[RFC PATCH v3 16/19] Update dimm state on reset

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



in case of hot-remove failure on a guest that does not implement _OST,
the dimm bitmaps in qemu and Seabios show the dimm as unplugged, but the dimm
is still present on the qdev/memory bus. To avoid this inconsistency, we set the
dimm state to active/hot-plugged on a reset of the associated acpi_pm device.
This way the dimm is still active after a VM reboot and dimm visibility has
always the same behaviour, regardless of _OST support in the guest.

Signed-off-by: Vasilis Liaskovitis <vasilis.liaskovitis@xxxxxxxxxxxxxxxx>
---
 hw/acpi_piix4.c |    1 +
 hw/dimm.c       |   20 ++++++++++++++++++++
 hw/dimm.h       |    1 +
 3 files changed, 22 insertions(+), 0 deletions(-)

diff --git a/hw/acpi_piix4.c b/hw/acpi_piix4.c
index f7220d4..8bf58a6 100644
--- a/hw/acpi_piix4.c
+++ b/hw/acpi_piix4.c
@@ -373,6 +373,7 @@ static void piix4_reset(void *opaque)
         pci_conf[0x5B] = 0x02;
     }
     piix4_update_hotplug(s);
+    dimm_state_sync();
 }
 
 static void piix4_powerdown(void *opaque, int irq, int power_failing)
diff --git a/hw/dimm.c b/hw/dimm.c
index 1521462..b993668 100644
--- a/hw/dimm.c
+++ b/hw/dimm.c
@@ -182,6 +182,26 @@ static DimmDevice *dimm_find_from_idx(uint32_t idx)
     return NULL;
 }
 
+void dimm_state_sync(void)
+{
+    DimmBus *bus = main_memory_bus;
+    DimmDevice *slot;
+
+    /* if a hot-remove operation is pending on reset, it means the hot-remove
+     * operation has failed, but the guest hasn't notified us e.g. because the
+     * guest does not provide _OST notifications. The device is still present on
+     * the dimmbus, but the qemu and Seabios dimm bitmaps show this device as
+     * unplugged. To avoid this inconsistency, we set the dimm bits to active
+     * i.e. hot-plugged for each dimm present on the dimmbus.
+     */
+    QTAILQ_FOREACH(slot, &bus->dimmlist, nextdimm) {
+        if (slot->pending == DIMM_REMOVE_PENDING) {
+            if (bus->dimm_revert)
+                bus->dimm_revert(bus->dimm_hotplug_qdev, slot, 0);
+        }
+    }
+}
+
 /* used to create a dimm device, only on incoming migration of a hotplugged
  * RAMBlock
  */
diff --git a/hw/dimm.h b/hw/dimm.h
index a6c6e6f..ce091fe 100644
--- a/hw/dimm.h
+++ b/hw/dimm.h
@@ -95,5 +95,6 @@ void main_memory_bus_create(Object *parent);
 void dimm_config_create(char *id, uint64_t size, uint64_t node,
         uint32_t dimm_idx, uint32_t populated);
 uint64_t get_hp_memory_total(void);
+void dimm_state_sync(void);
 
 #endif
-- 
1.7.9

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux