January 2010 - Devel - libvirt List Archives

[libvirt] Schedule for libvirt-0.7.6
by Daniel Veillard 29 Jan '10

29 Jan '10

So the plan is still to start the feature freeze end of next friday the 22 Jan and try to push 0.7.6 on the 29th. This mean there is still one week to push features :-) Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

2 1

[libvirt] [PATCH] Allow a per-PCI passthrough device permissive attribute
by Chris Lalancette 28 Jan '10

28 Jan '10

Currently there is a global tag to let the administrator turn off system-wide ACS checking when doing PCI device passthrough. However, this is too coarse-grained of an attribute, since it doesn't allow setups where certain guests are trusted while other ones are untrusted. Allow more complicated setups by making the device checking a per-device setting. The more detailed explanation of why this is necessary delves deep into PCIe internals. Ideally we'd like to be able to probe devices and figure out whether it is safe to assign them. In practice, this isn't possible because PCIe allows devices to have "hidden" bridges that software can't discover. If you were to have two devices assigned to two different domains behind one of these hidden bridges, they could do P2P traffic and bypass all of the VT-d/IOMMU checks. The next thing we could try to do is to have a whitelist of devices that we know to be safe. For instance, instead of a "hidden" bridge, PCI devices can multiplex functions instead, which causes all traffic to head to an upstream bridge before P2P can take place. Additionally, some "hidden" PCI bridges may have ACS on-board. In both of these cases it's safe to passthrough the device(s), since they can't P2P without the IOMMU getting involved. However, even if we did have a whitelist, I think we still need a permissive attribute. For one thing, the whitelist will always be out of date with respect to new hardware, so we'd need to allow administrators to temporarily override the whitelist restriction until a new version of the whitelist came out. Also, we want to support the case where the administrator knows it is safe to assign possibly unsafe devices to a domain he trusts. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- docs/schemas/domain.rng | 6 ++++++ src/conf/domain_conf.c | 14 +++++++++++--- src/conf/domain_conf.h | 1 + src/libvirt_private.syms | 2 ++ src/qemu/qemu_driver.c | 1 + src/util/pci.c | 13 ++++++++++++- src/util/pci.h | 3 +++ 7 files changed, 36 insertions(+), 4 deletions(-) diff --git a/docs/schemas/domain.rng b/docs/schemas/domain.rng index 827ff6f..9f90f4d 100644 --- a/docs/schemas/domain.rng +++ b/docs/schemas/domain.rng @@ -1175,6 +1175,12 @@ <value>no</value> </choice> </attribute> + <attribute name="permissive"> + <choice> + <value>yes</value> + <value>no</value> + </choice> + </attribute> </optional> <group> <element name="source"> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index e548d1d..899967d 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -2929,7 +2929,7 @@ virDomainHostdevDefParseXML(virConnectPtr conn, xmlNodePtr cur; virDomainHostdevDefPtr def; - char *mode, *type = NULL, *managed = NULL; + char *mode, *type = NULL, *managed = NULL, *permissive = NULL; if (VIR_ALLOC(def) < 0) { virReportOOMError(conn); @@ -2968,6 +2968,13 @@ virDomainHostdevDefParseXML(virConnectPtr conn, VIR_FREE(managed); } + permissive = virXMLPropString(node, "permissive"); + if (permissive != NULL) { + if (STREQ(permissive, "yes")) + def->permissive = 1; + VIR_FREE(permissive); + } + cur = node->children; while (cur != NULL) { if (cur->type == XML_ELEMENT_NODE) { @@ -5243,8 +5250,9 @@ virDomainHostdevDefFormat(virConnectPtr conn, return -1; } - virBufferVSprintf(buf, " <hostdev mode='%s' type='%s' managed='%s'>\n", - mode, type, def->managed ? "yes" : "no"); + virBufferVSprintf(buf, " <hostdev mode='%s' type='%s' managed='%s' permissive='%s'>\n", + mode, type, def->managed ? "yes" : "no", + def->permissive ? "yes" : "no"); virBufferAddLit(buf, " <source>\n"); if (def->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB) { diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 7be090d..57ab4b4 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -475,6 +475,7 @@ typedef virDomainHostdevDef *virDomainHostdevDefPtr; struct _virDomainHostdevDef { int mode; /* enum virDomainHostdevMode */ unsigned int managed : 1; + unsigned int permissive : 1; union { struct { int type; /* enum virDomainHostdevBusType */ diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index d56fb7d..49e222e 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -441,6 +441,8 @@ pciWaitForDeviceCleanup; pciResetDevice; pciDeviceSetManaged; pciDeviceGetManaged; +pciDeviceSetPermissive; +pciDeviceGetPermissive; pciDeviceListNew; pciDeviceListFree; pciDeviceListAdd; diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 5bf6743..9848f1c 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2137,6 +2137,7 @@ qemuGetPciHostDeviceList(virConnectPtr conn, } pciDeviceSetManaged(dev, hostdev->managed); + pciDeviceSetPermissive(dev, hostdev->permissive); } return list; diff --git a/src/util/pci.c b/src/util/pci.c index 0274806..42a0be6 100644 --- a/src/util/pci.c +++ b/src/util/pci.c @@ -64,6 +64,7 @@ struct _pciDevice { unsigned has_flr : 1; unsigned has_pm_reset : 1; unsigned managed : 1; + unsigned permissive : 1; }; struct _pciDeviceList { @@ -1076,6 +1077,16 @@ unsigned pciDeviceGetManaged(pciDevice *dev) return dev->managed; } +void pciDeviceSetPermissive(pciDevice *dev, unsigned permissive) +{ + dev->permissive = !!permissive; +} + +unsigned pciDeviceGetPermissive(pciDevice *dev) +{ + return dev->permissive; +} + pciDeviceList * pciDeviceListNew(virConnectPtr conn) { @@ -1353,7 +1364,7 @@ int pciDeviceIsAssignable(virConnectPtr conn, return 0; if (ret) { - if (!strict_acs_check) { + if (!strict_acs_check || dev->permissive) { VIR_DEBUG("%s %s: strict ACS check disabled; device assignment allowed", dev->id, dev->name); } else { diff --git a/src/util/pci.h b/src/util/pci.h index e6ab137..3f98374 100644 --- a/src/util/pci.h +++ b/src/util/pci.h @@ -44,6 +44,9 @@ int pciResetDevice (virConnectPtr conn, void pciDeviceSetManaged(pciDevice *dev, unsigned managed); unsigned pciDeviceGetManaged(pciDevice *dev); +void pciDeviceSetPermissive(pciDevice *dev, + unsigned permissive); +unsigned pciDeviceGetPermissive(pciDevice *dev); pciDeviceList *pciDeviceListNew (virConnectPtr conn); void pciDeviceListFree (virConnectPtr conn, -- 1.6.6

2 3

[libvirt] [PATCH 1/3] Clarify why some controllers don't generate -device string in QEMU driver
by Matthew Booth 28 Jan '10

28 Jan '10

The QEMU driver contained code to generate a -device string for piix4-ide, but wasn't using it. This change removes this string generation. It also adds a comment explaining why IDE and FDC controllers don't generate -device strings. The change also generates an error if a sata controller is specified for a QEMU domain, as this isn't supported. * src/qemu/qemu_conf.c: Remove VIR_DOMAIN_CONTROLLER_TYPE_IDE handler in qemuBuildControllerDevStr(). Ignore IDE and FDC controllers. Error if SATA controller is discovered. Add comments. --- src/qemu/qemu_conf.c | 26 ++++++++++++++++++-------- 1 files changed, 18 insertions(+), 8 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index f4a6c08..3b7793f 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2121,11 +2121,8 @@ qemuBuildControllerDevStr(virDomainControllerDefPtr def) virBufferVSprintf(&buf, ",id=scsi%d", def->idx); break; + /* We always get an IDE controller, whether we want it or not. */ case VIR_DOMAIN_CONTROLLER_TYPE_IDE: - virBufferAddLit(&buf, "piix4-ide"); - virBufferVSprintf(&buf, ",id=ide%d", def->idx); - break; - default: goto error; } @@ -3141,16 +3138,29 @@ int qemudBuildCommandLine(virConnectPtr conn, if (qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) { for (i = 0 ; i < def->ncontrollers ; i++) { - char *scsi; - if (def->controllers[i]->type != VIR_DOMAIN_CONTROLLER_TYPE_SCSI) + virDomainControllerDefPtr cont = def->controllers[i]; + + /* We don't add an explicit IDE or FD controller because the + * provided PIIX4 device already includes one. It isn't possible to + * remove the PIIX4. */ + if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE || + cont->type == VIR_DOMAIN_CONTROLLER_TYPE_FDC) continue; + /* QEMU doesn't implement a SATA driver */ + if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_SATA) { + qemudReportError(conn, NULL, NULL, VIR_ERR_NO_SUPPORT, + "%s", _("SATA is not supported with this QEMU binary")); + goto error; + } + ADD_ARG_LIT("-device"); - if (!(scsi = qemuBuildControllerDevStr(def->controllers[i]))) + char *devstr; + if (!(devstr = qemuBuildControllerDevStr(def->controllers[i]))) goto no_memory; - ADD_ARG(scsi); + ADD_ARG(devstr); } } -- 1.6.6

1 2

[libvirt] Libvirt guide draft published on libvirt.org
by David Jorm 28 Jan '10

28 Jan '10

I am a tech writer who recently joined the Red Hat team. I have been tasked with assisting in the improvement of libvirt documentation where possible and co-ordinating the development of the libvirt Application Development Guide. The guide was previously in the hands of Dani Coulson, who has since left Red Hat. She got it to a draft stage with a skeletal structure, but as far as I can tell nothing ever reached a publishable state. I've picked up where she left off and re-built the latest guide from the DocBook XML in git. It is now up at: http://libvirt.org/guide/ If you look in the guide, you will notice an awful lot of "TBD" stubs. Contributions to fill these would be greatly appreciated - please email them to me directly. I will chase up with the people who were originally nominated as the responsible parties to try and get some content to flesh out the guide. I don't have a lot of spare temporal bandwidth at the moment, but if there are any docs-related BZs or libvirt issues, feel free to push them my way and i'll do what I can. I think i've made every mistake possible so far in submitting patches, so I know the process by virtue of what-not-to-do. Thanks David

2 3

[libvirt] Libvirt/LXC creation example
by Avi Weit 28 Jan '10

28 Jan '10

Hi, I am looking for a simple example for creating a Linux container (LXC) with a network connection that can be accessed from the outside world. I first tried to set up a bridge and link it with the interface eth0 configured on host via: brctl addif virbr0 eth0 but each time I did that - I lost the connection to the host and could not even ping it. The connection was resumed only when removing the host interface of eth0 from the bridge: brctl delif virbr0 eth0 How can I define the network interfaces to be configured inside the container? How can I setup container IPs? Any full XML example would be appreciated. Thanks - Avi

1 0

[libvirt] [PATCH 0/1] Clean up udev init
by David Allan 28 Jan '10

28 Jan '10

The following patch contains cleanup for the udev init error handling and Matthias' fix for the case in which priv is NULL when shutdown is called. Dave

3 3

[libvirt] [PATCH 0/7] Update hotplug code to use device_add
by Daniel P. Berrange 27 Jan '10

27 Jan '10

I previously changed the command line syntax to use -device for starting guests. The key reason for this was to allow unique names and PCI addresses to be assigned to all devices. I forgot todo the same for hotplug :-( This series is not yet complete, but it goes some of the way towards fixing the hotplug code to use 'device_add' for creating devices. The main remaining problem is that I'm nt assigning aliases to devices, to they all get the name of '(null)'. Obviously I'll fix that before submitting again... Daniel

1 7

[libvirt] [PATCH] support XEN_SYSCTL_INTERFACE_VERSION
by Jim Fehlig 27 Jan '10

27 Jan '10

xen-unstable c/s 20762 bumped XEN_SYSCTL_INTERFACE_VERSION to 7. I don't see how the interface change affects libvirt, other than failing xenHypervisorInit() since version 7 is not tried. The attached patch accommodates the upcoming Xen 4.0 release by checking for XEN_SYSCTL_INTERFACE_VERSION 7. If found, it sets XEN_DOMCTL_INTERFACE_VERSION to 6, which is also new to Xen 4.0. Regards, Jim

2 2

[libvirt] [PATCH 0/1] Clean up udev init
by David Allan 27 Jan '10

27 Jan '10

Matthias' patch made me realize that I didn't write the udev init code as cleanly as I'd like, and the error handling was starting to get a little messy. The attached patch cleans it up. There should be no functional change. Dave

3 3

[libvirt] [PATCH] Add missing sata controller type to domain.rng
by Matthew Booth 27 Jan '10

27 Jan '10

* docs/schemas/domain.rng: Add sata controller type --- docs/schemas/domain.rng | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/docs/schemas/domain.rng b/docs/schemas/domain.rng index 49b57eb..827ff6f 100644 --- a/docs/schemas/domain.rng +++ b/docs/schemas/domain.rng @@ -529,6 +529,7 @@ <value>fdc</value> <value>ide</value> <value>scsi</value> + <value>sata</value> </choice> </attribute> </optional> -- 1.6.6

2 1