The API queries huge page info in the host and reports it back to the caller. This may be handy for management application to decide whether to run a domain with huge pages enabled or not. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- daemon/remote.c | 54 ++++++++++++++++++++++++++++++++++++++++++++ include/libvirt/libvirt.h.in | 6 +++++ src/driver.h | 7 ++++++ src/libvirt.c | 54 ++++++++++++++++++++++++++++++++++++++++++++ src/libvirt_public.syms | 5 ++++ src/remote/remote_driver.c | 48 +++++++++++++++++++++++++++++++++++++++ src/remote/remote_protocol.x | 19 +++++++++++++++- src/remote_protocol-structs | 13 +++++++++++ 8 files changed, 205 insertions(+), 1 deletion(-) diff --git a/daemon/remote.c b/daemon/remote.c index 34c96c9..95ff973 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -6115,6 +6115,60 @@ remoteDispatchDomainGetTime(virNetServerPtr server ATTRIBUTE_UNUSED, return rv; } +static int remoteDispatchNodeHugeTlb(virNetServerPtr server ATTRIBUTE_UNUSED, + virNetServerClientPtr client, + virNetMessagePtr msg ATTRIBUTE_UNUSED, + virNetMessageErrorPtr rerr, + remote_node_huge_tlb_args *args, + remote_node_huge_tlb_ret *ret) +{ + virTypedParameterPtr params = NULL; + int nparams = 0; + int rv = -1; + struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client); + + if (!priv->conn) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); + goto cleanup; + } + + if (args->nparams > REMOTE_NODE_MEMORY_PARAMETERS_MAX) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large")); + goto cleanup; + } + if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0) + goto cleanup; + nparams = args->nparams; + + if (virNodeHugeTLB(priv->conn, args->type, params, &nparams, args->flags) < 0) + goto cleanup; + + /* In this case, we need to send back the number of parameters + * supported + */ + if (args->nparams == 0) { + ret->nparams = nparams; + goto success; + } + + if (remoteSerializeTypedParameters(params, nparams, + &ret->params.params_val, + &ret->params.params_len, + args->flags) < 0) + goto cleanup; + + + success: + rv = 0; + + cleanup: + if (rv < 0) + virNetMessageSaveError(rerr); + virTypedParamsFree(params, nparams); + return rv; +} + + /*----- Helpers. -----*/ /* get_nonnull_domain and get_nonnull_network turn an on-wire diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in index 260c971..cb1cad9 100644 --- a/include/libvirt/libvirt.h.in +++ b/include/libvirt/libvirt.h.in @@ -858,6 +858,12 @@ int virNodeSetMemoryParameters(virConnectPtr conn, int nparams, unsigned int flags); +int virNodeHugeTLB(virConnectPtr conn, + int type, + virTypedParameterPtr params, + int *nparams, + unsigned int flags); + /* * node CPU map */ diff --git a/src/driver.h b/src/driver.h index 5ac89d6..4acde08 100644 --- a/src/driver.h +++ b/src/driver.h @@ -1070,6 +1070,12 @@ typedef int virTypedParameterPtr params, int nparams, unsigned int flags); +typedef int +(*virDrvNodeHugeTLB)(virConnectPtr conn, + int type, + virTypedParameterPtr params, + int *nparams, + unsigned int flags); typedef int (*virDrvNodeGetCPUMap)(virConnectPtr conn, @@ -1376,6 +1382,7 @@ struct _virDriver { virDrvDomainGetMetadata domainGetMetadata; virDrvNodeGetMemoryParameters nodeGetMemoryParameters; virDrvNodeSetMemoryParameters nodeSetMemoryParameters; + virDrvNodeHugeTLB nodeHugeTLB; virDrvNodeGetCPUMap nodeGetCPUMap; virDrvDomainFSTrim domainFSTrim; virDrvDomainSendProcessSignal domainSendProcessSignal; diff --git a/src/libvirt.c b/src/libvirt.c index f01b6dd..de20b0c 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -7561,6 +7561,60 @@ virNodeSetMemoryParameters(virConnectPtr conn, /** + * virNodeHugeTLB: + * @conn: pointer to the hypervisor connection + * @type: type + * @params: pointer to memory parameter object + * (return value, allocated by the caller) + * @nparams: pointer to number of memory parameters; input and output + * @flags: extra flags; not used yet, so callers should always pass 0 + * + * Get information about host's huge pages. On input, @nparams + * gives the size of the @params array; on output, @nparams gives + * how many slots were filled with parameter information, which + * might be less but will not exceed the input value. + * + * As a special case, calling with @params as NULL and @nparams + * as 0 on input will cause @nparams on output to contain the + * number of parameters supported by the hypervisor. The caller + * should then allocate @params array, i.e. + * (sizeof(@virTypedParameter) * @nparams) bytes and call the API + * again. See virDomainGetMemoryParameters() for an equivalent + * usage example. + * + * Returns 0 in case of success, and -1 in case of failure. + */ +int +virNodeHugeTLB(virConnectPtr conn, + int type, + virTypedParameterPtr params, + int *nparams, + unsigned int flags) +{ + VIR_DEBUG("conn=%p, type=%d, params=%p, nparams=%p, flags=%x", + conn, type, params, nparams, flags); + + virResetLastError(); + + virCheckNonNullArgGoto(nparams, error); + virCheckNonNegativeArgGoto(*nparams, error); + + if (conn->driver->nodeHugeTLB) { + int ret; + ret = conn->driver->nodeHugeTLB(conn, type, params, nparams, flags); + if (ret < 0) + goto error; + return ret; + } + virReportUnsupportedError(); + + error: + virDispatchError(conn); + return -1; +} + + +/** * virDomainGetSchedulerType: * @domain: pointer to domain object * @nparams: pointer to number of scheduler parameters, can be NULL diff --git a/src/libvirt_public.syms b/src/libvirt_public.syms index cce6bdf..0339e70 100644 --- a/src/libvirt_public.syms +++ b/src/libvirt_public.syms @@ -658,5 +658,10 @@ LIBVIRT_1.2.5 { virDomainSetTime; } LIBVIRT_1.2.3; +LIBVIRT_1.2.6 { + global: + virNodeHugeTLB; +} LIBVIRT_1.2.5; + # .... define new API here using predicted next version number .... diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index 85fe597..9136b78 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -7469,6 +7469,53 @@ remoteDomainGetTime(virDomainPtr dom, } +static int +remoteNodeHugeTLB(virConnectPtr conn, + int type, + virTypedParameterPtr params, + int *nparams, + unsigned int flags) +{ + int rv = -1; + remote_node_huge_tlb_args args; + remote_node_huge_tlb_ret ret; + struct private_data *priv = conn->privateData; + + remoteDriverLock(priv); + + args.type = type; + args.nparams = *nparams; + args.flags = flags; + + memset(&ret, 0, sizeof(ret)); + if (call(conn, priv, 0, REMOTE_PROC_NODE_HUGE_TLB, + (xdrproc_t) xdr_remote_node_huge_tlb_args, (char *) &args, + (xdrproc_t) xdr_remote_node_huge_tlb_ret, (char *) &ret) == -1) + goto done; + + if (*nparams == 0) { + *nparams = ret.nparams; + rv = 0; + goto cleanup; + } + + if (remoteDeserializeTypedParameters(ret.params.params_val, + ret.params.params_len, + REMOTE_NODE_MEMORY_PARAMETERS_MAX, + ¶ms, + nparams) < 0) + goto cleanup; + + rv = 0; + + cleanup: + xdr_free((xdrproc_t) xdr_remote_node_huge_tlb_ret, (char *) &ret); + done: + remoteDriverUnlock(priv); + return rv; +} + + /* get_nonnull_domain and get_nonnull_network turn an on-wire * (name, uuid) pair into virDomainPtr or virNetworkPtr object. * These can return NULL if underlying memory allocations fail, @@ -7805,6 +7852,7 @@ static virDriver remote_driver = { .domainFSThaw = remoteDomainFSThaw, /* 1.2.5 */ .domainGetTime = remoteDomainGetTime, /* 1.2.5 */ .domainSetTime = remoteDomainSetTime, /* 1.2.5 */ + .nodeHugeTLB = remoteNodeHugeTLB, /* 1.2.6 */ }; static virNetworkDriver network_driver = { diff --git a/src/remote/remote_protocol.x b/src/remote/remote_protocol.x index 1f9d583..4d6cc15 100644 --- a/src/remote/remote_protocol.x +++ b/src/remote/remote_protocol.x @@ -2836,6 +2836,17 @@ struct remote_node_get_memory_parameters_ret { int nparams; }; +struct remote_node_huge_tlb_args { + int type; + int nparams; + unsigned int flags; +}; + +struct remote_node_huge_tlb_ret { + remote_typed_param params<REMOTE_NODE_MEMORY_PARAMETERS_MAX>; + int nparams; +}; + struct remote_node_get_cpu_map_args { int need_map; int need_online; @@ -5338,5 +5349,11 @@ enum remote_procedure { * @generate: both * @acl: domain:set_time */ - REMOTE_PROC_DOMAIN_SET_TIME = 338 + REMOTE_PROC_DOMAIN_SET_TIME = 338, + + /** + * @generate: none + * @acl: connect:read + */ + REMOTE_PROC_NODE_HUGE_TLB = 339 }; diff --git a/src/remote_protocol-structs b/src/remote_protocol-structs index 5b22049..e3fcdea 100644 --- a/src/remote_protocol-structs +++ b/src/remote_protocol-structs @@ -2271,6 +2271,18 @@ struct remote_node_get_memory_parameters_ret { } params; int nparams; }; +struct remote_node_huge_tlb_args { + int type; + int nparams; + u_int flags; +}; +struct remote_node_huge_tlb_ret { + struct { + u_int params_len; + remote_typed_param * params_val; + } params; + int nparams; +}; struct remote_node_get_cpu_map_args { int need_map; int need_online; @@ -2802,4 +2814,5 @@ enum remote_procedure { REMOTE_PROC_DOMAIN_FSTHAW = 336, REMOTE_PROC_DOMAIN_GET_TIME = 337, REMOTE_PROC_DOMAIN_SET_TIME = 338, + REMOTE_PROC_NODE_HUGE_TLB = 339, }; -- 1.9.3

The API is exposed as 'hugepage' command. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- tools/virsh-host.c | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tools/virsh.pod | 7 ++++++ 2 files changed, 69 insertions(+) diff --git a/tools/virsh-host.c b/tools/virsh-host.c index cac6086..e71a341 100644 --- a/tools/virsh-host.c +++ b/tools/virsh-host.c @@ -644,6 +644,62 @@ cmdHostname(vshControl *ctl, const vshCmd *cmd ATTRIBUTE_UNUSED) } /* + * "hugepages" command + */ +static const vshCmdInfo info_hugepages[] = { + {.name = "help", + .data = N_("print hugepages info") + }, + {.name = "desc", + .data = N_("print hugepages info") + }, + {.name = NULL} +}; + +static const vshCmdOptDef opts_hugepages[] = { + {.name = "node", + .type = VSH_OT_INT, + .help = N_("NUMA node") + }, + {.name = NULL} +}; + +static bool +cmdHugepages(vshControl *ctl, const vshCmd *cmd) +{ + virTypedParameterPtr params = NULL; + int nparams = 0; + int node = -1; /* Overall stats */ + unsigned int flags = 0; + bool ret = false; + size_t i; + + if (vshCommandOptInt(cmd, "node", &node) < 0) + return false; + + if (virNodeHugeTLB(ctl->conn, node, NULL, &nparams, flags) < 0) + goto cleanup; + + params = vshCalloc(ctl, nparams, sizeof(*params)); + + if (virNodeHugeTLB(ctl->conn, node, params, &nparams, flags) < 0) + goto cleanup; + + vshPrint(ctl, _("Supported huge page sizes:\n")); + for (i = 0; i < nparams; i++) { + char *str = vshGetTypedParamValue(ctl, ¶ms[i]); + vshPrint(ctl, "\t%-15s\t%s\n", params[i].field, str); + VIR_FREE(str); + } + + ret = 0; + cleanup: + virTypedParamsFree(params, nparams); + return ret; +} + + +/* * "uri" command */ static const vshCmdInfo info_uri[] = { @@ -964,6 +1020,12 @@ const vshCmdDef hostAndHypervisorCmds[] = { .info = info_hostname, .flags = 0 }, + {.name = "hugepages", + .handler = cmdHugepages, + .opts = opts_hugepages, + .info = info_hugepages, + .flags = 0 + }, {.name = "maxvcpus", .handler = cmdMaxvcpus, .opts = opts_maxvcpus, diff --git a/tools/virsh.pod b/tools/virsh.pod index 02671b4..5ba591e 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -279,6 +279,13 @@ Prints the hypervisor canonical URI, can be useful in shell mode. Print the hypervisor hostname. +=item B<hugepages> [I<node>] + +Print information on host huge pages. By default overall statistics are printed. +To narrow down selection to a single NUMA node use pass it's number to I<node>. +As a special case, if NUMA node number equals to value of -1, the aggregated +stats are printed out. + =item B<sysinfo> Print the XML representation of the hypervisor sysinfo, if available. -- 1.9.3

These functions will handle PCIe devices and their link capabilities to query some info about it. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- src/libvirt_private.syms | 3 ++ src/util/virpci.c | 81 +++++++++++++++++++++++++++++++++++++++++++++++- src/util/virpci.h | 8 +++++ 3 files changed, 91 insertions(+), 1 deletion(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 1d97c89..649da16 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1690,6 +1690,8 @@ virPCIDeviceFree; virPCIDeviceGetDriverPathAndName; virPCIDeviceGetIOMMUGroupDev; virPCIDeviceGetIOMMUGroupList; +virPCIDeviceGetLinkCap; +virPCIDeviceGetLinkSta; virPCIDeviceGetManaged; virPCIDeviceGetName; virPCIDeviceGetRemoveSlot; @@ -1698,6 +1700,7 @@ virPCIDeviceGetStubDriver; virPCIDeviceGetUnbindFromStub; virPCIDeviceGetUsedBy; virPCIDeviceIsAssignable; +virPCIDeviceIsPCIExpress; virPCIDeviceListAdd; virPCIDeviceListAddCopy; virPCIDeviceListCount; diff --git a/src/util/virpci.c b/src/util/virpci.c index e0f2344..8ad28b8 100644 --- a/src/util/virpci.c +++ b/src/util/virpci.c @@ -130,7 +130,13 @@ struct _virPCIDeviceList { /* PCIe20 7.8.3 Device Capabilities Register (Offset 04h) */ #define PCI_EXP_DEVCAP 0x4 /* Device capabilities */ -#define PCI_EXP_DEVCAP_FLR (1<<28) /* Function Level Reset */ +#define PCI_EXP_LNKCAP 0xc /* Link Capabilities */ +#define PCI_EXP_LNKSTA 0x12 /* Link Status */ +#define PCI_EXP_DEVCAP_FLR (1<<28) /* Function Level Reset */ +#define PCI_EXP_LNKCAP_SPEED 0x0000f /* Maximum Link Speed */ +#define PCI_EXP_LNKCAP_WIDTH 0x003f0 /* Maximum Link Width */ +#define PCI_EXP_LNKSTA_SPEED 0x000f /* Negotiated Link Speed */ +#define PCI_EXP_LNKSTA_WIDTH 0x03f0 /* Negotiated Link Width */ /* Header type 1 BR12 3.2 PCI-to-PCI Bridge Configuration Space Header Format */ #define PCI_PRIMARY_BUS 0x18 /* BR12 3.2.5.2 Primary bus number */ @@ -2750,3 +2756,76 @@ virPCIGetVirtualFunctionInfo(const char *vf_sysfs_device_path ATTRIBUTE_UNUSED, return -1; } #endif /* __linux__ */ + +int +virPCIDeviceIsPCIExpress(virPCIDevicePtr dev) +{ + int fd; + int ret = -1; + + if ((fd = virPCIDeviceConfigOpen(dev, true)) < 0) + return ret; + + if (virPCIDeviceInit(dev, fd) < 0) + goto cleanup; + + ret = dev->pcie_cap_pos != 0; + + cleanup: + virPCIDeviceConfigClose(dev, fd); + return ret; +} + +int +virPCIDeviceGetLinkCap(virPCIDevicePtr dev, + int *port, + unsigned int *speed, + unsigned int *width) +{ + uint32_t t; + int fd; + int ret = -1; + + if ((fd = virPCIDeviceConfigOpen(dev, true) < 0)) + return ret; + + if (virPCIDeviceInit(dev, fd) < 0) + goto cleanup; + + t = virPCIDeviceRead32(dev, fd, dev->pcie_cap_pos + PCI_EXP_LNKCAP); + + *port = t >> 0x18; + *speed = t & PCI_EXP_LNKCAP_SPEED; + *width = (t & PCI_EXP_LNKCAP_WIDTH) >> 4; + ret = 0; + + cleanup: + virPCIDeviceConfigClose(dev, fd); + return ret; +} + +int +virPCIDeviceGetLinkSta(virPCIDevicePtr dev, + unsigned int *speed, + unsigned int *width) +{ + uint32_t t; + int fd; + int ret = -1; + + if ((fd = virPCIDeviceConfigOpen(dev, true) < 0)) + return ret; + + if (virPCIDeviceInit(dev, fd) < 0) + goto cleanup; + + t = virPCIDeviceRead32(dev, fd, dev->pcie_cap_pos + PCI_EXP_LNKSTA); + + *speed = t & PCI_EXP_LNKSTA_SPEED; + *width = (t & PCI_EXP_LNKSTA_WIDTH) >> 4; + ret = 0; + + cleanup: + virPCIDeviceConfigClose(dev, fd); + return ret; +} diff --git a/src/util/virpci.h b/src/util/virpci.h index 20ffe54..0a608c2 100644 --- a/src/util/virpci.h +++ b/src/util/virpci.h @@ -176,4 +176,12 @@ int virPCIDeviceGetDriverPathAndName(virPCIDevicePtr dev, char **path, char **name); +int virPCIDeviceIsPCIExpress(virPCIDevicePtr dev); +int virPCIDeviceGetLinkCap(virPCIDevicePtr dev, + int *port, + unsigned int *speed, + unsigned int *width); +int virPCIDeviceGetLinkSta(virPCIDevicePtr dev, + unsigned int *speed, + unsigned int *width); #endif /* __VIR_PCI_H__ */ -- 1.9.3

Currently it is not possible to determine the speed of an interface and whether a link is actually detected from the API. Orchestrating platforms want to be able to determine when the link has failed and where multiple speeds may be available which one the interface is actually connected at. This commit introduces an extension to our interface XML (without implementation to interface driver backends): <interface type='ethernet' name='eth0'> <start mode='none'/> <mac address='aa:bb:cc:dd:ee:ff'/> <link speed='1000' state='up'/> <mtu size='1492'/> ... </interface> Where @speed is negotiated link speed in Mbits per second, and state is the current NIC state (can be one of the following: "unknown", "notpresent", "down", "lowerlayerdown","testing", "dormant", "up"). Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- docs/schemas/interface.rng | 27 +++++++++++++++++ src/conf/interface_conf.c | 39 ++++++++++++++++++++++++- src/conf/interface_conf.h | 15 ++++++++++ src/libvirt_private.syms | 2 ++ tests/interfaceschemadata/bridge-no-address.xml | 1 + tests/interfaceschemadata/bridge.xml | 1 + tests/interfaceschemadata/ethernet-dhcp.xml | 1 + 7 files changed, 85 insertions(+), 1 deletion(-) diff --git a/docs/schemas/interface.rng b/docs/schemas/interface.rng index 3984b63..d980ef5 100644 --- a/docs/schemas/interface.rng +++ b/docs/schemas/interface.rng @@ -41,6 +41,7 @@ <attribute name="address"><ref name="macAddr"/></attribute> </element> </optional> + <ref name="link-speed-state"/> <!-- FIXME: Allow (some) ethtool options --> </define> @@ -271,6 +272,32 @@ </element> </define> + <define name="link-speed-state"> + <optional> + <element name="link"> + <optional> + <attribute name="speed"> + <ref name="unsignedInt"/> + </attribute> + </optional> + <optional> + <attribute name="state"> + <choice> + <value>unknown</value> + <value>notpresent</value> + <value>down</value> + <value>lowerlayerdown</value> + <value>testing</value> + <value>dormant</value> + <value>up</value> + </choice> + </attribute> + </optional> + </element> + </optional> + </define> + + <!-- Assignment of addresses to an interface, allowing for different protocols diff --git a/src/conf/interface_conf.c b/src/conf/interface_conf.c index 1f67446..227b9d9 100644 --- a/src/conf/interface_conf.c +++ b/src/conf/interface_conf.c @@ -31,6 +31,7 @@ #include "virxml.h" #include "viruuid.h" #include "virbuffer.h" +#include "virstring.h" #define VIR_FROM_THIS VIR_FROM_INTERFACE @@ -38,6 +39,13 @@ VIR_ENUM_IMPL(virInterface, VIR_INTERFACE_TYPE_LAST, "ethernet", "bridge", "bond", "vlan") +VIR_ENUM_IMPL(virInterfaceState, + VIR_INTERFACE_STATE_LAST, + "" /* value of zero means no state */, + "unknown", "notpresent", + "down", "lowerlayerdown", + "testing", "dormant", "up") + static virInterfaceDefPtr virInterfaceDefParseXML(xmlXPathContextPtr ctxt, int parentIfType); static int @@ -666,7 +674,7 @@ static virInterfaceDefPtr virInterfaceDefParseXML(xmlXPathContextPtr ctxt, int parentIfType) { virInterfaceDefPtr def; - int type; + int type, state; char *tmp; xmlNodePtr cur = ctxt->node; @@ -711,6 +719,26 @@ virInterfaceDefParseXML(xmlXPathContextPtr ctxt, int parentIfType) tmp = virXPathString("string(./mac/@address)", ctxt); if (tmp != NULL) def->mac = tmp; + + tmp = virXPathString("string(./link/@state)", ctxt); + if (tmp) { + if ((state = virInterfaceStateTypeFromString(tmp)) < 0) { + virReportError(VIR_ERR_XML_ERROR, + _("unknown link state: %s"), + tmp); + goto error; + } + def->state = state; + } + + tmp = virXPathString("string(./link/@speed)", ctxt); + if (tmp && virStrToLong_ul(tmp, NULL, 10, &def->speed) < 0) { + virReportError(VIR_ERR_XML_ERROR, + _("Unable to parse link speed: %s"), + tmp); + goto error; + } + if (parentIfType == VIR_INTERFACE_TYPE_LAST) { /* only recognize these in toplevel bond interfaces */ if (virInterfaceDefParseStartMode(def, ctxt) < 0) @@ -1088,6 +1116,15 @@ virInterfaceDefDevFormat(virBufferPtr buf, const virInterfaceDef *def) virInterfaceStartmodeDefFormat(buf, def->startmode); if (def->mac != NULL) virBufferAsprintf(buf, "<mac address='%s'/>\n", def->mac); + if (def->state || def->speed) { + virBufferAddLit(buf, "<link"); + if (def->speed) + virBufferAsprintf(buf, " speed='%lu'", def->speed); + if (def->state) + virBufferAsprintf(buf, " state='%s'", + virInterfaceStateTypeToString(def->state)); + virBufferAddLit(buf, "/>\n"); + } if (def->mtu != 0) virBufferAsprintf(buf, "<mtu size='%d'/>\n", def->mtu); virInterfaceProtocolDefFormat(buf, def); diff --git a/src/conf/interface_conf.h b/src/conf/interface_conf.h index b3c92b2..c8b3f6c 100644 --- a/src/conf/interface_conf.h +++ b/src/conf/interface_conf.h @@ -84,6 +84,19 @@ typedef enum { VIR_INTERFACE_BOND_ARP_ALL, /* validate all */ } virInterfaceBondArpValid; +typedef enum { + VIR_INTERFACE_STATE_UNKNOWN = 1, + VIR_INTERFACE_STATE_NOT_PRESENT, + VIR_INTERFACE_STATE_DOWN, + VIR_INTERFACE_STATE_LOWER_LAYER_DOWN, + VIR_INTERFACE_STATE_TESTING, + VIR_INTERFACE_STATE_DORMANT, + VIR_INTERFACE_STATE_UP, + VIR_INTERFACE_STATE_LAST +} virInterfaceState; + +VIR_ENUM_DECL(virInterfaceState) + struct _virInterfaceDef; /* forward declaration required for bridge/bond */ typedef struct _virInterfaceBridgeDef virInterfaceBridgeDef; @@ -146,6 +159,8 @@ struct _virInterfaceDef { char *name; /* interface name */ unsigned int mtu; /* maximum transmit size in byte */ char *mac; /* MAC address */ + virInterfaceState state; /* link state */ + unsigned long speed; /* link speed in Mb per second */ virInterfaceStartMode startmode; /* how it is started */ diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 649da16..5e908d6 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -495,6 +495,8 @@ virInterfaceObjListFree; virInterfaceObjLock; virInterfaceObjUnlock; virInterfaceRemove; +virInterfaceStateTypeFromString; +virInterfaceStateTypeToString; # conf/netdev_bandwidth_conf.h diff --git a/tests/interfaceschemadata/bridge-no-address.xml b/tests/interfaceschemadata/bridge-no-address.xml index 7757534..68b8c94 100644 --- a/tests/interfaceschemadata/bridge-no-address.xml +++ b/tests/interfaceschemadata/bridge-no-address.xml @@ -4,6 +4,7 @@ <bridge stp='off'> <interface type='ethernet' name='eth0'> <mac address='ab:bb:cc:dd:ee:ff'/> + <link speed='1000' state='up'/> </interface> <interface type='ethernet' name='eth1'> </interface> diff --git a/tests/interfaceschemadata/bridge.xml b/tests/interfaceschemadata/bridge.xml index 2535edf..c865116 100644 --- a/tests/interfaceschemadata/bridge.xml +++ b/tests/interfaceschemadata/bridge.xml @@ -7,6 +7,7 @@ <bridge stp='off' delay='0.01'> <interface type='ethernet' name='eth0'> <mac address='ab:bb:cc:dd:ee:ff'/> + <link speed='10'/> </interface> <interface type='ethernet' name='eth1'> </interface> diff --git a/tests/interfaceschemadata/ethernet-dhcp.xml b/tests/interfaceschemadata/ethernet-dhcp.xml index fe969df..c124372 100644 --- a/tests/interfaceschemadata/ethernet-dhcp.xml +++ b/tests/interfaceschemadata/ethernet-dhcp.xml @@ -1,6 +1,7 @@ <interface type='ethernet' name='eth0'> <start mode='none'/> <mac address='aa:bb:cc:dd:ee:ff'/> + <link state='down'/> <mtu size='1492'/> <protocol family='ipv4'> <dhcp peerdns='no'/> -- 1.9.3

In previous commit the interface XML is prepared for exporting information on NIC link speed and state. This commit implement actual logic for getting such info and writing it into XML. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- src/interface/interface_backend_udev.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/src/interface/interface_backend_udev.c b/src/interface/interface_backend_udev.c index c5353ea..c8940ce 100644 --- a/src/interface/interface_backend_udev.c +++ b/src/interface/interface_backend_udev.c @@ -1036,6 +1036,8 @@ udevGetIfaceDef(struct udev *udev, const char *name) const char *mtu_str; char *vlan_parent_dev = NULL; const char *devtype; + const char *operstate; + const char *link_speed; /* Allocate our interface definition structure */ if (VIR_ALLOC(ifacedef) < 0) @@ -1059,6 +1061,23 @@ udevGetIfaceDef(struct udev *udev, const char *name) udev_device_get_sysattr_value(dev, "address")) < 0) goto error; + operstate = udev_device_get_sysattr_value(dev, "operstate"); + if ((ifacedef->state = virInterfaceStateTypeFromString(operstate)) <= 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Unable to parse operstate: %s"), + operstate); + goto error; + } + + link_speed = udev_device_get_sysattr_value(dev, "speed"); + if (link_speed && + virStrToLong_ul(link_speed, NULL, 10, &ifacedef->speed) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Unable to parse speed: %s"), + link_speed); + goto error; + } + /* MTU */ mtu_str = udev_device_get_sysattr_value(dev, "mtu"); if (virStrToLong_ui(mtu_str, NULL, 10, &mtu) < 0) { -- 1.9.3

There's no need to check for these two host capabilities on each device attach or detach. It's sufficient to check them on the daemon start and then just query them from virCaps when needed. Moreover, this way it's fairly simple to expose them in capabilities XML. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- src/conf/capabilities.c | 46 ++++++++++++++++++++++++++++++++++++++++ src/conf/capabilities.h | 11 ++++++++++ src/libvirt_private.syms | 5 +++++ src/lxc/lxc_conf.c | 3 +++ src/nodeinfo.c | 15 +++++++++++++ src/nodeinfo.h | 1 + src/parallels/parallels_driver.c | 3 +++ src/phyp/phyp_driver.c | 3 +++ src/qemu/qemu_capabilities.c | 3 +++ src/qemu/qemu_driver.c | 16 ++++++++++++-- src/qemu/qemu_hostdev.c | 21 ++++++++++++++---- src/uml/uml_conf.c | 3 +++ src/vbox/vbox_tmpl.c | 3 +++ 13 files changed, 127 insertions(+), 6 deletions(-) diff --git a/src/conf/capabilities.c b/src/conf/capabilities.c index cf474d7..9561ba3 100644 --- a/src/conf/capabilities.c +++ b/src/conf/capabilities.c @@ -1068,3 +1068,49 @@ virCapabilitiesGetCpusForNodemask(virCapsPtr caps, return ret; } + + +int +virCapabilitiesGetKVMLegacy(virCapsPtr caps, + bool *legacy) +{ + if (!caps) + return -1; + + *legacy = caps->host.legacyKVMPassthrough; + return 0; +} + +int +virCapabilitiesSetKVMLegacy(virCapsPtr caps, + bool legacy) +{ + if (!caps) + return -1; + + caps->host.legacyKVMPassthrough = legacy; + return 0; +} + + +int +virCapabilitiesGetVFIO(virCapsPtr caps, + bool *vfio) +{ + if (!caps) + return -1; + + *vfio = caps->host.VFIOPassthrough; + return 0; +} + +int +virCapabilitiesSetVFIO(virCapsPtr caps, + bool vfio) +{ + if (!caps) + return -1; + + caps->host.VFIOPassthrough = vfio; + return 0; +} diff --git a/src/conf/capabilities.h b/src/conf/capabilities.h index ba99e1a..935cd14 100644 --- a/src/conf/capabilities.h +++ b/src/conf/capabilities.h @@ -144,6 +144,9 @@ struct _virCapsHost { virCPUDefPtr cpu; unsigned char host_uuid[VIR_UUID_BUFLEN]; + + bool legacyKVMPassthrough; + bool VFIOPassthrough; }; typedef int (*virDomainDefNamespaceParse)(xmlDocPtr, xmlNodePtr, @@ -274,4 +277,12 @@ virCapabilitiesFormatXML(virCapsPtr caps); virBitmapPtr virCapabilitiesGetCpusForNodemask(virCapsPtr caps, virBitmapPtr nodemask); +int virCapabilitiesGetKVMLegacy(virCapsPtr caps, + bool *legacy); +int virCapabilitiesSetKVMLegacy(virCapsPtr caps, + bool legacy); +int virCapabilitiesGetVFIO(virCapsPtr caps, + bool *vfio); +int virCapabilitiesSetVFIO(virCapsPtr caps, + bool vfio); #endif /* __VIR_CAPABILITIES_H */ diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 9a942ec..95f1599 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -58,9 +58,13 @@ virCapabilitiesFormatXML; virCapabilitiesFreeMachines; virCapabilitiesFreeNUMAInfo; virCapabilitiesGetCpusForNodemask; +virCapabilitiesGetKVMLegacy; +virCapabilitiesGetVFIO; virCapabilitiesHostSecModelAddBaseLabel; virCapabilitiesNew; virCapabilitiesSetHostCPU; +virCapabilitiesSetKVMLegacy; +virCapabilitiesSetVFIO; # conf/cpu_conf.h @@ -866,6 +870,7 @@ virLockManagerRelease; # nodeinfo.h +nodeCapsInitKVM; nodeCapsInitNUMA; nodeGetCellsFreeMemory; nodeGetCPUBitmap; diff --git a/src/lxc/lxc_conf.c b/src/lxc/lxc_conf.c index a35a5e0..9589290 100644 --- a/src/lxc/lxc_conf.c +++ b/src/lxc/lxc_conf.c @@ -82,6 +82,9 @@ virCapsPtr virLXCDriverCapsInit(virLXCDriverPtr driver) VIR_WARN("Failed to query host NUMA topology, disabling NUMA capabilities"); } + if (nodeCapsInitKVM(caps) < 0) + VIR_WARN("Failed to query KVM info"); + if (virNodeSuspendGetTargetMask(&caps->host.powerMgmt) < 0) VIR_WARN("Failed to get host power management capabilities"); diff --git a/src/nodeinfo.c b/src/nodeinfo.c index be63a30..d5ce75b 100644 --- a/src/nodeinfo.c +++ b/src/nodeinfo.c @@ -52,6 +52,7 @@ #include "virtypedparam.h" #include "virstring.h" #include "virnuma.h" +#include "virhostdev.h" #define VIR_FROM_THIS VIR_FROM_NONE @@ -1821,6 +1822,20 @@ nodeCapsInitNUMA(virCapsPtr caps) int +nodeCapsInitKVM(virCapsPtr caps) +{ + bool legacy = virHostdevHostSupportsPassthroughLegacy(); + bool vfio = virHostdevHostSupportsPassthroughVFIO(); + + if (virCapabilitiesSetKVMLegacy(caps, legacy) < 0 || + virCapabilitiesSetVFIO(caps, vfio) < 0) + return -1; + + return 0; +} + + +int nodeGetCellsFreeMemory(unsigned long long *freeMems, int startCell, int maxCells) diff --git a/src/nodeinfo.h b/src/nodeinfo.h index 625de39..52d74fc 100644 --- a/src/nodeinfo.h +++ b/src/nodeinfo.h @@ -28,6 +28,7 @@ int nodeGetInfo(virNodeInfoPtr nodeinfo); int nodeCapsInitNUMA(virCapsPtr caps); +int nodeCapsInitKVM(virCapsPtr caps); int nodeGetCPUStats(int cpuNum, virNodeCPUStatsPtr params, diff --git a/src/parallels/parallels_driver.c b/src/parallels/parallels_driver.c index ab59599..9bd7bef 100644 --- a/src/parallels/parallels_driver.c +++ b/src/parallels/parallels_driver.c @@ -126,6 +126,9 @@ parallelsBuildCapabilities(void) if (nodeCapsInitNUMA(caps) < 0) goto error; + if (nodeCapsInitKVM(caps) < 0) + VIR_WARN("Failed to query KVM info"); + if ((guest = virCapabilitiesAddGuest(caps, "hvm", VIR_ARCH_X86_64, "parallels", diff --git a/src/phyp/phyp_driver.c b/src/phyp/phyp_driver.c index 508e4c0..43e3aea 100644 --- a/src/phyp/phyp_driver.c +++ b/src/phyp/phyp_driver.c @@ -305,6 +305,9 @@ phypCapsInit(void) ("Failed to query host NUMA topology, disabling NUMA capabilities"); } + if (nodeCapsInitKVM(caps) < 0) + VIR_WARN("Failed to query KVM info"); + if ((guest = virCapabilitiesAddGuest(caps, "linux", caps->host.arch, diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 08c3d04..89f38e2 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -939,6 +939,9 @@ virCapsPtr virQEMUCapsInit(virQEMUCapsCachePtr cache) VIR_WARN("Failed to query host NUMA topology, disabling NUMA capabilities"); } + if (nodeCapsInitKVM(caps) < 0) + VIR_WARN("Failed to query KVM info"); + if (virQEMUCapsInitCPU(caps, hostarch) < 0) VIR_WARN("Failed to get host CPU"); diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index d8d0e89..9f97b56 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -11313,12 +11313,23 @@ qemuNodeDeviceDetachFlags(virNodeDevicePtr dev, int ret = -1; virNodeDeviceDefPtr def = NULL; char *xml = NULL; - bool legacy = virHostdevHostSupportsPassthroughLegacy(); - bool vfio = virHostdevHostSupportsPassthroughVFIO(); + virCapsPtr caps = NULL; + bool legacy; + bool vfio; virHostdevManagerPtr hostdev_mgr = driver->hostdevMgr; virCheckFlags(0, -1); + if (!(caps = virQEMUDriverGetCapabilities(driver, false))) + goto cleanup; + + if (virCapabilitiesGetKVMLegacy(caps, &legacy) < 0 || + virCapabilitiesGetVFIO(caps, &vfio) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Unable to get info on KVM of VFIO")); + goto cleanup; + } + xml = virNodeDeviceGetXMLDesc(dev, 0); if (!xml) goto cleanup; @@ -11379,6 +11390,7 @@ qemuNodeDeviceDetachFlags(virNodeDevicePtr dev, virPCIDeviceFree(pci); virNodeDeviceDefFree(def); VIR_FREE(xml); + virObjectUnref(caps); return ret; } diff --git a/src/qemu/qemu_hostdev.c b/src/qemu/qemu_hostdev.c index 39dacb2..1b2ae95 100644 --- a/src/qemu/qemu_hostdev.c +++ b/src/qemu/qemu_hostdev.c @@ -84,14 +84,22 @@ qemuUpdateActiveSCSIHostdevs(virQEMUDriverPtr driver, static bool -qemuPrepareHostdevPCICheckSupport(virDomainHostdevDefPtr *hostdevs, +qemuPrepareHostdevPCICheckSupport(virCapsPtr caps, + virDomainHostdevDefPtr *hostdevs, size_t nhostdevs, virQEMUCapsPtr qemuCaps) { - bool supportsPassthroughKVM = virHostdevHostSupportsPassthroughLegacy(); - bool supportsPassthroughVFIO = virHostdevHostSupportsPassthroughVFIO(); + bool supportsPassthroughKVM; + bool supportsPassthroughVFIO; size_t i; + if (virCapabilitiesGetKVMLegacy(caps, &supportsPassthroughKVM) < 0 || + virCapabilitiesGetVFIO(caps, &supportsPassthroughVFIO) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Unable to get info on KVM of VFIO")); + return false; + } + /* assign defaults for hostdev passthrough */ for (i = 0; i < nhostdevs; i++) { virDomainHostdevDefPtr hostdev = hostdevs[i]; @@ -156,15 +164,20 @@ qemuPrepareHostdevPCIDevices(virQEMUDriverPtr driver, unsigned int flags) { int ret = -1; + virCapsPtr caps = NULL; virHostdevManagerPtr hostdev_mgr = driver->hostdevMgr; - if (!qemuPrepareHostdevPCICheckSupport(hostdevs, nhostdevs, qemuCaps)) + if (!(caps = virQEMUDriverGetCapabilities(driver, false))) + goto out; + + if (!qemuPrepareHostdevPCICheckSupport(caps, hostdevs, nhostdevs, qemuCaps)) goto out; ret = virHostdevPreparePCIDevices(hostdev_mgr, QEMU_DRIVER_NAME, name, uuid, hostdevs, nhostdevs, flags); out: + virObjectUnref(caps); return ret; } diff --git a/src/uml/uml_conf.c b/src/uml/uml_conf.c index 464d56d..6fa0546 100644 --- a/src/uml/uml_conf.c +++ b/src/uml/uml_conf.c @@ -70,6 +70,9 @@ virCapsPtr umlCapsInit(void) VIR_WARN("Failed to query host NUMA topology, disabling NUMA capabilities"); } + if (nodeCapsInitKVM(caps) < 0) + VIR_WARN("Failed to query KVM info"); + if (virNodeSuspendGetTargetMask(&caps->host.powerMgmt) < 0) VIR_WARN("Failed to get host power management capabilities"); diff --git a/src/vbox/vbox_tmpl.c b/src/vbox/vbox_tmpl.c index e124e69..5a4e170 100644 --- a/src/vbox/vbox_tmpl.c +++ b/src/vbox/vbox_tmpl.c @@ -888,6 +888,9 @@ static virCapsPtr vboxCapsInit(void) if (nodeCapsInitNUMA(caps) < 0) goto no_memory; + if (nodeCapsInitKVM(caps) < 0) + VIR_WARN("Failed to query KVM info"); + if ((guest = virCapabilitiesAddGuest(caps, "hvm", caps->host.arch, -- 1.9.3

This piece of information may be useful for management application to decide if a domain with a device passthrough can be started on given host or not. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- Notes: I'm not very happy with the element names, but they're the best I could come up with so far. If you have any better suggestion I am all ears. docs/formatcaps.html.in | 8 +++++++- docs/schemas/capability.rng | 12 ++++++++++++ src/conf/capabilities.c | 4 ++++ tests/capabilityschemadata/caps-qemu-kvm.xml | 2 ++ tests/capabilityschemadata/caps-test.xml | 2 ++ tests/capabilityschemadata/caps-test2.xml | 2 ++ tests/capabilityschemadata/caps-test3.xml | 2 ++ tests/xencapsdata/xen-i686-pae-hvm.xml | 2 ++ tests/xencapsdata/xen-i686-pae.xml | 2 ++ tests/xencapsdata/xen-i686.xml | 2 ++ tests/xencapsdata/xen-ia64-be-hvm.xml | 2 ++ tests/xencapsdata/xen-ia64-be.xml | 2 ++ tests/xencapsdata/xen-ia64-hvm.xml | 2 ++ tests/xencapsdata/xen-ia64.xml | 2 ++ tests/xencapsdata/xen-ppc64.xml | 2 ++ tests/xencapsdata/xen-x86_64-hvm.xml | 2 ++ tests/xencapsdata/xen-x86_64.xml | 2 ++ 17 files changed, 51 insertions(+), 1 deletion(-) diff --git a/docs/formatcaps.html.in b/docs/formatcaps.html.in index d060a5b..eb8c905 100644 --- a/docs/formatcaps.html.in +++ b/docs/formatcaps.html.in @@ -35,6 +35,8 @@ BIOS you will see</p> <suspend_disk/> <suspend_hybrid/> <power_management/> + <kvm>true</kvm> + <vfio>true</vfio> </host></span> <!-- xen-3.0-x86_64 --> @@ -78,7 +80,11 @@ BIOS you will see</p> Suspend-to-Disk (S4) and Hybrid-Suspend (a combination of S3 and S4). In case the host does not support any such feature, then an empty <power_management/> - tag will be shown. </p> + tag will be shown. Then, two elements + <code><kvm/></code> and <code><vfio/></code> + expose the fact, whether the host supports legacy device + passthrough with IOMMU cooperation or newer Virtual function + I/O.</p> <p>The second block (in blue) indicates the paravirtualization support of the Xen support, you will see the os_type of xen to indicate a paravirtual kernel, then architecture diff --git a/docs/schemas/capability.rng b/docs/schemas/capability.rng index d2d9776..3b378eb 100644 --- a/docs/schemas/capability.rng +++ b/docs/schemas/capability.rng @@ -48,6 +48,18 @@ <zeroOrMore> <ref name='secmodel'/> </zeroOrMore> + <element name='kvm'> + <choice> + <value>false</value> + <value>true</value> + </choice> + </element> + <element name='vfio'> + <choice> + <value>false</value> + <value>true</value> + </choice> + </element> </element> </define> diff --git a/src/conf/capabilities.c b/src/conf/capabilities.c index 9561ba3..a91f37b 100644 --- a/src/conf/capabilities.c +++ b/src/conf/capabilities.c @@ -901,6 +901,10 @@ virCapabilitiesFormatXML(virCapsPtr caps) virBufferAddLit(&buf, "</secmodel>\n"); } + /* KVM and VFIO features */ + virBufferAsprintf(&buf, "<kvm>%s</kvm>\n", caps->host.legacyKVMPassthrough ? "true" : "false"); + virBufferAsprintf(&buf, "<vfio>%s</vfio>\n", caps->host.VFIOPassthrough ? "true" : "false"); + virBufferAdjustIndent(&buf, -2); virBufferAddLit(&buf, "</host>\n\n"); diff --git a/tests/capabilityschemadata/caps-qemu-kvm.xml b/tests/capabilityschemadata/caps-qemu-kvm.xml index 55faa16..1a7fca9 100644 --- a/tests/capabilityschemadata/caps-qemu-kvm.xml +++ b/tests/capabilityschemadata/caps-qemu-kvm.xml @@ -28,6 +28,8 @@ <baselabel type='kvm'>system_u:system_r:svirt_t:s0</baselabel> <baselabel type='qemu'>system_u:system_r:svirt_tcg_t:s0</baselabel> </secmodel> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/capabilityschemadata/caps-test.xml b/tests/capabilityschemadata/caps-test.xml index 64f9bb6..72bfafc 100644 --- a/tests/capabilityschemadata/caps-test.xml +++ b/tests/capabilityschemadata/caps-test.xml @@ -36,6 +36,8 @@ </cell> </cells> </topology> + <kvm>true</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/capabilityschemadata/caps-test2.xml b/tests/capabilityschemadata/caps-test2.xml index a99c1b8..04c9b09 100644 --- a/tests/capabilityschemadata/caps-test2.xml +++ b/tests/capabilityschemadata/caps-test2.xml @@ -33,6 +33,8 @@ <uri_transport>tcp</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>true</vfio> </host> <guest> diff --git a/tests/capabilityschemadata/caps-test3.xml b/tests/capabilityschemadata/caps-test3.xml index 7e21f85..9374077 100644 --- a/tests/capabilityschemadata/caps-test3.xml +++ b/tests/capabilityschemadata/caps-test3.xml @@ -85,6 +85,8 @@ <baselabel type='kvm'>107:107</baselabel> <baselabel type='qemu'>107:107</baselabel> </secmodel> + <kvm>true</kvm> + <vfio>true</vfio> </host> </capabilities> diff --git a/tests/xencapsdata/xen-i686-pae-hvm.xml b/tests/xencapsdata/xen-i686-pae-hvm.xml index 872e5f6..4f3c720 100644 --- a/tests/xencapsdata/xen-i686-pae-hvm.xml +++ b/tests/xencapsdata/xen-i686-pae-hvm.xml @@ -14,6 +14,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-i686-pae.xml b/tests/xencapsdata/xen-i686-pae.xml index 3dba6eb..f24306e 100644 --- a/tests/xencapsdata/xen-i686-pae.xml +++ b/tests/xencapsdata/xen-i686-pae.xml @@ -14,6 +14,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-i686.xml b/tests/xencapsdata/xen-i686.xml index 22d7685..b4f0a86 100644 --- a/tests/xencapsdata/xen-i686.xml +++ b/tests/xencapsdata/xen-i686.xml @@ -11,6 +11,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-ia64-be-hvm.xml b/tests/xencapsdata/xen-ia64-be-hvm.xml index 222de1d..951dbdd 100644 --- a/tests/xencapsdata/xen-ia64-be-hvm.xml +++ b/tests/xencapsdata/xen-ia64-be-hvm.xml @@ -11,6 +11,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-ia64-be.xml b/tests/xencapsdata/xen-ia64-be.xml index 017816c..0b90fea 100644 --- a/tests/xencapsdata/xen-ia64-be.xml +++ b/tests/xencapsdata/xen-ia64-be.xml @@ -11,6 +11,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-ia64-hvm.xml b/tests/xencapsdata/xen-ia64-hvm.xml index 33c4946..6b1780e 100644 --- a/tests/xencapsdata/xen-ia64-hvm.xml +++ b/tests/xencapsdata/xen-ia64-hvm.xml @@ -11,6 +11,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-ia64.xml b/tests/xencapsdata/xen-ia64.xml index 82ce965..1bb58e9 100644 --- a/tests/xencapsdata/xen-ia64.xml +++ b/tests/xencapsdata/xen-ia64.xml @@ -11,6 +11,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-ppc64.xml b/tests/xencapsdata/xen-ppc64.xml index 91401b9..fa61351 100644 --- a/tests/xencapsdata/xen-ppc64.xml +++ b/tests/xencapsdata/xen-ppc64.xml @@ -11,6 +11,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-x86_64-hvm.xml b/tests/xencapsdata/xen-x86_64-hvm.xml index 8de8cf4..7aece30 100644 --- a/tests/xencapsdata/xen-x86_64-hvm.xml +++ b/tests/xencapsdata/xen-x86_64-hvm.xml @@ -14,6 +14,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> diff --git a/tests/xencapsdata/xen-x86_64.xml b/tests/xencapsdata/xen-x86_64.xml index 0c3279b..3bbd40c 100644 --- a/tests/xencapsdata/xen-x86_64.xml +++ b/tests/xencapsdata/xen-x86_64.xml @@ -14,6 +14,8 @@ <uri_transport>xenmigr</uri_transport> </uri_transports> </migration_features> + <kvm>false</kvm> + <vfio>false</vfio> </host> <guest> -- 1.9.3

A PCI device can be associated with a specific NUMA node. Later, when a guest is pinned to one NUMA node the PCI device can be assigned on different NUMA node. This makes DMA transfers travel across nodes and thus results in suboptimal performance. We should expose the NUMA node locality for PCI devices so management applications can make better decisions. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- Notes: All the machines I have tried this on had only -1 in the numa_node file. From the kernel sources it seems that this is the default, so I'm not printing the <numa/> element into the XML in this case. But I'd like to hear your opinion. docs/formatnode.html.in | 6 +++ docs/schemas/nodedev.rng | 11 ++++++ src/conf/node_device_conf.c | 43 ++++++++++++++++++++++ src/conf/node_device_conf.h | 1 + src/node_device/node_device_udev.c | 7 ++++ tests/nodedevschemadata/pci_1002_71c4.xml | 1 + tests/nodedevschemadata/pci_8086_10c9_sriov_pf.xml | 1 + 7 files changed, 70 insertions(+) diff --git a/docs/formatnode.html.in b/docs/formatnode.html.in index 46ec2bc..3671c8b 100644 --- a/docs/formatnode.html.in +++ b/docs/formatnode.html.in @@ -124,6 +124,12 @@ GigaTransfers per second) and <code>width</code> for the number of lanes used. Since the port can't be negotiated, it's not exposed in <code>./pci-express/link/[@validity='sta']</code>. + <dt><code>numa</code></dt> + <dd> + This optional element contains information on the PCI device + with respect to NUMA. For example, the optional + <code>node</code> attribute tells which NUMA node is the PCI + device associated with. </dd> </dl> </dd> diff --git a/docs/schemas/nodedev.rng b/docs/schemas/nodedev.rng index 79e8fd2..10b716f 100644 --- a/docs/schemas/nodedev.rng +++ b/docs/schemas/nodedev.rng @@ -184,6 +184,17 @@ </zeroOrMore> </element> </optional> + + <optional> + <element name='numa'> + <optional> + <attribute name='node'> + <data type='int'/> + </attribute> + </optional> + </element> + </optional> + </define> <define name='capusbdev'> diff --git a/src/conf/node_device_conf.c b/src/conf/node_device_conf.c index 70634cc..57dccc4 100644 --- a/src/conf/node_device_conf.c +++ b/src/conf/node_device_conf.c @@ -380,6 +380,9 @@ char *virNodeDeviceDefFormat(const virNodeDeviceDef *def) } if (data->pci_dev.flags & VIR_NODE_DEV_CAP_FLAG_PCIE) virPCIEDeviceInfoFormat(&buf, data->pci_dev.pci_express); + if (data->pci_dev.numa_node >= 0) + virBufferAsprintf(&buf, "<numa node='%d'/>\n", + data->pci_dev.numa_node); break; case VIR_NODE_DEV_CAP_USB_DEV: virBufferAsprintf(&buf, "<bus>%d</bus>\n", data->usb_dev.bus); @@ -554,6 +557,41 @@ char *virNodeDeviceDefFormat(const virNodeDeviceDef *def) return NULL; } +/** + * virNodeDevCapsDefParseIntOptional: + * @xpath: XPath to evaluate + * @ctxt: Context + * @value: Where to store parsed value + * @def: Node device which is parsed + * @invalid_error_fmt: error message to print on invalid format + * + * Returns: -1 on error (invalid int format under @xpath) + * 0 if @xpath was not found (@value is untouched) + * 1 on success + */ +static int +virNodeDevCapsDefParseIntOptional(const char *xpath, + xmlXPathContextPtr ctxt, + int *value, + virNodeDeviceDefPtr def, + const char *invalid_error_fmt) +{ + int ret; + int val; + + ret = virXPathInt(xpath, ctxt, &val); + if (ret < -1) { + virReportError(VIR_ERR_INTERNAL_ERROR, + invalid_error_fmt, + def->name); + return -1; + } else if (ret == -1) { + return 0; + } + *value = val; + return 1; +} + static int virNodeDevCapsDefParseULong(const char *xpath, xmlXPathContextPtr ctxt, @@ -1222,6 +1260,11 @@ virNodeDevCapPCIDevParseXML(xmlXPathContextPtr ctxt, goto out; } + if (virNodeDevCapsDefParseIntOptional("number(./numa[1]/@node)", ctxt, + &data->pci_dev.numa_node, def, + _("invalid NUMA node ID supplied for '%s'")) < 0) + goto out; + ret = 0; out: ctxt->node = orignode; diff --git a/src/conf/node_device_conf.h b/src/conf/node_device_conf.h index 563bf6a..6abab5e 100644 --- a/src/conf/node_device_conf.h +++ b/src/conf/node_device_conf.h @@ -142,6 +142,7 @@ struct _virNodeDevCapsDef { size_t nIommuGroupDevices; unsigned int iommuGroupNumber; virPCIEDeviceInfoPtr pci_express; + int numa_node; } pci_dev; struct { unsigned int bus; diff --git a/src/node_device/node_device_udev.c b/src/node_device/node_device_udev.c index 9ed869b..1663edc 100644 --- a/src/node_device/node_device_udev.c +++ b/src/node_device/node_device_udev.c @@ -495,6 +495,13 @@ static int udevProcessPCI(struct udev_device *device, goto out; } + if (udevGetIntSysfsAttr(device, + "numa_node", + &data->pci_dev.numa_node, + 10) == PROPERTY_ERROR) { + goto out; + } + if (!virPCIGetPhysicalFunction(syspath, &data->pci_dev.physical_function)) data->pci_dev.flags |= VIR_NODE_DEV_CAP_FLAG_PCI_PHYSICAL_FUNCTION; diff --git a/tests/nodedevschemadata/pci_1002_71c4.xml b/tests/nodedevschemadata/pci_1002_71c4.xml index 6de09c1..6d5d85b 100644 --- a/tests/nodedevschemadata/pci_1002_71c4.xml +++ b/tests/nodedevschemadata/pci_1002_71c4.xml @@ -8,5 +8,6 @@ <function>0</function> <product id='0x71c4'>M56GL [Mobility FireGL V5200]</product> <vendor id='0x1002'>ATI Technologies Inc</vendor> + <numa node='1'/> </capability> </device> diff --git a/tests/nodedevschemadata/pci_8086_10c9_sriov_pf.xml b/tests/nodedevschemadata/pci_8086_10c9_sriov_pf.xml index eff8932..6e1dc86 100644 --- a/tests/nodedevschemadata/pci_8086_10c9_sriov_pf.xml +++ b/tests/nodedevschemadata/pci_8086_10c9_sriov_pf.xml @@ -12,5 +12,6 @@ <address domain='0x0000' bus='0x02' slot='0x00' function='0x0'/> <address domain='0x0000' bus='0x02' slot='0x00' function='0x1'/> </iommuGroup> + <numa node='0'/> </capability> </device> -- 1.9.3