? FWIW, should be able to use
>>> qemu from distributions without recompiling it.
>>>
>>> Thanks,
>>> Guenter
>>>
>>
>>
> 


-- 
Chen Gang

Open share and attitude like air water and life which God blessed