linux-kernel - Re: [PATCH v5 04/12] Provide special async page fault handler when async PF capability is detected

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

Message-ID: <4C72A05D.70603@redhat.com>
Date:	Mon, 23 Aug 2010 19:22:53 +0300
From:	Avi Kivity <avi@...hat.com>
To:	Rik van Riel <riel@...hat.com>
CC:	Gleb Natapov <gleb@...hat.com>, kvm@...r.kernel.org,
	linux-mm@...ck.org, linux-kernel@...r.kernel.org, mingo@...e.hu,
	a.p.zijlstra@...llo.nl, tglx@...utronix.de, hpa@...or.com,
	cl@...ux-foundation.org, mtosatti@...hat.com
Subject: Re: [PATCH v5 04/12] Provide special async page fault handler when
 async PF capability is detected

  On 08/23/2010 06:52 PM, Rik van Riel wrote:
> On 08/23/2010 11:48 AM, Avi Kivity wrote:
>
>> Do you need to match cpu here as well? Or is token globally unique?
>>
>> Perhaps we should make it locally unique to remove a requirement from
>> the host to synchronize? I haven't seen how you generate it yet.
>
> If a task goes to sleep on one VCPU, but that VCPU ends
> up not being runnable later on, it would be nice to wake
> the task up on on a different VCPU.
>
> I do not remember why it is safe to send this wakeup
> event as an exception rather than an interrupt...

Wakeup could definitely be an interrupt, but the apf needs to be an 
exception so we reuse it.

>
>> The other cpu might be waiting for us to yield. We can fix it later with
>> the the pv spinlock infrastructure.
>>
>> Or, we can avoid the allocation. If at most one apf can be pending (is
>> this true?), we can use a per-cpu variable for this dummy entry.
>
> Having a limit of just one APF pending kind of defeats
> the point.

Yes.  How about, one APF pending before it is seen by the guest - but 
how can we tell without an annoying xchg?

>
> At that point, a second one of these faults would put
> the VCPU to sleep, which prevents the first task from
> running once its pagefault (which started earlier)
> completes...
>


-- 
error compiling committee.c: too many arguments to function

--
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/