public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Jan Kiszka <jan.kiszka@siemens.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Avi Kivity <avi@redhat.com>,
	Marcelo Tosatti <mtosatti@redhat.com>, kvm <kvm@vger.kernel.org>,
	Eric Northup <digitaleric@google.com>
Subject: Re: [PATCH v4] KVM: Introduce direct MSI message injection for in-kernel irqchips
Date: Wed, 04 Apr 2012 11:22:43 +0200	[thread overview]
Message-ID: <4F7C12E3.3050702@siemens.com> (raw)
In-Reply-To: <20120404085359.GA3404@redhat.com>

On 2012-04-04 10:53, Michael S. Tsirkin wrote:
> On Wed, Apr 04, 2012 at 11:44:23AM +0300, Avi Kivity wrote:
>> On 04/04/2012 11:38 AM, Michael S. Tsirkin wrote:
>>>>
>>>>>
>>>>> A performance note: delivering an interrupt needs to search all vcpus
>>>>> for an APIC ID match.  The previous plan was to cache (or pre-calculate)
>>>>> this lookup in the irq routing table.  Now it looks like we'll need a
>>>>> separate cache for this.
>>>>
>>>> As this is non-existent until today, we don't regress here. And it can
>>>> still be added on top later on, transparently.
>>>
>>> I always worry about hash collisions and the cost of
>>> calculating good hash functions.
>>>
>>> We could instead return an index in the cache on injection, maintain in
>>> userspace and use it for fast path on the next injection.
>>
>> Ahem, that is almost the existing routing table to a T.
>>
>>> Will make it easy to use an array index instead of a hash here,
>>> and fallback to a slower ID lookup on mismatch.
>>
>> Need a free ioctl so we can reuse IDs.
> 
> No, it could be kernel controlled not userspace controlled. We get both
> and address and an index:
> 
> if (table[u.i].addr == u.addr && table[u.i].data == u.data) {
> 	return table[u.i].id;
> }
> 
> u.i = find_lru_idx(&table);
> table[u.i].addr = u.addr;
> table[u.i].data = u.data;
> table[u.i].id = find_id(u.addr, u.data);
> return table[u.i].id;
> 
> 
>>> Until we do have this fast path we can just fill this value with zeros,
>>> so kernel patch (almost) does not need to change for this -
>>> just the header.
>>
>> Partially implemented interfaces invite breakage.
> 
> Hmm true. OK scrap this idea then, it's not clear
> whether we are going to optimize this anyway.
> 

Also, the problem is that keeping that ID in userspace requires an
infrastructure like the MSIRoutingCache that I proposed originally. Not
much won /wrt invasiveness there. So we should really do the routing
optimization in the kernel - one day.

Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux

  reply	other threads:[~2012-04-04  9:22 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-03-28 17:47 [PATCH v2] KVM: Introduce direct MSI message injection for in-kernel irqchips Jan Kiszka
2012-03-28 17:52 ` Jan Kiszka
2012-03-28 19:58 ` Eric Northup
2012-03-28 20:21   ` Jan Kiszka
2012-03-29 15:39 ` Michael S. Tsirkin
2012-03-29 15:43   ` Jan Kiszka
2012-03-29 16:15 ` [PATCH v3] " Jan Kiszka
2012-03-29 16:46   ` Michael S. Tsirkin
2012-03-29 16:50     ` Jan Kiszka
2012-03-29 18:25   ` Jan Kiszka
2012-03-29 19:14   ` [PATCH v4] " Jan Kiszka
2012-03-29 19:41     ` Michael S. Tsirkin
2012-03-30  7:45       ` Jan Kiszka
2012-03-30 12:45         ` Michael S. Tsirkin
2012-04-03 16:27     ` Avi Kivity
2012-04-03 16:47       ` Jan Kiszka
2012-04-03 16:54         ` Avi Kivity
2012-04-03 17:24           ` Jan Kiszka
2012-04-04  8:47             ` Avi Kivity
2012-04-04  8:38         ` Michael S. Tsirkin
2012-04-04  8:44           ` Avi Kivity
2012-04-04  8:53             ` Michael S. Tsirkin
2012-04-04  9:22               ` Jan Kiszka [this message]
2012-04-04  9:36                 ` Avi Kivity
2012-04-04  9:38                   ` Jan Kiszka
2012-04-04  9:55                     ` Avi Kivity
2012-04-04 10:48                       ` Jan Kiszka
2012-04-04 11:50                         ` Avi Kivity
2012-04-04 12:01                           ` Jan Kiszka
2012-04-10 18:30       ` [PATCH] KVM: Introduce generic interrupt " Jan Kiszka
2012-04-23 14:44         ` Jan Kiszka
2012-04-23 15:17           ` Avi Kivity
2012-04-23 15:32         ` Avi Kivity
2012-04-23 15:55           ` Jan Kiszka
2012-04-24 11:54             ` Avi Kivity
2012-04-24 11:57     ` [PATCH v4] KVM: Introduce direct MSI message " Avi Kivity
2012-04-24 12:07       ` Jan Kiszka
2012-04-24 12:59         ` Avi Kivity
2012-04-24 13:24           ` Jan Kiszka
2012-04-11 22:10   ` [PATCH v3] " Marcelo Tosatti
2012-04-12  9:28     ` Jan Kiszka
2012-04-12 22:38       ` Marcelo Tosatti
2012-04-13 13:45         ` Jan Kiszka

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4F7C12E3.3050702@siemens.com \
    --to=jan.kiszka@siemens.com \
    --cc=avi@redhat.com \
    --cc=digitaleric@google.com \
    --cc=kvm@vger.kernel.org \
    --cc=mst@redhat.com \
    --cc=mtosatti@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox