Re: [Xen-devel] [PATCH v16 4/9] x86: add multiboot2 protocol support for EFI platforms

[Doug Goldstein <cardoe@xxxxxxxxxx>]

On 3/15/17 9:38 AM, Daniel Kiper wrote:
> On Wed, Mar 15, 2017 at 09:27:27AM -0500, Doug Goldstein wrote:
>> On 3/15/17 6:35 AM, Daniel Kiper wrote:
>>> On Thu, Mar 09, 2017 at 02:02:49PM -0600, Doug Goldstein wrote:
>>>
>>> [...]
>>>
>>>> Still missing 'xl info'.
>>>
>>> Got Intel NUC5i3MYHE (internally it is NUC5i3MYBE board) into my hands.
>>> I have put 8 GiB RAM and 500 GB SATA 3 into it. Updated BIOS/EFI to 0041
>>> version (it is the latest one). Installed latest Debian testing (Debian
>>> GNU/Linux 9 (stretch)), built GRUB2 and Xen, with and without relocation
>>> patches, on it. Everything works (I left machine working last night).
>>> Guest boots without any issue. Please take look at attached logs.
>>>
>>> Doug, could you tell me how exactly did you test your machine? I need OS
>>> type, version, C version (GCC, clang, anything else), bintuils version,
>>> etc. "xl dmesg", "xl info" and "dmesg" full outputs are welcome too.
>>>
>>> Daniel
>>>
>>
>> I thought I already responded to Konrad saying that latest staging +
>> relocation patches also comes up.
> 
> I do not remember it. Maybe I have missed that.
> 
>> My guess is that it is related to the IOMMU "fix" that Andrew and Jan
>> did by #if 0'ing out some of ebmalloc. But I'm not sure. I haven't had
> 
> I reenabled free_ebmalloc_unused_mem() during QEMU tests last week.
> It has not changed anything. I will do the same on my NUC.
> 
>> time to look at any of this stuff lately. I went to ELC and then
>> vacation and then managed to hurt myself on vacation so I've been away
>> from my computer a bit.
>>
>> All my branches are available in https://github.com/cardoe/xen and I've
>> been on Ubuntu 16.04.
> 
> I will try this too. Thanks for update.
> 
> Daniel
> 

Where's the branch you are using on that NUC? Because you can't be using
plain staging on there because the firmware version you reported dead
locks when the EFI GetTime() is called. Its a known issue by Intel. So
you must be patching that out of your Xen?

-- 
Doug Goldstein

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel