Re: [PATCH] delaying interrupts in mips [ 2.6.32]

From: David Daney
Date: Mon Aug 11 2014 - 18:28:33 EST


On 08/11/2014 03:13 PM, Sadasivan Shaiju wrote:
Hi ,

I work for Montavista (Cavium Inc) as a Technical Lead . I want to
push some of the kernel patches to rt community (2.6.32 kernel 2.6.33
rt patch) , so that It will go to the main line These patches are
reviewed and approved by our system Architect. I request you to
include in the main line .

It is already in the "main line":

commit 1bcfecc028686ea32e49b0f4f6e8a665917cb49a
Author: Yong Zhang <yong.zhang@xxxxxxxxxxxxx>
Date: Thu Jul 19 09:13:53 2012 +0200

MIPS: Octeon: delay enable irq to ->smp_finish()

To prepare for smoothing set_cpu_[active|online]() mess up

Signed-off-by: Yong Zhang <yong.zhang0@xxxxxxxxx>
Cc: Sergei Shtylyov <sshtylyov@xxxxxxxxxx>
Cc: David Daney <david.daney@xxxxxxxxxx>
Acked-by: David Daney <david.daney@xxxxxxxxxx>
Patchwork: https://patchwork.linux-mips.org/patch/3845/
Signed-off-by: Ralf Baechle <ralf@xxxxxxxxxxxxxx>

Why are you sending it again?

David Daney



These issues were reported by our
customer CISCO .

Problem Description:
When CONFIG_DEBUG_PREEMPT is enabled the following stack trace occurs.

[ 170.814470] BUG: using smp_processor_id() in preemptible [00000000]
code: sirq-timer/4/62
[ 170.814482] caller is hrtimer_run_pending+0x10/0x20
[ 170.814488] Call Trace:
[ 170.814496] [<ffffffff8010d844>] dump_stack+0x8/0x34
[ 170.814507] [<ffffffff803c2598>] debug_smp_processor_id+0xe0/0xf0
[ 170.814517] [<ffffffff80254158>] hrtimer_run_pending+0x10/0x20
[ 170.814528] [<ffffffff80242400>] run_timer_softirq+0x60/0x348
[ 170.814539] [<ffffffff8023ae80>] run_ksoftirqd+0x1c8/0x348
[ 170.814550] [<ffffffff802503f8>] kthread+0x88/0x90
[ 170.814561] [<ffffffff80206c80>] kernel_thread_helper+0x10/0x18

Root Cause:
Interrupt was occurring before the processor was completely up, and the
softirq
threads were unable to schedule on the processor and then ran on the wrong
CPU.

How Solved:
Enabling of interrupt has been delayed till smp_finish so that
kthread_bind
can safely bind threads to any possible CPU.

I request you to merge the above patch to the main line . If any
questions please contact me atsshaiju@xxxxxxxxxx (shaiju_sada@xxxxxxxxx)

Regards,
Shaiju.


0001-Interrupt-delaying-enabling-of-interrupt.patch


From 58512475cba93003c23f2b380b573e64eebcabd5 Mon Sep 17 00:00:00 2001
From: Sadasivan Shaiju<sshaiju@xxxxxxxxxx>
Date: Mon, 20 Feb 2012 13:25:50 -0800
Subject: [PATCH] Interrupt : delaying enabling of interrupt

Source: MontaVista Software, LLC
MR: 47157
Type: Defect Fix
Disposition: Local
ChangeID: 48c837329556b161f3111e6fded1c9857fa3a149
Description:

This patch is to delay the enabling of interrupt till
smp_finish . So that kthread_bind can safely bind
threads to any possible cpu. Without this change
interrupt should occur beofre the processor was
completely up, and the softirq threads were unable
to schedule on the processor and then ran on the
wrong CPU.

Signed-off-by: Sadasivan Shaiju<sshaiju@xxxxxxxxxx>
---
arch/mips/cavium-octeon/smp.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/arch/mips/cavium-octeon/smp.c b/arch/mips/cavium-octeon/smp.c
index ff21542..c7de7ac 100644
--- a/arch/mips/cavium-octeon/smp.c
+++ b/arch/mips/cavium-octeon/smp.c
@@ -308,7 +308,6 @@ static void octeon_init_secondary(void)
octeon_init_cvmcount();

octeon_irq_setup_secondary();
- raw_local_irq_enable();
}

/**
@@ -365,6 +364,8 @@ static void octeon_smp_finish(void)

/* to generate the first CPU timer interrupt */
write_c0_compare(read_c0_count() + mips_hpt_frequency / HZ);
+ /* enable local interrupts */
+ raw_local_irq_enable();
}

/**
-- 1.7.0.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/