xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
To: JBeulich@suse.com, ian.campbell@citrix.com,
	ian.jackson@eu.citrix.com, xen-devel@lists.xenproject.org,
	keir@xen.org, tim@xen.org
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Subject: [PATCH v6 for-xen-4.5 2/3] dpci: In hvm_dirq_assist stop using pt_pirq_iterate
Date: Mon, 22 Sep 2014 22:10:14 -0400	[thread overview]
Message-ID: <1411438215-8146-3-git-send-email-konrad.wilk@oracle.com> (raw)
In-Reply-To: <1411438215-8146-1-git-send-email-konrad.wilk@oracle.com>

When an interrupt for an PCI (or PCIe) passthrough device
is to be sent to a guest, we find the appropiate
'hvm_dirq_dpci' structure for the interrupt (PIRQ), set
a bit, and schedule an tasklet.

As we are now called from dpci_softirq with the outstanding
'struct hvm_pirq_dpci' there is no need to call pt_pirq_iterate
which will iterate over all of the PIRQs and call us with every
one that is mapped. And then _hvm_dirq_assist figuring out
which one to execute.

This is a inefficient and not fair as:
 - We iterate starting at PIRQ 0 and up every time. That means
   the PCIe devices that have lower PIRQs get to be called
   first.
 - If we have many PCIe devices passed in with many PIRQs and
   most of the time only the highest numbered PIRQ gets an
   interrupt - we iterate over many PIRQs.

Since we know which 'hvm_pirq_dpci' to check - as the tasklet is
called for a specific 'struct hvm_pirq_dpci' - we do that
in this patch.

Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
---
 xen/drivers/passthrough/io.c | 28 +++++++++++-----------------
 1 file changed, 11 insertions(+), 17 deletions(-)

diff --git a/xen/drivers/passthrough/io.c b/xen/drivers/passthrough/io.c
index 59b5c09..d2a1214 100644
--- a/xen/drivers/passthrough/io.c
+++ b/xen/drivers/passthrough/io.c
@@ -534,9 +534,14 @@ void hvm_dpci_msi_eoi(struct domain *d, int vector)
     spin_unlock(&d->event_lock);
 }
 
-static int _hvm_dirq_assist(struct domain *d, struct hvm_pirq_dpci *pirq_dpci,
-                            void *arg)
+static void hvm_dirq_assist(unsigned long _d)
 {
+    struct hvm_pirq_dpci *pirq_dpci = (struct hvm_pirq_dpci *)_d;
+    struct domain *d = pirq_dpci->dom;
+
+    ASSERT(d->arch.hvm_domain.irq.dpci);
+
+    spin_lock(&d->event_lock);
     if ( test_and_clear_bool(pirq_dpci->masked) )
     {
         struct pirq *pirq = dpci_pirq(pirq_dpci);
@@ -547,13 +552,13 @@ static int _hvm_dirq_assist(struct domain *d, struct hvm_pirq_dpci *pirq_dpci,
             send_guest_pirq(d, pirq);
 
             if ( pirq_dpci->flags & HVM_IRQ_DPCI_GUEST_MSI )
-                return 0;
+                goto out;
         }
 
         if ( pirq_dpci->flags & HVM_IRQ_DPCI_GUEST_MSI )
         {
             vmsi_deliver_pirq(d, pirq_dpci);
-            return 0;
+            goto out;
         }
 
         list_for_each_entry ( digl, &pirq_dpci->digl_list, list )
@@ -566,7 +571,7 @@ static int _hvm_dirq_assist(struct domain *d, struct hvm_pirq_dpci *pirq_dpci,
         {
             /* for translated MSI to INTx interrupt, eoi as early as possible */
             __msi_pirq_eoi(pirq_dpci);
-            return 0;
+            goto out;
         }
 
         /*
@@ -579,18 +584,7 @@ static int _hvm_dirq_assist(struct domain *d, struct hvm_pirq_dpci *pirq_dpci,
         ASSERT(pt_irq_need_timer(pirq_dpci->flags));
         set_timer(&pirq_dpci->timer, NOW() + PT_IRQ_TIME_OUT);
     }
-
-    return 0;
-}
-
-static void hvm_dirq_assist(unsigned long _d)
-{
-    struct domain *d = ((struct hvm_pirq_dpci *)_d)->dom;
-
-    ASSERT(d->arch.hvm_domain.irq.dpci);
-
-    spin_lock(&d->event_lock);
-    pt_pirq_iterate(d, _hvm_dirq_assist, NULL);
+ out:
     spin_unlock(&d->event_lock);
 }
 
-- 
1.9.3

  parent reply	other threads:[~2014-09-23  2:10 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-09-23  2:10 [PATCH v6] Fix interrupt latency of HVM PCI passthrough devices Konrad Rzeszutek Wilk
2014-09-23  2:10 ` [PATCH v6 for-xen-4.5 1/3] dpci: Move from domain centric model to hvm_dirq_dpci model Konrad Rzeszutek Wilk
2014-09-25 14:24   ` Jan Beulich
2014-09-25 14:48     ` Konrad Rzeszutek Wilk
2014-09-25 15:04       ` Jan Beulich
2014-09-27  1:32         ` Konrad Rzeszutek Wilk
2014-09-29  7:21           ` Jan Beulich
2014-10-07 15:40             ` Konrad Rzeszutek Wilk
2014-10-07 16:10               ` Jan Beulich
2014-09-23  2:10 ` Konrad Rzeszutek Wilk [this message]
2014-09-25 14:29   ` [PATCH v6 for-xen-4.5 2/3] dpci: In hvm_dirq_assist stop using pt_pirq_iterate Jan Beulich
2014-09-23  2:10 ` [PATCH v6 for-xen-4.5 3/3] dpci: Replace tasklet with an softirq (v6) Konrad Rzeszutek Wilk
2014-09-25 14:55   ` Jan Beulich
2014-09-25 15:27     ` Konrad Rzeszutek Wilk
2014-09-25 15:45       ` Jan Beulich
2014-09-25 16:05         ` Konrad Rzeszutek Wilk
2014-09-27  1:32         ` Konrad Rzeszutek Wilk

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=1411438215-8146-3-git-send-email-konrad.wilk@oracle.com \
    --to=konrad.wilk@oracle.com \
    --cc=JBeulich@suse.com \
    --cc=ian.campbell@citrix.com \
    --cc=ian.jackson@eu.citrix.com \
    --cc=keir@xen.org \
    --cc=tim@xen.org \
    --cc=xen-devel@lists.xenproject.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).