Am 03.06.2014 um 07:54 schrieb Paul Mackerras [off-list ref]:
=20

quoted

On Tue, May 20, 2014 at 01:25:11PM +0200, Alexander Graf wrote:
=20

quoted

On 20.05.14 10:30, Gavin Shan wrote:
If we detects frozen state on PE that has been passed to guest, we
needn't handle it. Instead, we rely on the guest to detect and recover
it. The patch avoid EEH event on the frozen passed PE so that the guest
can have chance to handle that.
=20
Signed-off-by: Gavin Shan <redacted>

=20
How does the guest learn about this failure? We'd need to inject an error=

quoted

into it, no?
=20
I think what you want is an irqfd that the in-kernel eeh code notifies wh=

en

quoted

it sees a failure. When such an fd exists, the kernel skips its own error=

quoted

handling.

=20
Well... we don't have irqfd support for book3s HV upstream yet.  The
way the current code is, we have to turn on GSI routing, which puts a
hard and relatively small limit on the hardware IRQ numbers we can use
as it uses a flat array indexed by hardware IRQ number.  Which is a
problem that I need to solve somehow,

Please sync up with the ARM folks on this - they were also unhappy about the=
 routing requirements for irqfd ;).

but it makes using an irqfd
unattractive in the short term.

For EEH it could as well be a dumb eventfd - really just a side channel that=
 can tell user space that something happened asynchronously :).


Alex