[vfio-users] Far Cry Primal

Tue Jun 14 17:25:16 UTC 2016

Overwatch doesn't have this problem. It is only sc2 and Heroes

On Tue, Jun 14, 2016, 12:25 Jayme Howard <g.prime at gmail.com> wrote:

> Actually, I also haven't seen further ones while I'm playing Overwatch.
> Just at startup.  I wonder if maybe I've got something misconfigured?
>
> On Tue, Jun 14, 2016 at 11:23 AM, Abdulla Bubshait <darkstego at gmail.com>
> wrote:
>
>> Thanks, I guess the "issue" is only from the blizzard games. It could
>> very well be that this is used as some anti-tamper tool in these blizzard
>> games, but I find it odd that it only does this in Win10 and that returning
>> 0 doesn't cause a problem.
>>
>> On Tue, Jun 14, 2016 at 12:06 PM Jayme Howard <g.prime at gmail.com> wrote:
>>
>>> I've only seen it call them at startup.  After playing for an hour or
>>> so, I didn't see more of them in my dmesg output.
>>>
>>> On Tue, Jun 14, 2016 at 9:38 AM, Abdulla Bubshait <darkstego at gmail.com>
>>> wrote:
>>>
>>>> For what its worth it seems Rise of the Tomb Raider also calls it, so
>>>> it might be Denuvo.
>>>>
>>>> The interesting thing is Rise of the Tomb Raider only calls it once. So
>>>> there is no performance issue. Heroes and Starcraft continuously make these
>>>> calls which causes the performance issue.
>>>>
>>>> Can anyone confirm if Doom only calls up the MSRS in one spot or does
>>>> it constantly keep calling the MSRS throughout the game?
>>>>
>>>> On Tue, Jun 14, 2016 at 8:55 AM thibaut noah <thibaut.noah at gmail.com>
>>>> wrote:
>>>>
>>>>> They have the warden which is constantly checking your system
>>>>>
>>>>> 2016-06-14 14:36 GMT+02:00 Jayme Howard <g.prime at gmail.com>:
>>>>>
>>>>>> I don't think Blizzard uses Denuvo.
>>>>>>
>>>>>> On Tue, Jun 14, 2016 at 7:29 AM, thibaut noah <thibaut.noah at gmail.com
>>>>>> > wrote:
>>>>>>
>>>>>>> That might also have to do with denuvo protection
>>>>>>>
>>>>>>> 2016-06-14 9:49 GMT+02:00 Abdulla Bubshait <darkstego at gmail.com>:
>>>>>>>
>>>>>>>> Yes, this is a win10 issue.
>>>>>>>> Changing the Virtual CPU model does not make a difference. Each
>>>>>>>> model has its own LBR register address, but one will be called.
>>>>>>>>
>>>>>>>> The only thing I have not tried is using an AMD processor. AMD CPUs
>>>>>>>> support lbrv which is virtualization of the LBR registers, so it is handled
>>>>>>>> in hardware rather than software.
>>>>>>>> If anyone has an AMD CPU and a win10 VM willing to confirm if their
>>>>>>>> setup does indeed work in these games, that would be great.
>>>>>>>>
>>>>>>>> On Tue, Jun 14, 2016 at 3:39 AM Ivan Volosyuk <
>>>>>>>> ivan.volosyuk at gmail.com> wrote:
>>>>>>>>
>>>>>>>>> 2 things:
>>>>>>>>>  - this msrs are not called from StarCraft on Win8.1.
>>>>>>>>>  - i wonder if forcing virtual CPU model to something older will
>>>>>>>>> disable them in Win10.
>>>>>>>>>
>>>>>>>>> This msrs issue is what holds me off upgrade to W10.
>>>>>>>>>
>>>>>>>>> On Tue, Jun 14, 2016 at 12:39 PM Jayme Howard <g.prime at gmail.com>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> For what it's worth, I believe I was getting them on Overwatch as
>>>>>>>>>> well.  It's not happening with EVERY game I have though.
>>>>>>>>>>
>>>>>>>>>> On Mon, Jun 13, 2016 at 8:28 PM, Abdulla Bubshait <
>>>>>>>>>> darkstego at gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On Mon, Jun 13, 2016 at 6:27 PM, Jayme Howard <g.prime at gmail.com>
>>>>>>>>>>> wrote:
>>>>>>>>>>> > That's the output for Doom.
>>>>>>>>>>> >
>>>>>>>>>>> > [1639445.044855] kvm [9487]: vcpu2 kvm_set_msr_common:
>>>>>>>>>>> MSR_IA32_DEBUGCTLMSR
>>>>>>>>>>> > 0x1, nop
>>>>>>>>>>> > [1639445.044861] kvm [9487]: vcpu2 ignored rdmsr: 0x1c9
>>>>>>>>>>> > [1639445.044862] kvm [9487]: vcpu2 ignored rdmsr: 0x680
>>>>>>>>>>> > [1639445.044863] kvm [9487]: vcpu2 ignored rdmsr: 0x6c0
>>>>>>>>>>> > [1639445.044890] kvm [9487]: vcpu1 kvm_set_msr_common:
>>>>>>>>>>> MSR_IA32_DEBUGCTLMSR
>>>>>>>>>>> > 0x1, nop
>>>>>>>>>>> > [1639445.044895] kvm [9487]: vcpu1 ignored rdmsr: 0x1c9
>>>>>>>>>>> > [1639445.044896] kvm [9487]: vcpu1 ignored rdmsr: 0x680
>>>>>>>>>>> > [1639445.044897] kvm [9487]: vcpu1 ignored rdmsr: 0x6c0
>>>>>>>>>>> > [1639445.044905] kvm [9487]: vcpu1 kvm_set_msr_common:
>>>>>>>>>>> MSR_IA32_DEBUGCTLMSR
>>>>>>>>>>> > 0x1, nop
>>>>>>>>>>> > [1639445.044907] kvm [9487]: vcpu1 ignored rdmsr: 0x1c9
>>>>>>>>>>> > [1639445.044908] kvm [9487]: vcpu1 ignored rdmsr: 0x680
>>>>>>>>>>> > [1639445.044909] kvm [9487]: vcpu1 ignored rdmsr: 0x6c0
>>>>>>>>>>> > [1639445.046195] kvm [9487]: vcpu2 kvm_set_msr_common:
>>>>>>>>>>> MSR_IA32_DEBUGCTLMSR
>>>>>>>>>>> > 0x1, nop
>>>>>>>>>>> > [1639445.046198] kvm [9487]: vcpu2 ignored rdmsr: 0x1c9
>>>>>>>>>>> > [1639445.046204] kvm [9487]: vcpu2 kvm_set_msr_common:
>>>>>>>>>>> MSR_IA32_DEBUGCTLMSR
>>>>>>>>>>> > 0x1, nop
>>>>>>>>>>> > [1639445.155114] kvm [9487]: vcpu0 kvm_set_msr_common:
>>>>>>>>>>> MSR_IA32_DEBUGCTLMSR
>>>>>>>>>>> > 0x1, nop
>>>>>>>>>>> > [1639445.155123] kvm [9487]: vcpu0 kvm_set_msr_common:
>>>>>>>>>>> MSR_IA32_DEBUGCTLMSR
>>>>>>>>>>> > 0x1, nop
>>>>>>>>>>> >
>>>>>>>>>>>
>>>>>>>>>>> Odd, this is another game that is contantly reading the LBR. I
>>>>>>>>>>> initially thought this was part of the StarCraft 2 and Heroes of the Storm
>>>>>>>>>>> code. But now I think this might be part of Windows 10. Something being
>>>>>>>>>>> called by these games is causing a ton of LBR reads. Maybe DX 12?
>>>>>>>>>>>
>>>>>>>>>>> I think if we can find what is causing these reads in the games
>>>>>>>>>>> we might be able to solve it. Because if this is the case we might be
>>>>>>>>>>> getting more and more games with performance problems because of these LBR
>>>>>>>>>>> checks.
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> _______________________________________________
>>>>>>>>>> vfio-users mailing list
>>>>>>>>>> vfio-users at redhat.com
>>>>>>>>>> https://www.redhat.com/mailman/listinfo/vfio-users
>>>>>>>>>>
>>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> vfio-users mailing list
>>>>>>>> vfio-users at redhat.com
>>>>>>>> https://www.redhat.com/mailman/listinfo/vfio-users
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/vfio-users/attachments/20160614/a52708dc/attachment.htm>