linux/kernel/irq, branch v2.6.28.4

genirq: __irq_set_trigger: change pr_warning to pr_debug

2008-11-13T10:59:48Z

Commit 0c5d1eb77a8be917b638344a22afe1398236482b (genirq: record trigger type) caused powerpc platforms that had no set_type() function in their struct irq_chip to spew out warnings about "No set_type function for IRQ...". This warning isn't necessarily justified though because the generic powerpc platform code calls set_irq_type() (which in turn calls __irq_set_trigger) with information from the device tree to establish the interrupt mappings, regardless of whether the PIC can actually set a type. A platform's irq_chip might not have a set_type function for a variety of reasons, for example: the platform may have the type essentially hard-coded, or as in the case for Cell interrupts are just messages past around that have no real concept of type, or the platform could even have a virtual PIC as on the PS3. Signed-off-by: Mark Nelson Signed-off-by: Ingo Molnar

irq: fix typo

2008-11-09T21:26:45Z

Impact: build fix fix build failure on UP. Signed-off-by: Ingo Molnar

genirq: fix the affinity setting in setup_irq

2008-11-09T21:23:54Z

The affinity setting in setup irq is called before the NO_BALANCING flag is checked and might therefore override affinity settings from the calling code with the default setting. Move the NO_BALANCING flag check before the call to the affinity setting. Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar

genirq: keep affinities set from userspace across free/request_irq()

2008-11-09T21:23:49Z

Impact: preserve user-modified affinities on interrupts Kumar Galak noticed that commit 18404756765c713a0be4eb1082920c04822ce588 (genirq: Expose default irq affinity mask (take 3)) overrides an already set affinity setting across a free / request_irq(). Happens e.g. with ifdown/ifup of a network device. Change the logic to mark the affinities as set and keep them intact. This also fixes the unlocked access to irq_desc in irq_select_affinity() when called from irq_affinity_proc_write() Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar

irq: make variable static

2008-10-22T05:37:17Z

This variable is only used in the source file, so make it static. Signed-off-by: Roel Kluin Signed-off-by: Ingo Molnar

genirq: NULL struct irq_desc's member 'name' in dynamic_irq_cleanup()

2008-10-21T13:59:21Z

If the member 'name' of the irq_desc structure happens to point to a character string that is resident within a kernel module, problems ensue if that module is rmmod'd (at which time dynamic_irq_cleanup() is called) and then later show_interrupts() is called by someone. It is also not a good thing if the character string resided in kmalloc'd space that has been kfree'd (after having called dynamic_irq_cleanup()). dynamic_irq_cleanup() fails to NULL the 'name' member and show_interrupts() references it on a few architectures (like h8300, sh and x86). Signed-off-by: Dean Nelson Signed-off-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar

genirq: fix set_irq_type() when recording trigger type

2008-10-21T08:10:08Z

Impact: fix boot hang on a G5 In set_irq_type() we want to pass the type rather than the current interrupt state. Signed-off-by: Chris Friesen Acked-by: Benjamin Herrenschmidt Acked-by: David Brownell Signed-off-by: Ingo Molnar

genirq: fix name space collision of nr_irqs in autoprobe.c

2008-10-16T14:53:30Z

probe_irq_off() is disfunctional as the local nr_irqs is referenced instead of the global one for the for_each_irq_desc() iterator. Signed-off-by: Thomas Gleixner

genirq: use iterators for irq_desc loops

2008-10-16T14:53:30Z

Use for_each_irq_desc[_reverse] for all the iteration loops. Signed-off-by: Thomas Gleixner

genirq: cleanup the sparseirq modifications

2008-10-16T14:53:29Z

Signed-off-by: Thomas Gleixner