2015-04-02 1:35 GMT+08:00 Laine Stump <laine@xxxxxxxxx>:
So I don't think the problem you said exists. This solution is OK.
Am I right?
On 04/01/2015 12:15 PM, Huanle Han wrote:
On 2015年03月31日 17:46, Martin Kletzander wrote:
On Thu, Mar 26, 2015 at 10:23:47PM +0800, Huanle Han wrote:
Fix for such a case:
1. Domain A and B xml contain the same SRIOV net hostdev(<interface
type='hostdev' /> with same pci address).
2. virsh start A (Successfully, and configure the SRIOV net with
custom mac)
3. virsh start B (Fail because of the hostdev used by domain A or other
reason.)
In step 3, 'virHostdevNetConfigRestore' is called for the hostdev
which is still used by domain A. It makes the mac/vlan of the SRIOV net
change.
Makes perfect sense, thanks for finding that out.
Someone else encountered this last month too, and posted a patch:
https://www.redhat.com/archives/libvir-list/2015-February/msg00394.html
The patch was NACKed because I didn't think the fix was correct (although it did work)
https://www.redhat.com/archives/libvir-list/2015-February/msg00976.html
I don't have the time to do a close analysis of this patch right now, but from the comments it sounds like it addresses the concern that I had with the previous patch (that it wouldn't be messing with any devices that were currently not in use by any domain, and hadn't yet been reached in the setup part). I wanted to point that out now so that whoever looks at the next version of this patch checks for that.
active virPCIDevice has member variable 'used_by_drvname' and
'used_by_domname' to mark its user. According to
’virHostdevPreparePCIDevices‘, only the pci is successfully prepared,
will the member 'used_by_XXX' be assigned.
The domain PCI state when reattach:
1.
If qemuProcessStart() fails before pci
setup(virHostdevPreparePCIDevices()), the pci will either be not active,
or be active but used by others.
2. if qemuProcessStart() fails during pci setup , virHostdevPreparePCIDevices rollbacks the change. The result will be the same with 1.
2. if qemuProcessStart() fails during pci setup , virHostdevPreparePCIDevices rollbacks the change. The result will be the same with 1.
3. if qemuProcessStart() fails after pci setup, the pci will be active and used by this domain.
In fucntion virHostdevReAttachPCIDevices,
1. pcidevs is from virHostdevGetActivePCIHostDeviceList(). That is say, pci in pcidevs is active ( used/prepared by any domains).
2. In loop 1, the pci used by other domains is removed from the pcidevs. After loop, all pcis in pcidevs are used/prepared by this domain.
3. The following reattach operation is on these 'pcidevs', not the pci used by nobody or other domains.
1. pcidevs is from virHostdevGetActivePCIHostDeviceList(). That is say, pci in pcidevs is active ( used/prepared by any domains).
2. In loop 1, the pci used by other domains is removed from the pcidevs. After loop, all pcis in pcidevs are used/prepared by this domain.
3. The following reattach operation is on these 'pcidevs', not the pci used by nobody or other domains.
Is comment as below OK?
Code Change in this fix:
1. As the pci used by other domain have been removed from
'pcidevs' in previous loop, we only restore the nic config for
the hostdev still in 'pcidevs'(used by this domain)
2. wrap a function 'virHostdevIsPCINetDevice', which detect whether the
hostdev is a pci net device or not.
3. update the comments to make it more clear
Signed-off-by: Huanle Han <hanxueluo@xxxxxxxxx>
---
src/util/virhostdev.c | 52 ++++++++++++++++++++++++++++++++++++---------------
1 file changed, 37 insertions(+), 15 deletions(-)
diff --git a/src/util/virhostdev.c b/src/util/virhostdev.c
index 83f567d..b04bae2 100644
--- a/src/util/virhostdev.c
+++ b/src/util/virhostdev.c
@@ -350,6 +350,14 @@ virHostdevNetDevice(virDomainHostdevDefPtr hostdev, char **linkdev,
return ret;
}
+static int
+virHostdevIsPCINetDevice(virDomainHostdevDefPtr hostdev)
+{
+ return hostdev->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
+ hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI &&
+ hostdev->parent.type == VIR_DOMAIN_DEVICE_NET &&
+ hostdev->parent.data.net;
+}
static int
virHostdevNetConfigVirtPortProfile(const char *linkdev, int vf,
@@ -481,10 +489,7 @@ virHostdevNetConfigRestore(virDomainHostdevDefPtr hostdev,
/* This is only needed for PCI devices that have been defined
* using <interface type='hostdev'>. For all others, it is a NOP.
*/
- if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS ||
- hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI ||
- hostdev->parent.type != VIR_DOMAIN_DEVICE_NET ||
- !hostdev->parent.data.net)
+ if (!virHostdevIsPCINetDevice(hostdev))
return 0;
isvf = virHostdevIsVirtualFunction(hostdev);
@@ -781,19 +786,20 @@ virHostdevReAttachPCIDevices(virHostdevManagerPtr hostdev_mgr,
goto cleanup;
}
- /* Again 4 loops; mark all devices as inactive before reset
+ /* Here are 4 loops; mark all devices as inactive before reset
* them and reset all the devices before re-attach.
* Attach mac and port profile parameters to devices
*/
+
+ /* Loop 1: delete the copy of the dev from pcidevs if it's used by
+ * other domain. Or delete it from activePCIHostDevs if it had
+ * been used by this domain.
+ */
i = 0;
while (i < virPCIDeviceListCount(pcidevs)) {
virPCIDevicePtr dev = virPCIDeviceListGet(pcidevs, i);
virPCIDevicePtr activeDev = NULL;
- /* delete the copy of the dev from pcidevs if it's used by
- * other domain. Or delete it from activePCIHostDevs if it had
- * been used by this domain.
- */
activeDev = virPCIDeviceListFind(hostdev_mgr->activePCIHostdevs, dev);
if (activeDev) {
const char *usedby_drvname;
@@ -814,14 +820,27 @@ virHostdevReAttachPCIDevices(virHostdevManagerPtr hostdev_mgr,
* pcidevs, but has been removed from activePCIHostdevs.
*/
- /*
- * For SRIOV net host devices, unset mac and port profile before
- * reset and reattach device
+ /* Loop 2: before reset and reattach device,
+ * unset mac and port profile for SRIOV net host devices used by this domain
*/
The comments were formatted as a sentence, it would be nice to keep it
that way.
/*
* Loop 2: For SRIOV net host devices used by this domain, unset mac
* and port profile before reset and reattach device
*/
Is this OK?- for (i = 0; i < nhostdevs; i++)
- virHostdevNetConfigRestore(hostdevs[i], hostdev_mgr->stateDir,
- oldStateDir);
+ for (i = 0; i < nhostdevs; i++) {
+ virDomainHostdevDefPtr hostdev = hostdevs[i];
+ if (virHostdevIsPCINetDevice(hostdev)) {
+ virDomainHostdevSubsysPCIPtr pcisrc = &hostdev->source.subsys.u.pci;
+ virPCIDevicePtr dev = NULL;
+ dev = virPCIDeviceNew(pcisrc->addr.domain, pcisrc->addr.bus,
+ pcisrc->addr.slot, pcisrc->addr.function);
+
The daemon will crash if this function returns NULL. There should be
check for that, but it shouldn't be fatal, so we can clean up as much
as possible (see Loop 3 for example on how to handle that).
if (dev) {
if (virPCIDeviceListFind(pcidevs, dev)) {
virHostdevNetConfigRestore(hostdev, hostdev_mgr->stateDir,
oldStateDir);
}
} else {
virErrorPtr err = virGetLastError();
VIR_ERROR(_("Failed to new PCI device: %s"),
err ? err->message : _("unknown error"));
virResetError(err);
}
virPCIDeviceFree(dev);
+ if (virPCIDeviceListFind(pcidevs, dev)) {
+ virHostdevNetConfigRestore(hostdev, hostdev_mgr->stateDir,
+ oldStateDir);
+ }
+ virPCIDeviceFree(dev);
+ }
+ }
+
+ /* Loop 3: reset pci device used by this domain */
for (i = 0; i < virPCIDeviceListCount(pcidevs); i++) {
virPCIDevicePtr dev = virPCIDeviceListGet(pcidevs, i);
@@ -834,6 +853,9 @@ virHostdevReAttachPCIDevices(virHostdevManagerPtr hostdev_mgr,
}
}
+ /* Loop 4: reattach pci devices used by this domain
+ * and steal all the devices from pcidevs
+ */
while (virPCIDeviceListCount(pcidevs) > 0) {
virPCIDevicePtr dev = virPCIDeviceListStealIndex(pcidevs, 0);
virHostdevReattachPCIDevice(dev, hostdev_mgr);
--
1.9.1
Other than that, the patch looks fine.
Martin
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list