tip is broken with NOHZ: restart tick device from irq_enter()

From: Yinghai Lu
Date: Mon Oct 20 2008 - 19:12:50 EST

all my servers are broken with following commit.

kernel/tip/linux-2.6> git bisect bad
fb02fbc14d17837b4b7b02dbb36142c16a7bf208 is first bad commit
commit fb02fbc14d17837b4b7b02dbb36142c16a7bf208
Author: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Date: Fri Oct 17 10:01:23 2008 +0200

NOHZ: restart tick device from irq_enter()

We did not restart the tick device from irq_enter() to avoid double
reprogramming and extra events in the return immediate to idle case.

But long lasting softirqs can lead to a situation where jiffies become

tick stopped (reprogrammed to next pending timer)
jiffies updated from irq_enter()
interrupt handler
softirq function 1 runs 20ms
softirq function 2 arms a 10ms timer with a stale jiffies value
jiffies updated from irq_exit()
timer wheel has now an already expired timer
(the one added in function 2)
timer fires and timer softirq runs

This was discovered when debugging a timer problem which happend only
when the ath5k driver is active. The debugging proved that there is a
softirq function running for more than 20ms, which is a bug by itself.

To solve this we restart the tick timer right from irq_enter(), but do
not go through the other functions which are necessary to return from
idle when need_resched() is set.

Reported-by: Elias Oltmanns <eo@xxxxxxxxxxxxxx>
Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Tested-by: Elias Oltmanns <eo@xxxxxxxxxxxxxx>

:040000 040000 d36218956b6a233bac4f56acfa0b106f6301c0bd
f9c5 M kernel
yhlu@linux-zpir:~/xx/xx/kernel/tip/linux-2.6> git bisect log
git-bisect start
# bad: [8600bfdb4112c49ad09e7339010221e4a531716d] Merge branch 'warnings/simple'
git-bisect bad 8600bfdb4112c49ad09e7339010221e4a531716d
# good: [9601fd2e889cda328dbe66c2a907973916567c11] Merge branch 'sched/urgent'
git-bisect good 9601fd2e889cda328dbe66c2a907973916567c11
# good: [0cfd81031a26717fe14380d18275f8e217571615] Merge
git-bisect good 0cfd81031a26717fe14380d18275f8e217571615
# good: [2414719fbef90730a7521efb0074b8fb5905557f] Merge branch 'sched/urgent'
git-bisect good 2414719fbef90730a7521efb0074b8fb5905557f
# bad: [49ee4ff800b1d624fcd8bfa1ea3a55d1227635c3] Merge branch 'x86/doc'
git-bisect bad 49ee4ff800b1d624fcd8bfa1ea3a55d1227635c3
# bad: [0c4a0feaeb867cfa9486cef7555e1f23ceca2609] Merge branch 'tracing/urgent'
git-bisect bad 0c4a0feaeb867cfa9486cef7555e1f23ceca2609
# bad: [a597cb549361f6211d98edc99e672555899838f9] manual merge of
git-bisect bad a597cb549361f6211d98edc99e672555899838f9
# bad: [14fedb9d17f5cb35ba805e39b5db9b48cee44c7e] manual merge of timers/nohz
git-bisect bad 14fedb9d17f5cb35ba805e39b5db9b48cee44c7e
# good: [322acf6585f3c4e82ee32a246b0483ca0f6ad3f4] fix documentation
of sysrq-q really
git-bisect good 322acf6585f3c4e82ee32a246b0483ca0f6ad3f4
# good: [c34bec5a44e9486597d78e7a686b2f9088a0564c] NOHZ: split
git-bisect good c34bec5a44e9486597d78e7a686b2f9088a0564c
# bad: [fb02fbc14d17837b4b7b02dbb36142c16a7bf208] NOHZ: restart tick
device from irq_enter()
git-bisect bad fb02fbc14d17837b4b7b02dbb36142c16a7bf208
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/