On Wed, Mar 07, 2012 at 12:30:40AM +0100, Alexander Graf wrote:
>
> On 28.02.2012, at 04:18, David Gibson wrote:
>
> > The pseries "xics" interrupt controller, like most interrupt
> > controllers can support both message (i.e. edge sensitive) interrupts
> > and level sensitive interrupts, but it needs to know which are which.
> >
> > When I implemented the xics emulation for qemu, the only devices we
> > supported were the PAPR virtual IO devices. These devices only use
> > message interrupts, so they were the only ones I implemented in xics.
> >
> > Since then, however, we have added support for PCI devices, which use
> > level sensitive interrupts. It turns out the message interrupt logic
> > still actually works most of the time for these, but there are
> > circumstances where we can lost interrupts due to the incorrect
> > interrupt logic.
> >
> > This patch, therefore, implements the correct xics level-sensitive
> > interrupt logic. The type of the interrupt is set when a device
> > allocates a new xics interrupt.
> >
> > Signed-off-by: David Gibson <david@gib...@.dropbear.id.au>
>
> This looks wrong
Oops, I'll fix that up.
>
> > ---
> > hw/spapr.c | 4 +-
> > hw/spapr.h | 12 +++++-
> > hw/spapr_pci.c | 2 +-
> > hw/spapr_vio.c | 2 +-
> > hw/xics.c | 122
> > +++++++++++++++++++++++++++++++++++++++++---------------
> > hw/xics.h | 2 +-
> > 6 files changed, 106 insertions(+), 38 deletions(-)
> >
> > diff --git a/hw/spapr.c b/hw/spapr.c
> > index dffb6a2..828bc53 100644
> > --- a/hw/spapr.c
> > +++ b/hw/spapr.c
> > @@ -83,7 +83,7 @@
> >
> > sPAPREnvironment *spapr;
> >
> > -qemu_irq spapr_allocate_irq(uint32_t hint, uint32_t *irq_num)
> > +qemu_irq spapr_allocate_irq(uint32_t hint, uint32_t *irq_num, int lsi)
> > {
> > uint32_t irq;
> > qemu_irq qirq;
> > @@ -95,7 +95,7 @@ qemu_irq spapr_allocate_irq(uint32_t hint, uint32_t
> > *irq_num)
> > irq = spapr->next_irq++;
> > }
> >
> > - qirq = xics_find_qirq(spapr->icp, irq);
> > + qirq = xics_assign_irq(spapr->icp, irq, lsi);
>
> Find becomes assign? Is this on purpose?
Yes. Previously xics_find_qirq() just returned the qmeu_irq
associated with a given xics irq number, without changing any state.
Now it also flags whether the irq is an lsi or msi, so the old name
didn't seem right any more.
> While at it, could you make the "lsi" thing an enum? That'd make
> things easier to track :).
Ok, will do.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
