The patch titled
Improve behaviour of spurious IRQ detect
has been added to the -mm tree. Its filename is
improve-behaviour-of-spurious-irq-detect.patch
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
------------------------------------------------------
Subject: Improve behaviour of spurious IRQ detect
From: Alan Cox <alan@xxxxxxxxxxxxxxxxxxx>
Currently we handle spurious IRQ activity based upon seeing a lot of
invalid interrupts, and we clear things back on the base of lots of valid
interrupts.
Unfortunately in some cases you get legitimate invalid interrupts caused by
timing asynchronicity between the PCI bus and the APIC bus when disabling
interrupts and pulling other tricks. In this case although the spurious
IRQs are not a problem our unhandled counters didn't clear and they act as
a slow running timebomb. (This is effectively what the serial port/tty
problem that was fixed by clearing counters when registering a handler
showed up)
It's easy enough to add a second parameter - time. This means that if we
see a regular stream of harmless spurious interrupts which are not harming
processing we don't go off and do something stupid like disable the IRQ
after a month of running. OTOH lockups and performance killers show up a
lot more than 10/second
Signed-off-by: Alan Cox <alan@xxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/irq.h | 1 +
kernel/irq/spurious.c | 10 +++++++++-
2 files changed, 10 insertions(+), 1 deletion(-)
diff -puN include/linux/irq.h~improve-behaviour-of-spurious-irq-detect include/linux/irq.h
--- a/include/linux/irq.h~improve-behaviour-of-spurious-irq-detect
+++ a/include/linux/irq.h
@@ -161,6 +161,7 @@ struct irq_desc {
unsigned int wake_depth; /* nested wake enables */
unsigned int irq_count; /* For detecting broken IRQs */
unsigned int irqs_unhandled;
+ unsigned long last_unhandled; /* Aging timer for unhandled count */
spinlock_t lock;
#ifdef CONFIG_SMP
cpumask_t affinity;
diff -puN kernel/irq/spurious.c~improve-behaviour-of-spurious-irq-detect kernel/irq/spurious.c
--- a/kernel/irq/spurious.c~improve-behaviour-of-spurious-irq-detect
+++ a/kernel/irq/spurious.c
@@ -172,7 +172,15 @@ void note_interrupt(unsigned int irq, st
irqreturn_t action_ret)
{
if (unlikely(action_ret != IRQ_HANDLED)) {
- desc->irqs_unhandled++;
+ /* If we are seeing only the odd spurious IRQ caused by
+ bus asynchronicity then don't eventually trigger an error,
+ otherwise the couter becomes a doomsday timer for otherwise
+ working systems */
+ if (jiffies - desc->last_unhandled > HZ/10)
+ desc->irqs_unhandled = 1;
+ else
+ desc->irqs_unhandled++;
+ desc->last_unhandled = jiffies;
if (unlikely(action_ret != IRQ_NONE))
report_bad_irq(irq, desc, action_ret);
}
_
Patches currently in -mm which might be from alan@xxxxxxxxxxxxxxxxxxx are
lots-of-architectures-enable-arbitary-speed-tty-support.patch
ia64-arbitary-speed-tty-ioctl-support.patch
git-libata-all.patch
pata-add-the-mcp73-77-support-to-pata-driver.patch
libata-correct-abuse-of-language.patch
libata-core-sff-fix-multiple-assumptions-about-dma.patch
libata-fix-hopefully-all-the-remaining-problems-with.patch
testing-patch-for-ali-pata-fixes-hopefully-for-the-problems-with-atapi-dma.patch
pata_ali-more-work.patch
tty-add-the-new-ioctls-and-definitionto-the-mips.patch
add-pci_try_set_mwi.patch
git-scsi-misc.patch
add-pci_try_set_mwi-prism54pci.patch
xtensa-enable-arbitary-tty-speed-setting-ioctls.patch
blackfin-enable-arbitary-speed-serial-setting.patch
h8300-enable-arbitary-speed-tty-port-setup.patch
arm26-enable-arbitary-speed-tty-ioctls-and-split.patch
m32r-enable-arbitary-speed-tty-rate-setting.patch
etrax-enable-arbitary-speed-setting-on-tty-ports.patch
v850-enable-arbitary-speed-tty-ioctls.patch
doc-kernel-parameters-use-x86-32-tag-instead-of-ia-32.patch
make-proc-tty-drivers-use-seq_list_xxx-helpers.patch
update-zilog-timeout.patch
edd-switch-to-pci_get-based-api.patch
mpu401-warning-fixes.patch
char-tty_ioctl-use-wait_event_interruptible_timeout.patch
char-tty_ioctl-little-whitespace-cleanup.patch
intel-rng-undo-mess-made-by-an-80-column-extremist.patch
improve-behaviour-of-spurious-irq-detect.patch
i2o_cfg_passthru-cleanup.patch
i2o_cfg_passthru-cleanup-fix.patch
wrong-memory-access-in-i2o_block_device_lock.patch
i2o-message-leak-in-i2o_msg_post_wait_mem.patch
i2o-proc-reading-oops.patch
i2o-debug-output-cleanup.patch
coredump-masking-bound-suid_dumpable-sysctl.patch
coredump-masking-reimplementation-of-dumpable-using-two-flags.patch
coredump-masking-reimplementation-of-dumpable-using-two-flags-fix.patch
coredump-masking-add-an-interface-for-core-dump-filter.patch
coredump-masking-elf-enable-core-dump-filtering.patch
coredump-masking-elf-fdpic-remove-an-unused-argument.patch
coredump-masking-elf-fdpic-enable-core-dump-filtering.patch
coredump-masking-documentation-for-proc-pid-coredump_filter.patch
-
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
