| commit | 4712c158c5276fd3c401152f4bb5c3fccf185946 | |
| author | Jan Kiszka <jan.kiszka@siemens.com> | |
| Thu, 5 Apr 2018 17:41:47 +0000 (5 19:41 +0200) | ||
| committer | Jason Wang <jasowang@redhat.com> | |
| Fri, 20 Jul 2018 00:30:48 +0000 (20 08:30 +0800) | ||
| tree | 228699be967677c540afa6b0652d059740aa5c25 | treesnapshot (tar.gz zip) |
| parent | 9f2b67e1ca43c84ed37ebd027e7e77a0f2f8ef65 | commitdiff |
e1000e: Prevent MSI/MSI-X storms
Only signal MSI/MSI-X events on rising edges. So far we re-triggered the
interrupt sources even if the guest did no consumed the pending one,
easily causing interrupt storms.
Issue was observable with Linux 4.16 e1000e driver when MSI-X was used.
Vector 2 was causing interrupt storms after the driver activated the
device.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Only signal MSI/MSI-X events on rising edges. So far we re-triggered the
interrupt sources even if the guest did no consumed the pending one,
easily causing interrupt storms.
Issue was observable with Linux 4.16 e1000e driver when MSI-X was used.
Vector 2 was causing interrupt storms after the driver activated the
device.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
| hw/net/e1000e_core.c | diffblobblamehistory | |
| hw/net/e1000e_core.h | diffblobblamehistory |