[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 28/29] apic: Use 32bit APIC ID for migration instance ID
From: |
Juan Quintela |
Subject: |
[PULL 28/29] apic: Use 32bit APIC ID for migration instance ID |
Date: |
Tue, 14 Jan 2020 10:26:05 +0100 |
From: Peter Xu <address@hidden>
Migration is silently broken now with x2apic config like this:
-smp 200,maxcpus=288,sockets=2,cores=72,threads=2 \
-device intel-iommu,intremap=on,eim=on
After migration, the guest kernel could hang at anything, due to
x2apic bit not migrated correctly in IA32_APIC_BASE on some vcpus, so
any operations related to x2apic could be broken then (e.g., RDMSR on
x2apic MSRs could fail because KVM would think that the vcpu hasn't
enabled x2apic at all).
The issue is that the x2apic bit was never applied correctly for vcpus
whose ID > 255 when migrate completes, and that's because when we
migrate APIC we use the APICCommonState.id as instance ID of the
migration stream, while that's too short for x2apic.
Let's use the newly introduced initial_apic_id for that.
Signed-off-by: Peter Xu <address@hidden>
Reviewed-by: Juan Quintela <address@hidden>
Reviewed-by: Eduardo Habkost <address@hidden>
Signed-off-by: Juan Quintela <address@hidden>
---
hw/intc/apic_common.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/hw/intc/apic_common.c b/hw/intc/apic_common.c
index 54b8731fca..b5dbeb6206 100644
--- a/hw/intc/apic_common.c
+++ b/hw/intc/apic_common.c
@@ -268,7 +268,10 @@ static void apic_common_realize(DeviceState *dev, Error
**errp)
APICCommonState *s = APIC_COMMON(dev);
APICCommonClass *info;
static DeviceState *vapic;
- uint32_t instance_id = s->id;
+ uint32_t instance_id = s->initial_apic_id;
+
+ /* Normally initial APIC ID should be no more than hundreds */
+ assert(instance_id != VMSTATE_INSTANCE_ID_ANY);
info = APIC_COMMON_GET_CLASS(s);
info->realize(dev, errp);
--
2.24.1
- [PULL 18/29] migration/postcopy: set all_zero to true on the first target page, (continued)
- [PULL 18/29] migration/postcopy: set all_zero to true on the first target page, Juan Quintela, 2020/01/14
- [PULL 19/29] migration/postcopy: enable random order target page arrival, Juan Quintela, 2020/01/14
- [PULL 20/29] migration/postcopy: enable compress during postcopy, Juan Quintela, 2020/01/14
- [PULL 21/29] migration/multifd: clean pages after filling packet, Juan Quintela, 2020/01/14
- [PULL 22/29] migration/multifd: not use multifd during postcopy, Juan Quintela, 2020/01/14
- [PULL 23/29] migration/multifd: fix nullptr access in terminating multifd threads, Juan Quintela, 2020/01/14
- [PULL 24/29] migration/multifd: fix destroyed mutex access in terminating multifd threads, Juan Quintela, 2020/01/14
- [PULL 25/29] Bug #1829242 correction., Juan Quintela, 2020/01/14
- [PULL 26/29] migration: Define VMSTATE_INSTANCE_ID_ANY, Juan Quintela, 2020/01/14
- [PULL 27/29] migration: Change SaveStateEntry.instance_id into uint32_t, Juan Quintela, 2020/01/14
- [PULL 28/29] apic: Use 32bit APIC ID for migration instance ID,
Juan Quintela <=
- [PULL 29/29] migration: Support QLIST migration, Juan Quintela, 2020/01/14
- Re: [PULL 00/29] Migration pull patches (second try), Peter Maydell, 2020/01/14
- Re: [PULL 00/29] Migration pull patches (second try), no-reply, 2020/01/14
- Re: [PULL 00/29] Migration pull patches (second try), Peter Maydell, 2020/01/14