From: Vineet.Gupta1@synopsys.com (Vineet Gupta)
To: linux-snps-arc@lists.infradead.org
Subject: [PATCH 1/5] ARCv2: SMP: Emulate IPI to self using software triggered interrupt
Date: Tue, 23 Feb 2016 14:41:38 +0530 [thread overview]
Message-ID: <1456218702-11911-2-git-send-email-vgupta@synopsys.com> (raw)
In-Reply-To: <1456218702-11911-1-git-send-email-vgupta@synopsys.com>
ARConnect/MCIP Inter-Core-Interrupt module can't send interrupt to
local core. So use core intc capability to trigger software
interrupt to self, using an unsued IRQ #21.
This showed up as csd deadlock with LTP trace_sched on a dual core
system. This test acts as scheduler fuzzer, triggering all sorts of
schedulting activity. Trouble starts with IPI to self, which doesn't get
delivered (effectively lost due to H/w capability), but the msg intended
to be sent remain enqueued in per-cpu @ipi_data.
All subsequent IPIs to this core from other cores get elided due to the
IPI coalescing optimization in ipi_send_msg_one() where a pending msg
implies an IPI already sent and assumes other core is yet to ack it.
After the elided IPI, other core simply goes into csd_lock_wait()
but never comes out as this core never sees the interrupt.
Fixes STAR 9001008624
Cc: Peter Zijlstra <peterz at infradead.org>
Cc: <stable at vger.kernel.org> [4.2]
Signed-off-by: Vineet Gupta <vgupta at synopsys.com>
---
arch/arc/include/asm/irqflags-arcv2.h | 11 +++++++++++
arch/arc/kernel/mcip.c | 15 +++++++++++++++
2 files changed, 26 insertions(+)
diff --git a/arch/arc/include/asm/irqflags-arcv2.h b/arch/arc/include/asm/irqflags-arcv2.h
index 1fc18ee06cf2..37c2f751eebf 100644
--- a/arch/arc/include/asm/irqflags-arcv2.h
+++ b/arch/arc/include/asm/irqflags-arcv2.h
@@ -22,6 +22,7 @@
#define AUX_IRQ_CTRL 0x00E
#define AUX_IRQ_ACT 0x043 /* Active Intr across all levels */
#define AUX_IRQ_LVL_PEND 0x200 /* Pending Intr across all levels */
+#define AUX_IRQ_HINT 0x201 /* For generating Soft Interrupts */
#define AUX_IRQ_PRIORITY 0x206
#define ICAUSE 0x40a
#define AUX_IRQ_SELECT 0x40b
@@ -115,6 +116,16 @@ static inline int arch_irqs_disabled(void)
return arch_irqs_disabled_flags(arch_local_save_flags());
}
+static inline void arc_softirq_trigger(int irq)
+{
+ write_aux_reg(AUX_IRQ_HINT, irq);
+}
+
+static inline void arc_softirq_clear(int irq)
+{
+ write_aux_reg(AUX_IRQ_HINT, 0);
+}
+
#else
.macro IRQ_DISABLE scratch
diff --git a/arch/arc/kernel/mcip.c b/arch/arc/kernel/mcip.c
index 002c5fcf8947..9e1bd03b87a6 100644
--- a/arch/arc/kernel/mcip.c
+++ b/arch/arc/kernel/mcip.c
@@ -11,9 +11,12 @@
#include <linux/smp.h>
#include <linux/irq.h>
#include <linux/spinlock.h>
+#include <asm/irqflags-arcv2.h>
#include <asm/mcip.h>
#include <asm/setup.h>
+#define SOFTIRQ_IRQ 21
+
static char smp_cpuinfo_buf[128];
static int idu_detected;
@@ -22,6 +25,7 @@ static DEFINE_RAW_SPINLOCK(mcip_lock);
static void mcip_setup_per_cpu(int cpu)
{
smp_ipi_irq_setup(cpu, IPI_IRQ);
+ smp_ipi_irq_setup(cpu, SOFTIRQ_IRQ);
}
static void mcip_ipi_send(int cpu)
@@ -29,6 +33,12 @@ static void mcip_ipi_send(int cpu)
unsigned long flags;
int ipi_was_pending;
+ /* ARConnect can only send IPI to others */
+ if (unlikely(cpu == raw_smp_processor_id())) {
+ arc_softirq_trigger(SOFTIRQ_IRQ);
+ return;
+ }
+
/*
* NOTE: We must spin here if the other cpu hasn't yet
* serviced a previous message. This can burn lots
@@ -63,6 +73,11 @@ static void mcip_ipi_clear(int irq)
unsigned long flags;
unsigned int __maybe_unused copy;
+ if (unlikely(irq == SOFTIRQ_IRQ)) {
+ arc_softirq_clear(irq);
+ return;
+ }
+
raw_spin_lock_irqsave(&mcip_lock, flags);
/* Who sent the IPI */
--
2.5.0
next prev parent reply other threads:[~2016-02-23 9:11 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-23 9:11 [PATCH 0/5] ARC IPI related fix Vineet Gupta
2016-02-23 9:11 ` Vineet Gupta [this message]
2016-02-23 9:11 ` [PATCH 2/5] ARC: [intc-compact] Remove IPI setup from ARCompact port Vineet Gupta
2016-02-23 9:11 ` [PATCH 3/5] ARCv2: SMP: Push IPI_IRQ into IPI provider Vineet Gupta
2016-02-23 9:11 ` [PATCH 4/5] ARCv2: Elide sending new cross core intr if receiver didn't ack prev Vineet Gupta
2016-02-26 11:02 ` Vineet Gupta
2016-02-23 9:11 ` [PATCH 5/5] ARC: SMP: No need for CONFIG_ARC_IPI_DBG Vineet Gupta
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=1456218702-11911-2-git-send-email-vgupta@synopsys.com \
--to=vineet.gupta1@synopsys.com \
--cc=linux-snps-arc@lists.infradead.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).