[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v3 6/8] ARM: VGIC: factor out vgic_get_hw_irq_desc()

On 31/01/18 16:24, Andre Przywara wrote:

On 31/01/18 16:16, Julien Grall wrote:

On 24/01/18 18:10, Andre Przywara wrote:
At the moment we happily access the VGIC internal struct pending_irq
(which describes a virtual IRQ) in irq.c.
Factor out the actually needed functionality to learn the associated
hardware IRQ and move that into gic-vgic.c to improve abstraction.

Signed-off-by: Andre Przywara <andre.przywara@xxxxxxxxxx>
Acked-by: Stefano Stabellini <sstabellini@xxxxxxxxxx>
   xen/arch/arm/gic-vgic.c    | 15 +++++++++++++++
   xen/arch/arm/irq.c         |  7 ++-----
   xen/include/asm-arm/vgic.h |  2 ++
   3 files changed, 19 insertions(+), 5 deletions(-)

diff --git a/xen/arch/arm/gic-vgic.c b/xen/arch/arm/gic-vgic.c
index d44e4dacd3..3ad98dcd3a 100644
--- a/xen/arch/arm/gic-vgic.c
+++ b/xen/arch/arm/gic-vgic.c
@@ -411,6 +411,21 @@ void gic_dump_vgic_info(struct vcpu *v)
           printk("Pending irq=%d\n", p->irq);
   +struct irq_desc *vgic_get_hw_irq_desc(struct domain *d, struct vcpu
+                                      unsigned int virq)
+    struct pending_irq *p;
+    if ( !v )
+        v = d->vcpu[0];

I dislike this new function in the current state. Let's imagine someone
decide to pass a PPI but no vCPU. The vCPU would now be default to vCPU0
and potentially returning the wrong irq_desc (imagine PPI passthrough
such as for the vtimer).

So the code needs at least checking the vCPU is passed in the case of a
PPI. I would be happy with an ASSERT().

Yeah, good point. Something like ASSERT(!v && virq >= 32), I guess?

Yes. It should be enough.



Julien Grall

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.