On 2021-03-09 05:14, Jürgen Groß wrote:

On 08.03.21 21:33, Boris Ostrovsky wrote:

quoted

On 3/6/21 11:18 AM, Juergen Gross wrote:

quoted

An event channel should be kept masked when an eoi is pending for it.
When being migrated to another cpu it might be unmasked, though.

In order to avoid this keep three different flags for each event channel
to be able to distinguish "normal" masking/unmasking from eoi related
masking/unmasking and temporary masking. The event channel should only
be able to generate an interrupt if all flags are cleared.

Cc: stable@vger.kernel.org
Fixes: 54c9de89895e0a36047 ("xen/events: add a new late EOI evtchn framework")
Reported-by: Julien Grall <redacted>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Julien Grall <redacted>
---
V2:
- introduce a lock around masking/unmasking
- merge patch 3 into this one (Jan Beulich)
V4:
- don't set eoi masking flag in lateeoi_mask_ack_dynirq()

Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>


Ross, are you planning to test this?

Just as another data point: With the previous version of the patches
a reboot loop of a guest needed max 33 reboots to loose network in
my tests (those were IIRC 6 test runs). With this patch version I
stopped the test after about 1300 reboots without having seen any
problems.

Thanks, I'll test it today and get back to you.

Ross