Re: [Xen-devel] Hpet interrupt overflow

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 13.08.13 at 11:03, Andrew Cooper <andrew.cooper3@xxxxxxxxxx> wrote:
> We had another crash, this time with a proper stack trace.  (This was
> using an early version stack trace improvements series)
> 
> From the stack trace (now correctly with frame pointers), we see 9 calls
> to handle_hpet_broadcast().
> 
> This indicates that the current logic does not correctly prevent
> repeated delivery of interrupts.

And this was with a 1:1 CPU <-> HPET channel mapping (not
visible from just the stack trace)?

In any case, could you try moving the call to ack_APIC_irq() from
hpet_msi_ack() to hpet_msi_end() (the latter may need to be
re-created depending on the Xen version you do this with). Or,
as another alternative, call hpet_msi_{,un}mask() from the two
functions respectively (albeit I think this might result in lost
interrupts).

Potentially hpet_msi_end() would then also need to disable
interrupts before doing either of these actions.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel