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]
Date:   Thu, 7 Sep 2017 16:25:39 -0700
From:   Paul Burton <paul.burton@...tec.com>
To:     Thomas Gleixner <tglx@...utronix.de>,
        Ralf Baechle <ralf@...ux-mips.org>
CC:     <dianders@...omium.org>, James Hogan <james.hogan@...tec.com>,
        Brian Norris <briannorris@...omium.org>,
        Jason Cooper <jason@...edaemon.net>,
        <jeffy.chen@...k-chips.com>, Marc Zyngier <marc.zyngier@....com>,
        <linux-kernel@...r.kernel.org>, <linux-mips@...ux-mips.org>,
        <tfiga@...omium.org>, Paul Burton <paul.burton@...tec.com>
Subject: [RFC PATCH v1 6/9] MIPS: perf: percpu_devid interrupt support

The MIPS CPU performance counter overflow interrupt is really a percpu
interrupt, but up until now we have not used the percpu interrupt APIs
to configure & control it. In preparation for doing so, introduce
support for percpu_devid interrupts in the MIPS perf implementation.

We switch from using request_irq() to using either setup_irq() or
setup_percpu_irq() with an explicit struct irqaction such that we can
set the flags, handler & name for that struct irqaction once rather than
needing to duplicate them in calls to request_irq() and
request_percpu_irq().

The IRQF_NOAUTOEN flag is passed because percpu interrupts
automatically get IRQ_NOAUTOEN set by irq_set_percpu_devid_flags(). We
opt into accepting this behaviour & explicitly enable the interrupt in
mipspmu_enable() right after configuring the local performance counters.

Signed-off-by: Paul Burton <paul.burton@...tec.com>
Cc: James Hogan <james.hogan@...tec.com>
Cc: Jason Cooper <jason@...edaemon.net>
Cc: Marc Zyngier <marc.zyngier@....com>
Cc: Ralf Baechle <ralf@...ux-mips.org>
Cc: Thomas Gleixner <tglx@...utronix.de>
Cc: linux-kernel@...r.kernel.org
Cc: linux-mips@...ux-mips.org
---

 arch/mips/kernel/perf_event_mipsxx.c | 30 +++++++++++++++++++-----------
 1 file changed, 19 insertions(+), 11 deletions(-)

diff --git a/arch/mips/kernel/perf_event_mipsxx.c b/arch/mips/kernel/perf_event_mipsxx.c
index cae36ca400e9..af7bae79dc51 100644
--- a/arch/mips/kernel/perf_event_mipsxx.c
+++ b/arch/mips/kernel/perf_event_mipsxx.c
@@ -514,6 +514,11 @@ static void mipspmu_enable(struct pmu *pmu)
 	write_unlock(&pmuint_rwlock);
 #endif
 	resume_local_counters();
+
+	if (irq_is_percpu_devid(mipspmu.irq))
+		enable_percpu_irq(mipspmu.irq, IRQ_TYPE_NONE);
+	else
+		enable_irq(mipspmu.irq);
 }
 
 /*
@@ -538,24 +543,27 @@ static void mipspmu_disable(struct pmu *pmu)
 static atomic_t active_events = ATOMIC_INIT(0);
 static DEFINE_MUTEX(pmu_reserve_mutex);
 
+static struct irqaction c0_perf_irqaction = {
+	.handler = mipsxx_pmu_handle_irq,
+	.flags = IRQF_PERCPU | IRQF_TIMER | IRQF_SHARED | IRQF_NOAUTOEN,
+	.name = "mips_perf_pmu",
+	.percpu_dev_id = &mipspmu,
+};
+
 static int mipspmu_get_irq(void)
 {
-	int err;
+	if (irq_is_percpu_devid(mipspmu.irq))
+		return setup_percpu_irq(mipspmu.irq, &c0_perf_irqaction);
 
-	err = request_irq(mipspmu.irq, mipsxx_pmu_handle_irq,
-			  IRQF_PERCPU | IRQF_NOBALANCING |
-			  IRQF_NO_THREAD | IRQF_NO_SUSPEND |
-			  IRQF_SHARED,
-			  "mips_perf_pmu", &mipspmu);
-	if (err)
-		pr_warn("Unable to request IRQ%d for MIPS performance counters!\n",
-			mipspmu.irq);
-	return err;
+	return setup_irq(mipspmu.irq, &c0_perf_irqaction);
 }
 
 static void mipspmu_free_irq(void)
 {
-	free_irq(mipspmu.irq, &mipspmu);
+	if (irq_is_percpu_devid(mipspmu.irq))
+		remove_percpu_irq(mipspmu.irq, &c0_perf_irqaction);
+	else
+		remove_irq(mipspmu.irq, &c0_perf_irqaction);
 }
 
 /*
-- 
2.14.1

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ