28 Oct, 2010
1 commit
-
In /proc/stat, the number of per-IRQ event is shown by making a sum each
irq's events on all cpus. But we can make use of kstat_irqs().kstat_irqs() do the same calculation, If !CONFIG_GENERIC_HARDIRQ,
it's not a big cost. (Both of the number of cpus and irqs are small.)If a system is very big and CONFIG_GENERIC_HARDIRQ, it does
for_each_irq()
for_each_cpu()
- look up a radix tree
- read desc->irq_stat[cpu]
This seems not efficient. This patch adds kstat_irqs() for
CONFIG_GENRIC_HARDIRQ and change the calculation asfor_each_irq()
look up radix tree
for_each_cpu()
- read desc->irq_stat[cpu]This reduces cost.
A test on (4096cpusp, 256 nodes, 4592 irqs) host (by Jack Steiner)
%time cat /proc/stat > /dev/null
Before Patch: 2.459 sec
After Patch : .561 sec[akpm@linux-foundation.org: unexport kstat_irqs, coding-style tweaks]
[akpm@linux-foundation.org: fix unused variable 'per_irq_sum']
Signed-off-by: KAMEZAWA Hiroyuki
Tested-by: Jack Steiner
Acked-by: Jack Steiner
Cc: Yinghai Lu
Cc: Ingo Molnar
Cc: Thomas Gleixner
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
13 Oct, 2010
1 commit
-
This option can be set to verify the full conversion to the new chip
functions. Fix the fallout of the patch rework, so the core code
compiles and works with it.Signed-off-by: Thomas Gleixner
12 Oct, 2010
18 commits
-
The allocator functions are now called outside of preempt disabled
regions. Switch to GFP_KERNEL.Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
No callers from atomic regions.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
The move_irq_desc() function was only used due to the problem that the
allocator did not free the old descriptors. So the descriptors had to
be moved in create_irq_nr(). That's history.The code would have never been able to move active interrupt
descriptors on affinity settings. That can be done in a completely
different way w/o all this horror.Remove all of it.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Use the cleanup functions of the dynamic allocator. No need to have
separate implementations.Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
This function should have not been there in the first place.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
sparse irq sets up NR_IRQS_LEGACY irq descriptors and archs then go
ahead and allocate more.Use the unused return value of arch_probe_nr_irqs() to let the
architecture return the number of early allocations. Fix up all users.Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Make irq_to_desc_alloc_node() a wrapper around the new allocator.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Mark a range of interrupts as allocated. In the SPARSE_IRQ=n case we
need this to update the bitmap for the legacy irqs so the enumerator
via irq_get_next_irq() works.Signed-off-by: Thomas Gleixner
-
Use the allocator bitmap to lookup active interrupts.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
/proc/irq never removes any entries, but when irq descriptors can be
freed for real this is necessary. Otherwise we'd reference a freed
descriptor in /proc/irq/NSigned-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
The current sparse_irq allocator has several short comings due to
failures in the design or the lack of it:- Requires iteration over the number of active irqs to find a free slot
(Some architectures have grown their own workarounds for this)
- Removal of entries is not possible
- Racy between create_irq_nr and destroy_irq (plugged by horrible
callbacks)
- Migration of active irq descriptors is not possible
- No bulk allocation of irq ranges
- Sprinkeled irq_desc references all over the place outside of kernel/irq/
(The previous chip functions series is addressing this issue)Implement a sane allocator which fixes the above short comings (though
migration of active descriptors needs a full tree wide cleanup of the
direct and mostly unlocked access to irq_desc).The new allocator still uses a radix_tree, but uses a bitmap for
keeping track of allocated irq numbers. That allows:- Fast lookup of a free slot
- Allows the removal of descriptors
- Prevents the create/destroy race
- Bulk allocation of consecutive irq ranges
- Basic design is ready for migration of life descriptors after
further cleanupsThe bitmap is also used in the SPARSE_IRQ=n case for lookup and
raceless (de)allocation of irq numbers. So it removes the requirement
for looping through the descriptor array to find slots.Right now it uses sparse_irq_lock to protect the bitmap and the radix
tree, but after cleaning up all users we should be able convert that
to a mutex and to switch the radix_tree and decriptor allocations to
GFP_KERNEL.[ Folded in a bugfix from Yinghai Lu ]
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Arch code sets it's own irq_desc.status flags right after boot and for
dynamically allocated interrupts. That might involve iterating over a
huge array.Allow ARCH_IRQ_INIT_FLAGS to set separate flags aside of IRQ_DISABLED
which is the default.Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
The statistics accessor is only used by proc/stats and
show_interrupts(). Both are compiled in.Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
early_init_irq_lock_class() is called way before anything touches the
irq descriptors. In case of SPARSE_IRQ=y this is a NOP operation
because the radix tree is empty at this point. For the SPARSE_IRQ=n
case it's sufficient to set the lock class in early_init_irq().Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
kernel/irq/handle.c has become a dumpground for random code in random
order. Split out the irq descriptor management and the dummy irq_chip
implementation into separate files. Cleanup the include maze while at
it.No code change.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Get the data structure from the core and provide inline wrappers to
access the irq_data members.Provide accessor inlines for irq_data as well.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Provide a irq_desc.status modifier function to cleanup the direct
access to irq_desc in arch and driver code.Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Move irq_desc and internal functions out of irq.h
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar
04 Oct, 2010
18 commits
-
This option covers now the old chip functions and the irq_desc data
fields which are moving to struct irq_data. More stuff will follow.Pretty handy for testing a conversion, whether something broke or not.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Wrap the old chip function retrigger() until the migration is complete
and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function set_wake() until the migration is complete
and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function set_type() until the migration is complete
and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function set_affinity() until the migration is
complete and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function startup() until the migration is complete and
the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip functions disable() and shutdown() until the
migration is complete and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function enable() until the migration is complete and
the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function eoi() until the migration is complete and
the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function mask_ack() until the migration is complete
and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function ack() until the migration is complete and
the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function unmask() until the migration is complete
and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip function mask() until the migration is complete and
the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
Wrap the old chip functions for bus_lock/bus_sync_unlock until the
migration is complete and the old chip functions are removed.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
The compat functions go away when the core code is converted.
Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Convert all references in the core code to orq, chip, handler_data,
chip_data, msi_desc, affinity to irq_data.*Signed-off-by: Thomas Gleixner
Reviewed-by: Ingo Molnar -
Low level chip functions need access to irq_desc->handler_data,
irq_desc->chip_data and irq_desc->msi_desc. We hand down the irq
number to the low level functions, so they need to lookup irq_desc.
With sparse irq this means a radix tree lookup.We could hand down irq_desc itself, but low level chip functions have
no need to fiddle with it directly and we want to restrict access to
irq_desc further.Preparatory patch for new chip functions.
Note, that the ugly anon union/struct is there to avoid a full tree
wide clean up for now. This is not going to last 3 years like __do_IRQ()Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar -
The generic irq Kconfig options are copied around all archs. Provide a
generic Kconfig file which can be included.Signed-off-by: Thomas Gleixner
Cc: Peter Zijlstra
LKML-Reference:
Reviewed-by: H. Peter Anvin
Reviewed-by: Ingo Molnar
24 Sep, 2010
1 commit
-
3 years transition phase is enough. Cleanup the last users and remove
the cruft.Signed-off-by: Thomas Gleixner
Cc: Leo Chen
Cc: Hirokazu Takata
Cc: Chris Metcalf
Cc: Jeff Dike
Cc: Chris Zankel
29 Jul, 2010
1 commit
-
A small number of users of IRQF_TIMER are using it for the implied no
suspend behaviour on interrupts which are not timer interrupts.Therefore add a new IRQF_NO_SUSPEND flag, rename IRQF_TIMER to
__IRQF_TIMER and redefine IRQF_TIMER in terms of these new flags.Signed-off-by: Ian Campbell
Cc: Jeremy Fitzhardinge
Cc: Dmitry Torokhov
Cc: Benjamin Herrenschmidt
Cc: Paul Mackerras
Cc: Grant Likely
Cc: xen-devel@lists.xensource.com
Cc: linux-input@vger.kernel.org
Cc: linuxppc-dev@ozlabs.org
Cc: devicetree-discuss@lists.ozlabs.org
LKML-Reference:
Signed-off-by: Thomas Gleixner
