From: Isaku Yamahata <yamahata@valinux.co.jp>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: qemu-devel@nongnu.org
Subject: [Qemu-devel] Re: [PATCH v3 3/3] piix_pci: optimize set irq path
Date: Mon, 21 Mar 2011 21:56:56 +0900 [thread overview]
Message-ID: <20110321125656.GC9998@valinux.co.jp> (raw)
In-Reply-To: <20110321123111.GA29012@redhat.com>
On Mon, Mar 21, 2011 at 02:31:11PM +0200, Michael S. Tsirkin wrote:
> On Mon, Mar 21, 2011 at 09:10:32PM +0900, Isaku Yamahata wrote:
> > On Mon, Mar 21, 2011 at 01:37:07PM +0200, Michael S. Tsirkin wrote:
> > > > +static int piix3_post_load(void *opaque, int version_id)
> > > > +{
> > > > + PIIX3State *piix3 = opaque;
> > > > + piix3_update_irq_levels(piix3);
> > >
> > > Couldn't figure out why would we not want to
> > > propagate the interrupts here.
> > > Could you explain please?
> > > What happens if we do propagate them?
> > > Nothing bad, right?
> >
> > I wanted to be just conservative.
> > If you are brave enough to change the behavior, I'm fine with propagating
> > interrupts.
> >
> > If we propagate the interrupts, guest OS may see interrupts
> > unnecessarily/spuriously injected after load.
> > Probably such interrupts doesn't harm OSes, so there is nothing
> > bad in theory as you said.
> > On the other hand, I hesitated to change the existing behavior because
> > it would be very difficult to debug it and to test many OSes.
>
> I expect it won't change the behaviour because the interrupts
> are level: at the moment e.g. pci devices already reassert
> interrupts on load.
>
> But I agree it better be a separate patch, and needs a lot of testing.
Like this?
diff --git a/hw/piix_pci.c b/hw/piix_pci.c
index f07e19d..8052c1e 100644
--- a/hw/piix_pci.c
+++ b/hw/piix_pci.c
@@ -280,8 +280,7 @@ static void piix3_set_irq_pic(PIIX3State *piix3, int pic_irq)
((PIIX_NUM_PIRQS - 1) << (pic_irq * PIIX_NUM_PIRQS))));
}
-static void piix3_set_irq_level(PIIX3State *piix3, int irq_num, int level,
- bool propagate)
+static void piix3_set_irq_level(PIIX3State *piix3, int irq_num, int level)
{
int pic_irq;
uint64_t mask;
@@ -295,15 +294,13 @@ static void piix3_set_irq_level(PIIX3State *piix3, int irq_num, int level,
piix3->pic_levels &= ~mask;
piix3->pic_levels |= mask * !!level;
- if (propagate) {
- piix3_set_irq_pic(piix3, pic_irq);
- }
+ piix3_set_irq_pic(piix3, pic_irq);
}
static void piix3_set_irq(void *opaque, int irq_num, int level)
{
PIIX3State *piix3 = opaque;
- piix3_set_irq_level(piix3, irq_num, level, true);
+ piix3_set_irq_level(piix3, irq_num, level);
}
/* irq routing is changed. so rebuild bitmap */
@@ -314,8 +311,7 @@ static void piix3_update_irq_levels(PIIX3State *piix3)
piix3->pic_levels = 0;
for (pirq = 0; pirq < PIIX_NUM_PIRQS; pirq++) {
piix3_set_irq_level(piix3, pirq,
- pci_bus_get_irq_level(piix3->dev.bus, pirq),
- false);
+ pci_bus_get_irq_level(piix3->dev.bus, pirq));
}
}
--
yamahata
next prev parent reply other threads:[~2011-03-21 12:57 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-03-19 13:24 [Qemu-devel] [PATCH v3 0/3] piix_pci: optimize irq data path Isaku Yamahata
2011-03-19 13:24 ` [Qemu-devel] [PATCH v3 1/3] pci: add accessor function to get irq levels Isaku Yamahata
2011-03-19 13:24 ` [Qemu-devel] [PATCH v3 2/3] piix_pci: eliminate PIIX3State::pci_irq_levels Isaku Yamahata
2011-03-19 13:24 ` [Qemu-devel] [PATCH v3 3/3] piix_pci: optimize set irq path Isaku Yamahata
2011-03-21 11:37 ` [Qemu-devel] " Michael S. Tsirkin
2011-03-21 12:10 ` Isaku Yamahata
2011-03-21 12:31 ` Michael S. Tsirkin
2011-03-21 12:56 ` Isaku Yamahata [this message]
2011-03-21 13:01 ` Michael S. Tsirkin
2011-03-21 12:26 ` Isaku Yamahata
2011-03-21 14:10 ` Michael S. Tsirkin
2011-03-22 0:50 ` Isaku Yamahata
2011-03-22 13:40 ` Michael S. Tsirkin
2011-03-22 14:07 ` Isaku Yamahata
2011-03-22 16:24 ` Michael S. Tsirkin
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=20110321125656.GC9998@valinux.co.jp \
--to=yamahata@valinux.co.jp \
--cc=mst@redhat.com \
--cc=qemu-devel@nongnu.org \
/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;
as well as URLs for NNTP newsgroup(s).