linux-kernel - Re: [PATCH ] Documentation/kvm: Update cpuid documentation for steal time and pv eoi

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:	Mon, 26 Aug 2013 17:03:00 +0530
From:	Raghavendra K T <raghavendra.kt@...ux.vnet.ibm.com>
To:	"Michael S. Tsirkin" <mst@...hat.com>, mtosatti@...hat.com
CC:	Gleb Natapov <gleb@...hat.com>,
	Paolo Bonzini <pbonzini@...hat.com>, linux-doc@...r.kernel.org,
	linux-kernel@...r.kernel.org, kvm@...r.kernel.org,
	Rob Landley <rob@...dley.net>
Subject: Re: [PATCH ] Documentation/kvm: Update cpuid documentation for steal
 time and pv eoi

On 08/26/2013 12:37 PM, Michael S. Tsirkin wrote:
> I would change the description to merely say what the CPUID bits
> mean, and what they mean is exactly that an MSR is valid.
> Use KVM_FEATURE_ASYNC_PF as a template.

Thank you for the review.
Changing the doc accordingly by adding msr info. Please refer below.

>> +KVM_FEATURE_STEAL_TIME             ||     5 || guest accounts fine granularity
>> +                                   ||       || task steal time.
>
> I'm not sure what this phrase means.
> Steal time is a host feature, not a guest feature:
> IIUC if this bit is set, the hypervisor can pass the guest information
> about how much time was spent running other processes outside the VM.

Okay. I guess I need some help here.

I took this from the PARAVIRT_TIME_ACCOUNTING config help. also I saw
that guest is  actually returning the steal time in kvm_steal_clock().

>
>> enabled when
>> +                                   ||       || shedstat or task delay accounting
>> +                                   ||       || is supported by the host.
>
> I think it's enabled by guest, not by host.

true. My understanding was, Guest enables it when host has schedstat or
task delay accounting  on.

I referred to this hunk in kvm/cpuid.c

if (sched_info_on())
    entry->eax |= (1 << KVM_FEATURE_STEAL_TIME);
and sched_info_on() is true when schedstat or task delay accounting is
on.

Does this look good?

"Enabled by writing to msr 0x4b564d03. The feature is
enabled by guest when host has schedstat or task delay accounting
support."

>> +KVM_FEATURE_PV_EOI                 ||     6 || overrides the generic EOI
>> +                                   ||       || implementation with an optimized
>> +                                   ||       || version.
>
> More exactly "with a paravirtualized version".

Okay. So how does this sound?

"overrides the generic EOI implementation with a paravirtualized
version. This feature is enabled by writing to msr  0x4b564d04."

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/