From: Li Zhong <zhong@linux.vnet.ibm.com>
To: kvm-ppc@vger.kernel.org,
PowerPC email list <linuxppc-dev@lists.ozlabs.org>
Cc: agraf@suse.com, Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Paul Mackerras <paulus@samba.org>,
Michael Ellerman <mpe@ellerman.id.au>,
Boqun Feng <boqun.feng@linux.vnet.ibm.com>
Subject: [RFC PATCH 2/2] KVM: PPC: Don't take lock when check irq's resend flag
Date: Mon, 16 May 2016 15:02:13 +0800 [thread overview]
Message-ID: <1463382133.19947.10.camel@TP420> (raw)
In-Reply-To: <1463381898.19947.6.camel@TP420>
It seems that we don't need to take the lock before evaluating irq's
resend flag. What needed is to make sure when we clear the ics's bit
in the icp's resend_map, we don't miss the resend flag of the irqs
that set the bit.
And seems this could be ordered through the barrier in test_and_clear_bit(),
and an newly added wmb when setting irq's resend flag, and icp's resend_map.
Signed-off-by: Li Zhong <zhong@linux.vnet.ibm.com>
---
arch/powerpc/kvm/book3s_hv_rm_xics.c | 14 +++++++-------
arch/powerpc/kvm/book3s_xics.c | 22 +++++++---------------
2 files changed, 14 insertions(+), 22 deletions(-)
diff --git a/arch/powerpc/kvm/book3s_hv_rm_xics.c b/arch/powerpc/kvm/book3s_hv_rm_xics.c
index c3f604e..c3fa386 100644
--- a/arch/powerpc/kvm/book3s_hv_rm_xics.c
+++ b/arch/powerpc/kvm/book3s_hv_rm_xics.c
@@ -39,14 +39,9 @@ static void ics_rm_check_resend(struct kvmppc_xics *xics,
for (i = 0; i < KVMPPC_XICS_IRQ_PER_ICS; i++) {
struct ics_irq_state *state = &ics->irq_state[i];
- arch_spin_lock(&state->lock);
-
- if (!state->resend) {
- arch_spin_unlock(&state->lock);
+ if (!state->resend)
continue;
- }
- arch_spin_unlock(&state->lock);
icp_rm_deliver_irq(xics, icp, state->number);
}
}
@@ -374,8 +369,13 @@ static void icp_rm_deliver_irq(struct kvmppc_xics *xics, struct kvmppc_icp *icp,
* We failed to deliver the interrupt we need to set the
* resend map bit and mark the ICS state as needing a resend
*/
- set_bit(ics->icsid, icp->resend_map);
state->resend = 1;
+ /*
+ * Make sure when checking resend, we don't miss the resend if
+ * resend_map bit is seen and cleared.
+ */
+ smp_wmb();
+ set_bit(ics->icsid, icp->resend_map);
/*
* If the need_resend flag got cleared in the ICP some time
diff --git a/arch/powerpc/kvm/book3s_xics.c b/arch/powerpc/kvm/book3s_xics.c
index b9dbf75..420c4fc 100644
--- a/arch/powerpc/kvm/book3s_xics.c
+++ b/arch/powerpc/kvm/book3s_xics.c
@@ -110,30 +110,17 @@ static void ics_check_resend(struct kvmppc_xics *xics, struct kvmppc_ics *ics,
{
int i;
- unsigned long flags;
-
- local_irq_save(flags);
-
for (i = 0; i < KVMPPC_XICS_IRQ_PER_ICS; i++) {
struct ics_irq_state *state = &ics->irq_state[i];
- arch_spin_lock(&state->lock);
-
- if (!state->resend) {
- arch_spin_unlock(&state->lock);
+ if (!state->resend)
continue;
- }
XICS_DBG("resend %#x prio %#x\n", state->number,
state->priority);
- arch_spin_unlock(&state->lock);
- local_irq_restore(flags);
icp_deliver_irq(xics, icp, state->number);
- local_irq_save(flags);
}
-
- local_irq_restore(flags);
}
static bool write_xive(struct kvmppc_xics *xics, struct kvmppc_ics *ics,
@@ -474,8 +461,13 @@ static void icp_deliver_irq(struct kvmppc_xics *xics, struct kvmppc_icp *icp,
* We failed to deliver the interrupt we need to set the
* resend map bit and mark the ICS state as needing a resend
*/
- set_bit(ics->icsid, icp->resend_map);
state->resend = 1;
+ /*
+ * Make sure when checking resend, we don't miss the resend if
+ * resend_map bit is seen and cleared.
+ */
+ smp_wmb();
+ set_bit(ics->icsid, icp->resend_map);
/*
* If the need_resend flag got cleared in the ICP some time
--
1.8.3.1
next prev parent reply other threads:[~2016-05-16 7:11 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-16 6:58 [RFC PATCH 1/2] KVM: PPC: divide the ics lock into smaller ones for each irq Li Zhong
2016-05-16 7:02 ` Li Zhong [this message]
2016-06-20 5:27 ` [RFC PATCH 2/2] KVM: PPC: Don't take lock when check irq's resend flag Paul Mackerras
2016-06-21 2:57 ` Li Zhong
2016-06-20 5:25 ` [RFC PATCH 1/2] KVM: PPC: divide the ics lock into smaller ones for each irq Paul Mackerras
2016-06-21 2:47 ` Li Zhong
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=1463382133.19947.10.camel@TP420 \
--to=zhong@linux.vnet.ibm.com \
--cc=agraf@suse.com \
--cc=benh@kernel.crashing.org \
--cc=boqun.feng@linux.vnet.ibm.com \
--cc=kvm-ppc@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=mpe@ellerman.id.au \
--cc=paulus@samba.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).