Re: Can't boot new 4.4 kernel with IOMMU enabled

[Mark Hounschell <markh-n2QNKt385d+sTnJN9+BGXg@public.gmane.org>]

On 01/21/2016 09:50 AM, Mark Hounschell wrote:
> On 01/20/2016 01:26 PM, Joerg Roedel wrote:
>> On Wed, Jan 20, 2016 at 01:19:52PM -0500, Mark Hounschell wrote:
>>>
>>> Would you like me to do anything more right now?
>>
>> Would be great if you find the time to bisect the issue.
>>
>> To speed thing up you can test commit b67ad2f and use it as the bad one
>> if it shows the same problem.
>>
>> As good commit you can use the v4.3 tag, this will only bisect in the
>> iommu related changes which went into v4.4-rc1.
>>
>> If commit b67ad2f works for you, then please bisect between v4.3 and
>> v4.4.
>>
>>
>>     Joerg
>>
>>
>
> Here ya go.
>
> 150952f96941cb13371770ce0cf47906647d8bba is the first bad commit
> commit 150952f96941cb13371770ce0cf47906647d8bba
> Author: Joerg Roedel <jroedel-l3A5Bk7waGM@public.gmane.org>
> Date:   Tue Oct 20 17:33:35 2015 +0200
>
>      iommu/amd: Do not iterate over alias-list in __[attach|detach]_device
>
>      The alias list is handled aleady by iommu core code. No need
>      anymore to handle it in this part of the AMD IOMMU code
>
>      Signed-off-by: Joerg Roedel <jroedel-l3A5Bk7waGM@public.gmane.org>
>
> :040000 040000 334ffdc0d7ff9389478d3ce6096eb81227d8eafc
> cb6fce5d7e94ce3e44a6a64092ba8eb32dd0978b M      drivers
>

FYI, this commit also had tons of IO_PAGE_FAULTS booting but eventually 
all the disks but the 2 in question actually ended up being mounted.

Mark