Thread (102 messages) 102 messages, 10 authors, 2007-02-01

Re: [RFC/PATCH 14/16] MPIC MSI backend

From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: 2007-01-26 20:42:35

On Thu, 2007-01-25 at 23:43 -0700, Grant Grundler wrote:
On Thu, Jan 25, 2007 at 07:34:16PM +1100, Michael Ellerman wrote:
quoted
MPIC MSI backend. Based on code from Segher, heavily hacked by me.
Renamed to mpic_htmsi, as it only deals with MSI over Hypertransport.
...
quoted
+		/* FIXME should we save the existing type */
+		set_irq_type(virq, IRQ_TYPE_EDGE_RISING);
What exactly does the "virq" represent here?
I'd like to understand if the FIXME comment could be dropped (or not).

I don't get the impression it's related to a PCI IRQ line.
Maybe irq_create_mapping() has comments that describe hwirq and virq?
If not, it would be useful if those terms were described.
Well, this is a ppc specific backend, so I'm not sure we need to
describe in there the way ppc interrupts work but heh ;-) Basically, on
powerpc nowadays, we disconnect "linux" irqs (virtual irqs) and
"hardware" irq numbers.

linux irqs are allocated dynamically and bound to a given PIC/hw irq
pair via irq_create_mapping() or one of the other superset of that
function.

In the case of something like the MPIC MSI backend, we first allocate a
HW vector (we have a bitmap of free vectors), then we map it to a new
virq with irq_create_mapping(). I think the FIXME is not needed.

Ben.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help