Every time a cpu is added via hotplug, we allocate the per-cpu MONDO
queues but we never free them up. Freeing isn't easy since the first
cpu gets this memory from bootmem.
Therefore, the simplest thing to do to fix this bug is to allocate the
queues for all possible cpus at boot time.
Signed-off-by: David S. Miller <davem@davemloft.net>
The dev_handle and dev_ino fields don't match up exactly to
the traditional IMAP_IGN and IMAP_INO masks.
So store them away in a table and look them up directly.
Signed-off-by: David S. Miller <davem@davemloft.net>
dr-cpu unconfigure requests will walk throught he enabled
IRQs and trigger ->set_affinity so that the going-down
cpu no longer has INOs targetted to it.
Signed-off-by: David S. Miller <davem@davemloft.net>
This is also a partial workaround for a bug in the LDOM firmware which
double-transmits RX inos during high load. Without this, such an
event causes the kernel to loop forever in the interrupt call chain
ACK'ing but never actually running the IRQ handler (and thus clearing
the interrupt condition in the device).
There is still a bad potential effect when double INOs occur,
not covered by this changeset. Namely, if the INO is already on
the per-cpu INO vector list, we still blindly re-insert it and
thus we can end up losing interrupts already linked in after
it.
We could deal with that by traversing the list before insertion,
but that's too expensive for this edge case.
Signed-off-by: David S. Miller <davem@davemloft.net>
We were doing the wrong call to turn them on, and also
when enabling we need to forcefully set the state to IDLE.
Signed-off-by: David S. Miller <davem@davemloft.net>
Some minor refactoring in the generic code was necessary for
this:
1) This controller requires 8-byte access to the interrupt map
and clear register. They are 64-bits on all the other
SBUS and PCI controllers anyways, so this was easy to cure.
2) The IMAP register has a different layout and some bits that we
need to preserve, so use a read/modify/write when making
changes to the IMAP register in generic code.
3) Flushing the entire IOMMU TLB is best done with a single write
to a register on this PCI controller, add a iommu->iommu_flushinv
for this.
Still lacks MSI support, that will come later.
Signed-off-by: David S. Miller <davem@davemloft.net>
Things were scattered all over the place, split between
SMP and non-SMP.
Unify it all so that dyntick support is easier to add.
Signed-off-by: David S. Miller <davem@davemloft.net>
This is kind of hokey, we could use the hardware provided facilities
much better.
MSIs are assosciated with MSI Queues. MSI Queues generate interrupts
when any MSI assosciated with it is signalled. This suggests a
two-tiered IRQ dispatch scheme:
MSI Queue interrupt --> queue interrupt handler
MSI dispatch --> driver interrupt handler
But we just get one-level under Linux currently. What I'd like to do
is possibly stick the IRQ actions into a per-MSI-Queue data structure,
and dispatch them form there, but the generic IRQ layer doesn't
provide a way to do that right now.
So, the current kludge is to "ACK" the interrupt by processing the
MSI Queue data structures and ACK'ing them, then we run the actual
handler like normal.
We are wasting a lot of useful information, for example the MSI data
and address are provided with ever MSI, as well as a system tick if
available. If we could pass this into the IRQ handler it could help
with certain things, in particular for PCI-Express error messages.
The MSI entries on sparc64 also tell you exactly which bus/device/fn
sent the MSI, which would be great for error handling when no
registered IRQ handler can service the interrupt.
We override the disable/enable IRQ chip methods in sun4v_msi, so we
have to call {mask,unmask}_msi_irq() directly from there. This is
another ugly wart.
Signed-off-by: David S. Miller <davem@davemloft.net>
Otherwise we can't use the generic MSI code.
Furthermore, properly use the {get,set}_irq_foo() abstracted
interfaces instead of direct accesses to irq_desc[]->foo.
Signed-off-by: David S. Miller <davem@davemloft.net>
Consolidation: remove the irq_affinity[NR_IRQS] array and move it into the
irq_desc[NR_IRQS].affinity field.
[akpm@osdl.org: sparc64 build fix]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch-queue improves the generic IRQ layer to be truly generic, by adding
various abstractions and features to it, without impacting existing
functionality.
While the queue can be best described as "fix and improve everything in the
generic IRQ layer that we could think of", and thus it consists of many
smaller features and lots of cleanups, the one feature that stands out most is
the new 'irq chip' abstraction.
The irq-chip abstraction is about describing and coding and IRQ controller
driver by mapping its raw hardware capabilities [and quirks, if needed] in a
straightforward way, without having to think about "IRQ flow"
(level/edge/etc.) type of details.
This stands in contrast with the current 'irq-type' model of genirq
architectures, which 'mixes' raw hardware capabilities with 'flow' details.
The patchset supports both types of irq controller designs at once, and
converts i386 and x86_64 to the new irq-chip design.
As a bonus side-effect of the irq-chip approach, chained interrupt controllers
(master/slave PIC constructs, etc.) are now supported by design as well.
The end result of this patchset intends to be simpler architecture-level code
and more consolidation between architectures.
We reused many bits of code and many concepts from Russell King's ARM IRQ
layer, the merging of which was one of the motivations for this patchset.
This patch:
rename desc->handler to desc->chip.
Originally i did not want to do this, because it's a big patch. But having
both "desc->handler", "desc->handle_irq" and "action->handler" caused a
large degree of confusion and made the code appear alot less clean than it
truly is.
I have also attempted a dual approach as well by introducing a
desc->chip alias - but that just wasnt robust enough and broke
frequently.
So lets get over with this quickly. The conversion was done automatically
via scripts and converts all the code in the kernel.
This renaming patch is the first one amongst the patches, so that the
remaining patches can stay flexible and can be merged and split up
without having some big monolithic patch act as a merge barrier.
[akpm@osdl.org: build fix]
[akpm@osdl.org: another build fix]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is the long overdue conversion of sparc64 over to
the generic IRQ layer.
The kernel image is slightly larger, but the BSS is ~60K
smaller due to the reduced size of struct ino_bucket.
A lot of IRQ implementation details, including ino_bucket,
were moved out of asm-sparc64/irq.h and are now private to
arch/sparc64/kernel/irq.c, and most of the code in irq.c
totally disappeared.
One thing that's different at the moment is IRQ distribution,
we do it at enable_irq() time. If the cpu mask is ALL then
we round-robin using a global rotating cpu counter, else
we pick the first cpu in the mask to support single cpu
targetting. This is similar to what powerpc's XICS IRQ
support code does.
This works fine on my UP SB1000, and the SMP build goes
fine and runs on that machine, but lots of testing on
different setups is needed.
Signed-off-by: David S. Miller <davem@davemloft.net>
Inspired by PowerPC XICS interrupt support code.
All IRQs are virtualized in order to keep NR_IRQS from needing
to be too large. Interrupts on sparc64 are arbitrary 11-bit
values, but we don't need to define NR_IRQS to 2048 if we
virtualize the IRQs.
As PCI and SBUS controller drivers build device IRQs, we divy
out virtual IRQ numbers incrementally starting at 1. Zero is
a special virtual IRQ used for the timer interrupt.
So device drivers all see virtual IRQs, and all the normal
interfaces such as request_irq(), enable_irq(), etc. translate
that into a real IRQ number in order to configure the IRQ.
At this point knowledge of the struct ino_bucket is almost
entirely contained within arch/sparc64/kernel/irq.c There are
a few small bits in the PCI controller drivers that need to
be swept away before we can remove ino_bucket's definition
out of asm-sparc64/irq.h and privately into kernel/irq.c
Signed-off-by: David S. Miller <davem@davemloft.net>
And reuse that struct member for virt_irq, which will
be used in future changesets for the implementation of
mapping between real and virtual IRQ numbers.
This nicely kills off a ton of SBUS and PCI controller
PIL assignment code which is no longer necessary.
Signed-off-by: David S. Miller <davem@davemloft.net>
Only pil0_dummy_bucket had a pil of zero and we just killed that
off, so we can delete all special case code that used bp->pil==0
as a way to identify a dummy bucket.
Signed-off-by: David S. Miller <davem@davemloft.net>
This is the first in a series of cleanups that will hopefully
allow a seamless attempt at using the generic IRQ handling
infrastructure in the Linux kernel.
Define PIL_DEVICE_IRQ and vector all device interrupts through
there.
Get rid of the ugly pil0_dummy_{bucket,desc}, instead vector
the timer interrupt directly to a specific handler since the
timer interrupt is the only event that will be signaled on
PIL 14.
The irq_worklist is now in the per-cpu trap_block[].
Signed-off-by: David S. Miller <davem@davemloft.net>
When we stop allocating percpu memory for not-possible CPUs we must not touch
the percpu data for not-possible CPUs at all. The correct way of doing this
is to test cpu_possible() or to use for_each_cpu().
This patch is a kernel-wide sweep of all instances of NR_CPUS. I found very
few instances of this bug, if any. But the patch converts lots of open-coded
test to use the preferred helper macros.
Cc: Mikael Starvik <starvik@axis.com>
Cc: David Howells <dhowells@redhat.com>
Acked-by: Kyle McMartin <kyle@parisc-linux.org>
Cc: Anton Blanchard <anton@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Christian Zankel <chris@zankel.net>
Cc: Philippe Elie <phil.el@wanadoo.fr>
Cc: Nathan Scott <nathans@sgi.com>
Cc: Jens Axboe <axboe@suse.de>
Cc: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
this patch converts arch/sparc64 to kzalloc usage.
Crosscompile tested with allyesconfig.
Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
We need to use the real hardware processor ID when
targetting interrupts, not the "define to 0" thing
the uniprocessor build gives us.
Also, fill in the Node-ID and Agent-ID fields properly
on sun4u/Safari.
Signed-off-by: David S. Miller <davem@davemloft.net>
The sibling cpu bringup is extremely fragile. We can only
perform the most basic calls until we take over the trap
table from the firmware/hypervisor on the new cpu.
This means no accesses to %g4, %g5, %g6 since those can't be
TLB translated without our trap handlers.
In order to achieve this:
1) Change sun4v_init_mondo_queues() so that it can operate in
several modes.
It can allocate the queues, or install them in the current
processor, or both.
The boot cpu does both in it's call early on.
Later, the boot cpu allocates the sibling cpu queue, starts
the sibling cpu, then the sibling cpu loads them in.
2) init_cur_cpu_trap() is changed to take the current_thread_info()
as an argument instead of reading %g6 directly on the current
cpu.
3) Create a trampoline stack for the sibling cpus. We do our basic
kernel calls using this stack, which is locked into the kernel
image, then go to our proper thread stack after taking over the
trap table.
4) While we are in this delicate startup state, we put 0xdeadbeef
into %g4/%g5/%g6 in order to catch accidental accesses.
5) On the final prom_set_trap_table*() call, we put &init_thread_union
into %g6. This is a hack to make prom_world(0) work. All that
wants to do is restore the %asi register using
get_thread_current_ds().
Longer term we should just do the OBP calls to set the trap table by
hand just like we do for everything else. This would avoid that silly
prom_world(0) issue, then we can remove the init_thread_union hack.
Signed-off-by: David S. Miller <davem@davemloft.net>
On SUN4V, force IRQ state to idle in enable_irq(). However,
I'm still not sure this is %100 correct.
Call add_interrupt_randomness() on SUN4V too.
Signed-off-by: David S. Miller <davem@davemloft.net>
We have to use bootmem during init_IRQ and page alloc
for sibling cpu calls.
Also, fix incorrect hypervisor call return value
checks in the hypervisor SMP cpu mondo send code.
Signed-off-by: David S. Miller <davem@davemloft.net>
