Impact: new Xen-internal API
Signed-off-by: NIan Campbell <ian.campbell@citrix.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

When retrieving the event channel number from irq, the irq
number may not be valid under some conditions.

So far that can be when we suspend/resume and irq ends with -1.
Validate and return sanitized irq and provide diagnostics information.

[v3: added unlikely on the WARN path]
[v2: reworded the commit message]
Signed-off-by: NJoe Jin <joe.jin@oracle.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Tested-by: NGurudas Pai <gurudas.pai@oracle.com>
Cc: Ian Campbell <ian.campbell@citrix.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Andrew Morton <akpm@linux-foundation.org>

__this_cpu_inc_return reduces code and simplifies code.

Cc: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NChristoph Lameter <cl@linux.com>

Use this_cpu_ops to reduce code size and simplify things in various places.

V3->V4:
	Move instance of this_cpu_inc_return to a later patchset so that
	this patch can be applied without infrastructure changes.

Cc: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NChristoph Lameter <cl@linux.com>
Signed-off-by: NTejun Heo <tj@kernel.org>

The "find_unbound_irq" is a bit unusual - it allocates
virtual IRQ (event channels) in reverse order. This means
starting at the "top" of the available IRQs (nr_irqs) down
to the GSI/MSI IRQs (nr_irqs_gsi). Lets document this and
also make the variables easier to understand.
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Allocate irq descs on any NUMA node (we don't care) rather than
specifically node 0, which may not exist.

(At the moment NUMA is meaningless within a domain, so any info
the kernel has is just from an SRAT table we haven't suppressed/disabled.)
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

Re-map and re-bind all the pirqs at resume time.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>

When remapping MSIs into pirqs for PV on HVM guests, qemu is responsible
for doing the actual mapping and unmapping.
We only give qemu the desired pirq number when we ask to do the mapping
the first time, after that we should be reading back the pirq number
from qemu every time we want to re-enable the MSI.

This fixes a bug in xen_hvm_setup_msi_irqs that manifests itself when
trying to enable the same MSI for the second time: the old MSI to pirq
mapping is still valid at this point but xen_hvm_setup_msi_irqs would
try to assign a new pirq anyway.
A simple way to reproduce this bug is to assign an MSI capable network
card to a PV on HVM guest, if the user brings down the corresponding
ethernet interface and up again, Linux would fail to enable MSIs on the
device.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>

Use the new hypercall PHYSDEVOP_get_free_pirq to ask Xen to allocate a
pirq. Remove the unsupported PHYSDEVOP_get_nr_pirqs hypercall to get the
amount of pirq available.

This fixes find_unbound_pirq that otherwise would return a number
starting from nr_irqs that might very well be out of range in Xen.

The symptom of this bug is that when you passthrough an MSI capable pci
device to a PV on HVM guest, Linux would fail to enable MSIs on the
device.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>

When we allocate a vector for MSI/MSI-X we save away the PIRQ, and the
vector value. When we unmap (de-allocate) the MSI/MSI-X vector(s) we
need to provide the PIRQ and the vector value. What we did instead
was to provide the GSI (which was zero) and the vector value, and we
got these unhappy error messages:

(XEN) irq.c:1575: dom0: pirq 0 not mapped
[    7.733415] unmap irq failed -22

This patches fixes this and we use the PIRQ value instead of the GSI
value.

CC: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

The per-cpu event channel masks can be updated unlocked from multiple
CPUs, so use the locked variant.
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

To bind all event channels to CPU#0, it is not sufficient to set all
of its cpu_evtchn_mask[] bits; all other CPUs also need to get their
bits cleared. Otherwise, evtchn_do_upcall() will start handling
interrupts on CPUs they're not intended to run on, which can be
particularly bad for per-CPU ones.

[ linux-2.6.18-xen.hg 7de7453dee36 ]
Signed-off-by: NJan Beulich <jbeulich@novell.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

The IRQ core code will take care of disabling and reenabling
interrupts over suspend resume automatically, therefore we do not need
to do this in the Xen event channel code.

The only exception is those event channels marked IRQF_NO_SUSPEND
which the IRQ core ignores. We must unmask these ourselves, taking
care to obey the current IRQ_DISABLED status. Failure check for
IRQ_DISABLED leads to enabling polled only event channels, such as
that associated with the pv spinlocks, which must never be enabled:

[   21.970432] ------------[ cut here ]------------
[   21.970432] kernel BUG at arch/x86/xen/spinlock.c:343!
[   21.970432] invalid opcode: 0000 [#1] SMP
[   21.970432] last sysfs file: /sys/devices/virtual/net/lo/operstate
[   21.970432] Modules linked in:
[   21.970432]
[   21.970432] Pid: 0, comm: swapper Not tainted (2.6.32.24-x86_32p-xen-01034-g787c727 #34)
[   21.970432] EIP: 0061:[<c102e209>] EFLAGS: 00010046 CPU: 3
[   21.970432] EIP is at dummy_handler+0x3/0x7
[   21.970432] EAX: 0000021c EBX: dfc16880 ECX: 0000001a EDX: 00000000
[   21.970432] ESI: dfc02c00 EDI: 00000001 EBP: dfc47e10 ESP: dfc47e10
[   21.970432]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0069
[   21.970432] Process swapper (pid: 0, ti=dfc46000 task=dfc39440 task.ti=dfc46000)
[   21.970432] Stack:
[   21.970432]  dfc47e30 c10a39f0 0000021c 00000000 00000000 dfc16880 0000021c 00000001
[   21.970432] <0> dfc47e40 c10a4f08 0000021c 00000000 dfc47e78 c12240a7 c1839284 c1839284
[   21.970432] <0> 00000200 00000000 00000000 f5720000 c1f3d028 c1f3d02c 00000180 dfc47e90
[   21.970432] Call Trace:
[   21.970432]  [<c10a39f0>] ? handle_IRQ_event+0x5f/0x122
[   21.970432]  [<c10a4f08>] ? handle_percpu_irq+0x2f/0x55
[   21.970432]  [<c12240a7>] ? __xen_evtchn_do_upcall+0xdb/0x15f
[   21.970432]  [<c122481e>] ? xen_evtchn_do_upcall+0x20/0x30
[   21.970432]  [<c1030d47>] ? xen_do_upcall+0x7/0xc
[   21.970432]  [<c102007b>] ? apic_reg_read+0xd3/0x22d
[   21.970432]  [<c1002227>] ? hypercall_page+0x227/0x1005
[   21.970432]  [<c102d30b>] ? xen_force_evtchn_callback+0xf/0x14
[   21.970432]  [<c102da7c>] ? check_events+0x8/0xc
[   21.970432]  [<c102da3b>] ? xen_irq_enable_direct_end+0x0/0x1
[   21.970432]  [<c105e485>] ? finish_task_switch+0x62/0xba
[   21.970432]  [<c14e3f84>] ? schedule+0x808/0x89d
[   21.970432]  [<c1084dc5>] ? hrtimer_start_expires+0x1a/0x22
[   21.970432]  [<c1085154>] ? tick_nohz_restart_sched_tick+0x15a/0x162
[   21.970432]  [<c102f43a>] ? cpu_idle+0x6d/0x6f
[   21.970432]  [<c14db29e>] ? cpu_bringup_and_idle+0xd/0xf
[   21.970432] Code: 5d 0f 95 c0 0f b6 c0 c3 55 66 83 78 02 00 89 e5 5d 0f 95 \
c0 0f b6 c0 c3 55 b2 01 86 10 31 c0 84 d2 89 e5 0f 94 c0 5d c3 55 89 e5 <0f> 0b \
eb fe 55 80 3d 4c ce 84 c1 00 89 e5 57 56 89 c6 53 74 15
[   21.970432] EIP: [<c102e209>] dummy_handler+0x3/0x7 SS:ESP 0069:dfc47e10
[   21.970432] ---[ end trace c0b71f7e12cf3011 ]---
Signed-off-by: NIan Campbell <ian.campbell@citrix.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

Use the console hypercalls for dom0 console.

[ Impact: Add Xen dom0 console ]
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NJuan Quintela <quintela@redhat.com>
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Implement xen_create_msi_irq to create an msi and remap it as pirq.
Use xen_create_msi_irq to implement an initial domain specific version
of setup_msi_irqs.
Signed-off-by: NQing He <qing.he@intel.com>
Signed-off-by: NYunhong Jiang <yunhong.jiang@intel.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Implement xen_register_gsi to setup the correct triggering and polarity
properties of a gsi.
Implement xen_register_pirq to register a particular gsi as pirq and
receive interrupts as events.
Call xen_setup_pirqs to register all the legacy ISA irqs as pirqs.
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Map MSIs into pirqs, writing 0 in the MSI vector data field and the pirq
number in the MSI destination id field.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Disable pcifront when running on HVM: it is meant to be used with pv
guests that don't have PCI bus.

Use acpi_register_gsi_xen_hvm to remap GSIs into pirqs.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

xen_hvm_register_pirq allows the kernel to map a GSI into a Xen pirq and
receive the interrupt as an event channel from that point on.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Use PHYSDEVOP_get_nr_pirqs to get the maximum number of pirqs from xen.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

PHYSDEVOP_map_pirq might return a pirq different from what we asked if
we are running as an HVM guest, so we need to be able to support pirqs
that are different from linux irqs.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

* Fix bitmask formatting on 64 bit by specifying correct field widths.

* Output both global and local masked and pending information.

* Indicate in list of pending interrupts whether they are pending in
  the L2, masked globally and/or masked locally.
Signed-off-by: NIan Campbell <ian.campbell@citrix.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

Make sure the irq is set up before binding a virq event channel to it.
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

All event channels startbound to VCPU 0 so ensure that cpu_evtchn_mask
is initialised to reflect this. Otherwise there is a race after registering an
event channel but before the affinity is explicitly set where the event channel
can be delivered. If this happens then the event channel remains pending in the
L1 (evtchn_pending) array but is cleared in L2 (evtchn_pending_sel), this means
the event channel cannot be reraised until another event channel happens to
trigger the same L2 entry on that VCPU.

sizeof(cpu_evtchn_mask(0))==sizeof(unsigned long*) which is not correct, and
causes only the first 32 or 64 event channels (depending on architecture) to be
initially bound to VCPU0. Use sizeof(struct cpu_evtchn_s) instead.
Signed-off-by: NIan Campbell <ian.campbell@citrix.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: stable@kernel.org

Change event delivery to:
 - mask+clear event in the upcall function
 - use handle_fasteoi_irq as the handler
 - unmask in the eoi function (and handle migration)
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

Thomas Gleixner cleaned up event handling to use the
sparse_irq handling, but the xen-pcifront patches utilized the
old mechanism. This fixes them to work with sparse_irq handling.
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

The frontend stub lives in arch/x86/pci/xen.c, alongside other
sub-arch PCI init code (e.g. olpc.c).

It provides a mechanism for Xen PCI frontend to setup/destroy
legacy interrupts, MSI/MSI-X, and PCI configuration operations.

[ Impact: add core of Xen PCI support ]
[ v2: Removed the IOMMU code and only focusing on PCI.]
[ v3: removed usage of pci_scan_all_fns as that does not exist]
[ v4: introduced pci_xen value to fix compile warnings]
[ v5: squished fixes+features in one patch, changed Reviewed-by to Ccs]
[ v7: added Acked-by]
Signed-off-by: NAlex Nixon <alex.nixon@citrix.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NIan Campbell <ian.campbell@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Acked-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Matthew Wilcox <willy@linux.intel.com>
Cc: Qing He <qing.he@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: x86@kernel.org

In driver/xen/events.c, whether bind_pirq is shareable or not is
determined by desc->action is NULL or not. But in __setup_irq,
startup(irq) is invoked before desc->action is assigned with
new action. So desc->action in startup_irq is always NULL, and
bind_pirq is always not shareable. This results in pt_irq_create_bind
failure when passthrough a device which shares irq to other devices.

This patch doesn't use probing_irq to determine if pirq is shareable
or not, instead set shareable flag in irq_info according to trigger
mode in xen_allocate_pirq. Set level triggered interrupts shareable.
Thus use this flag to set bind_pirq flag accordingly.

[v2: arch/x86/xen/pci.c no more, so file skipped]
Signed-off-by: NWeidong Han <weidong.han@intel.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

The 'xen_poll_irq_timeout' provides a method to pass in
the poll timeout for IRQs if requested. We also export
those two poll functions as Xen PCI fronted uses them.
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

In earlier Xen Linux kernels, the IRQ mapping was a straight 1:1 and the
find_unbound_irq started looking around 256 for open IRQs and up. IRQs
from 0 to 255 were reserved for PCI devices.  Previous to this patch,
the 'find_unbound_irq'  started looking at get_nr_hw_irqs() number.
For privileged  domain where the ACPI information is available that
returns the upper-bound of what the GSIs. For non-privileged PV domains,
where ACPI is no-existent the get_nr_hw_irqs() reports the IRQ_LEGACY (16).
With PCI passthrough enabled, and with PCI cards that have IRQs pinned
to a higher number than 16 we collide with previously allocated IRQs.
Specifically the PCI IRQs collide with the IPI's for Xen functions
(as they are allocated earlier).
For example:

00:00.11 USB Controller: ATI Technologies Inc SB700 USB OHCI1 Controller (prog-if 10 [OHCI])
	...
	Interrupt: pin A routed to IRQ 18

[root@localhost ~]# cat /proc/interrupts | head
           CPU0       CPU1       CPU2
 16:      38186          0          0   xen-dyn-virq      timer0
 17:        149          0          0   xen-dyn-ipi       spinlock0
 18:        962          0          0   xen-dyn-ipi       resched0

and when the USB controller is loaded, the kernel reports:
IRQ handler type mismatch for IRQ 18
current handler: resched0

One way to fix this is to reverse the logic when looking for un-used
IRQ numbers and start with the highest available number. With that,
we would get:

           CPU0       CPU1       CPU2
... snip ..
292:         35          0          0   xen-dyn-ipi       callfunc0
293:       3992          0          0   xen-dyn-ipi       resched0
294:        224          0          0   xen-dyn-ipi       spinlock0
295:      57183          0          0   xen-dyn-virq      timer0
NMI:          0          0          0   Non-maskable interrupts
.. snip ..

And interrupts for PCI cards are now accessible.

This patch also includes the fix, found by Ian Campbell, titled
"xen: fix off-by-one error in find_unbound_irq."

[v2: Added an explanation in the code]
[v3: Rebased on top of tip/irq/core]
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

Sometimes cpu_evtchn_mask_p can get used early, before it has been
allocated.  Statically initialize it with an initdata version to catch
any early references.
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Impact: cleanup

Make pirq show useful information in /proc/interrupts

[v2: Removed the parts for arch/x86/xen/pci.c ]
Signed-off-by: NGerd Hoffmann <kraxel@xeni.home.kraxel.org>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Dynamically allocate the irq_info and evtchn_to_irq arrays, so that
1) the irq_info array scales to the actual number of possible irqs,
and 2) we don't needlessly increase the static size of the kernel
when we aren't running under Xen.

Derived on patch from Mike Travis <travis@sgi.com>.

[Impact: reduce memory usage ]
[v2: Conflict in drivers/xen/events.c: Replaced alloc_bootmen with kcalloc ]
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Impact: preserve compat with native

Reserve the lower irq range for use for hardware interrupts so we
can identity-map them.

[v2: Rebased on top tip/irq/core]
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

A privileged PV Xen domain can get direct access to hardware.  In
order for this to be useful, it must be able to get hardware
interrupts.

Being a PV Xen domain, all interrupts are delivered as event channels.
PIRQ event channels are bound to a pirq number and an interrupt
vector.  When a IO APIC raises a hardware interrupt on that vector, it
is delivered as an event channel, which we can deliver to the
appropriate device driver(s).

This patch simply implements the infrastructure for dealing with pirq
event channels.

[ Impact: integrate hardware interrupts into Xen's event scheme ]
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>

There seems to be more cleanups possible, but that's left to the xen
experts :)
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Reviewed-by: NIngo Molnar <mingo@elte.hu>
Cc: Jeremy Fitzhardinge <jeremy@xensource.com>

Xen events are logically edge triggered, as Xen only calls the event
upcall when an event is newly set, but not continuously as it remains set.
As a result, use handle_edge_irq rather than handle_level_irq.

This has the important side-effect of fixing a long-standing bug of
events getting lost if:
 - an event's interrupt handler is running
 - the event is migrated to a different vcpu
 - the event is re-triggered

The most noticable symptom of these lost events is occasional lockups
of blkfront.

Many thanks to Tom Kopec and Daniel Stodden in tracking this down.
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Tom Kopec <tek@acm.org>
Cc: Daniel Stodden <daniel.stodden@citrix.com>
Cc: Stable Kernel <stable@kernel.org>

IPIs and VIRQs are inherently per-cpu event types, so treat them as such:
 - use a specific percpu irq_chip implementation, and
 - handle them with handle_percpu_irq

This makes the path for delivering these interrupts more efficient
(no masking/unmasking, no locks), and it avoid problems with attempts
to migrate them.
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stable Kernel <stable@kernel.org>

This patch introduce a CONFIG_XEN_PVHVM compile time option to
enable/disable Xen PV on HVM support.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>

In general the semantics of IPIs are that they are are expected to
continue functioning after dpm_suspend_noirq().

Specifically I have seen a deadlock between the callfunc IPI and the
stop machine used by xen's do_suspend() routine. If one CPU has already
called dpm_suspend_noirq() then there is a window where it can be sent
a callfunc IPI before all the other CPUs have entered stop_cpu().

If this happens then the first CPU ends up spinning in stop_cpu()
waiting for the other to rendezvous in state STOPMACHINE_PREPARE while
the other is spinning in csd_lock_wait().
Signed-off-by: NIan Campbell <ian.campbell@citrix.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: xen-devel@lists.xensource.com
LKML-Reference: <1280398595-29708-4-git-send-email-ian.campbell@citrix.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>