- Devel - libvirt List Archives

[libvirt] [PATCH 1/2 V2] cleanup: make nlComm commonly available
by Stefan Berger 22 Jun '11

22 Jun '11

In a first cleanup step, make nlComm from macvtap.c commonly available for other code to use. Since nlComm uses Linux-specific structures as parameters it's prototype is only visible on Linux. v2: - making nlComm visible also on non-Linux platforms Signed-off-by: Stefan Berger <stefanb(a)linux.vnet.ibm.com> --- src/Makefile.am | 1 src/libvirt_private.syms | 4 + src/util/macvtap.c | 98 ------------------------------ src/util/netlink.c | 152 +++++++++++++++++++++++++++++++++++++++++++++++ src/util/netlink.h | 18 +++++ 5 files changed, 176 insertions(+), 97 deletions(-) Index: libvirt-acl/src/Makefile.am =================================================================== --- libvirt-acl.orig/src/Makefile.am +++ libvirt-acl/src/Makefile.am @@ -63,6 +63,7 @@ UTIL_SOURCES = \ util/logging.c util/logging.h \ util/macvtap.c util/macvtap.h \ util/memory.c util/memory.h \ + util/netlink.c util/netlink.h \ util/pci.c util/pci.h \ util/processinfo.c util/processinfo.h \ util/hostusb.c util/hostusb.h \ Index: libvirt-acl/src/util/macvtap.c =================================================================== --- libvirt-acl.orig/src/util/macvtap.c +++ libvirt-acl/src/util/macvtap.c @@ -39,12 +39,8 @@ # include <sys/ioctl.h> # include <linux/if.h> -# include <linux/netlink.h> -# include <linux/rtnetlink.h> # include <linux/if_tun.h> -# include <netlink/msg.h> - /* Older kernels lacked this enum value. */ # if !HAVE_DECL_MACVLAN_MODE_PASSTHRU # define MACVLAN_MODE_PASSTHRU 8 @@ -64,6 +60,7 @@ # include "virterror_internal.h" # include "uuid.h" # include "files.h" +# include "netlink.h" # define VIR_FROM_THIS VIR_FROM_NET @@ -79,8 +76,6 @@ # define NLMSGBUF_SIZE 256 # define RATTBUF_SIZE 64 -# define NETLINK_ACK_TIMEOUT_S 2 - # define STATUS_POLL_TIMEOUT_USEC (10 * MICROSEC_PER_SEC) # define STATUS_POLL_INTERVL_USEC (MICROSEC_PER_SEC / 8) @@ -96,97 +91,6 @@ enum virVirtualPortOp { }; -/** - * nlComm: - * @nlmsg: pointer to netlink message - * @respbuf: pointer to pointer where response buffer will be allocated - * @respbuflen: pointer to integer holding the size of the response buffer - * on return of the function. - * @nl_pid: the pid of the process to talk to, i.e., pid = 0 for kernel - * - * Send the given message to the netlink layer and receive response. - * Returns 0 on success, -1 on error. In case of error, no response - * buffer will be returned. - */ -static -int nlComm(struct nl_msg *nl_msg, - unsigned char **respbuf, unsigned int *respbuflen, - int nl_pid) -{ - int rc = 0; - struct sockaddr_nl nladdr = { - .nl_family = AF_NETLINK, - .nl_pid = nl_pid, - .nl_groups = 0, - }; - ssize_t nbytes; - struct timeval tv = { - .tv_sec = NETLINK_ACK_TIMEOUT_S, - }; - fd_set readfds; - int fd; - int n; - struct nlmsghdr *nlmsg = nlmsg_hdr(nl_msg); - struct nl_handle *nlhandle = nl_handle_alloc(); - - if (!nlhandle) { - virReportSystemError(errno, - "%s", _("cannot allocate nlhandle for netlink")); - return -1; - } - - if (nl_connect(nlhandle, NETLINK_ROUTE) < 0) { - virReportSystemError(errno, - "%s", _("cannot connect to netlink socket")); - rc = -1; - goto err_exit; - } - - nlmsg_set_dst(nl_msg, &nladdr); - - nlmsg->nlmsg_pid = getpid(); - - nbytes = nl_send_auto_complete(nlhandle, nl_msg); - if (nbytes < 0) { - virReportSystemError(errno, - "%s", _("cannot send to netlink socket")); - rc = -1; - goto err_exit; - } - - fd = nl_socket_get_fd(nlhandle); - - FD_ZERO(&readfds); - FD_SET(fd, &readfds); - - n = select(fd + 1, &readfds, NULL, NULL, &tv); - if (n <= 0) { - if (n < 0) - virReportSystemError(errno, "%s", - _("error in select call")); - if (n == 0) - virReportSystemError(ETIMEDOUT, "%s", - _("no valid netlink response was received")); - rc = -1; - goto err_exit; - } - - *respbuflen = nl_recv(nlhandle, &nladdr, respbuf, NULL); - if (*respbuflen <= 0) { - virReportSystemError(errno, - "%s", _("nl_recv failed")); - rc = -1; - } -err_exit: - if (rc == -1) { - VIR_FREE(*respbuf); - *respbuf = NULL; - *respbuflen = 0; - } - - nl_handle_destroy(nlhandle); - return rc; -} # if WITH_MACVTAP Index: libvirt-acl/src/util/netlink.c =================================================================== --- /dev/null +++ libvirt-acl/src/util/netlink.c @@ -0,0 +1,152 @@ +/* + * Copyright (C) 2010-2011 Red Hat, Inc. + * Copyright (C) 2010 IBM Corporation + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + * Authors: + * Stefan Berger <stefanb(a)us.ibm.com> + * + * Notes: + * netlink: http://lovezutto.googlepages.com/netlink.pdf + * iproute2 package + * + */ + +#include <config.h> + +#include <errno.h> +#include <unistd.h> +#include <sys/types.h> + +#include "netlink.h" +#include "memory.h" +#include "virterror_internal.h" + +#define VIR_FROM_THIS VIR_FROM_NET + +# define netlinkError(code, ...) \ + virReportErrorHelper(VIR_FROM_NET, code, __FILE__, \ + __FUNCTION__, __LINE__, __VA_ARGS__) + +#define NETLINK_ACK_TIMEOUT_S 2 + +/** + * nlComm: + * @nlmsg: pointer to netlink message + * @respbuf: pointer to pointer where response buffer will be allocated + * @respbuflen: pointer to integer holding the size of the response buffer + * on return of the function. + * @nl_pid: the pid of the process to talk to, i.e., pid = 0 for kernel + * + * Send the given message to the netlink layer and receive response. + * Returns 0 on success, -1 on error. In case of error, no response + * buffer will be returned. + */ +#if __linux__ +int nlComm(struct nl_msg *nl_msg, + unsigned char **respbuf, unsigned int *respbuflen, + int nl_pid) +{ + int rc = 0; + struct sockaddr_nl nladdr = { + .nl_family = AF_NETLINK, + .nl_pid = nl_pid, + .nl_groups = 0, + }; + ssize_t nbytes; + struct timeval tv = { + .tv_sec = NETLINK_ACK_TIMEOUT_S, + }; + fd_set readfds; + int fd; + int n; + struct nlmsghdr *nlmsg = nlmsg_hdr(nl_msg); + struct nl_handle *nlhandle = nl_handle_alloc(); + + if (!nlhandle) { + virReportSystemError(errno, + "%s", _("cannot allocate nlhandle for netlink")); + return -1; + } + + if (nl_connect(nlhandle, NETLINK_ROUTE) < 0) { + virReportSystemError(errno, + "%s", _("cannot connect to netlink socket")); + rc = -1; + goto err_exit; + } + + nlmsg_set_dst(nl_msg, &nladdr); + + nlmsg->nlmsg_pid = getpid(); + + nbytes = nl_send_auto_complete(nlhandle, nl_msg); + if (nbytes < 0) { + virReportSystemError(errno, + "%s", _("cannot send to netlink socket")); + rc = -1; + goto err_exit; + } + + fd = nl_socket_get_fd(nlhandle); + + FD_ZERO(&readfds); + FD_SET(fd, &readfds); + + n = select(fd + 1, &readfds, NULL, NULL, &tv); + if (n <= 0) { + if (n < 0) + virReportSystemError(errno, "%s", + _("error in select call")); + if (n == 0) + virReportSystemError(ETIMEDOUT, "%s", + _("no valid netlink response was received")); + rc = -1; + goto err_exit; + } + + *respbuflen = nl_recv(nlhandle, &nladdr, respbuf, NULL); + if (*respbuflen <= 0) { + virReportSystemError(errno, + "%s", _("nl_recv failed")); + rc = -1; + } +err_exit: + if (rc == -1) { + VIR_FREE(*respbuf); + *respbuf = NULL; + *respbuflen = 0; + } + + nl_handle_destroy(nlhandle); + return rc; +} + +#else + +int nlComm(struct nl_msg *nl_msg ATTRIBUTE_UNUSED, + unsigned char **respbuf ATTRIBUTE_UNUSED, + unsigned int *respbuflen ATTRIBUTE_UNUSED, + int nl_pid ATTRIBUTE_UNUSED) +{ + netlinkError(VIR_ERR_INTERNAL_ERROR, "%s", + _("nlComm is not supported on non-linux platforms")); + return -1; +} + + +#endif /* __linux__ */ + Index: libvirt-acl/src/util/netlink.h =================================================================== --- /dev/null +++ libvirt-acl/src/util/netlink.h @@ -0,0 +1,18 @@ +#ifndef __VIR_NETLINK_H__ +# define __VIR_NETLINK_H__ + +# if __linux__ + +# include <netlink/msg.h> + +# else + +struct nl_msg; + +#endif /* __linux__ */ + +int nlComm(struct nl_msg *nl_msg, + unsigned char **respbuf, unsigned int *respbuflen, + int nl_pid); + +#endif /* __VIR_NETLINK_H__ */ Index: libvirt-acl/src/libvirt_private.syms =================================================================== --- libvirt-acl.orig/src/libvirt_private.syms +++ libvirt-acl/src/libvirt_private.syms @@ -668,6 +668,10 @@ virResizeN; virShrinkN; +#netlink.h +nlComm; + + # network.h virSocketAddrBroadcast; virSocketAddrBroadcastByPrefix;

2 1

[libvirt] [PATCH] Replace @SYSCONFDIR@ with ::SYSCONFDIR:: to avoid syntax-check warning
by Daniel P. Berrange 22 Jun '11

22 Jun '11

make syntax-check complains if any Makefile.am uses the old style @FOO@ for replacement, rather than $(FOO). Unfortunately the tools/Makefile.am intentionally uses the @FOO@ syntax for generating some scripts. Currently these are whitelisted in cfg.mk, but adding more to this whitelist makes the entire check rather worthless. This changes the tools/ directory to use ::FOO:: as the replacement syntax for virt-pki-validate.in and virt-xml-validate.in * tools/Makefile.am, tools/virt-pki-validate.in tools/virt-xml-validate.in: Change @FOO@ to ::FOO:: * cfg.mk: Remove _makefile_at_at_check_exceptions --- cfg.mk | 3 --- tools/Makefile.am | 4 ++-- tools/virt-pki-validate.in | 2 +- tools/virt-xml-validate.in | 2 +- 4 files changed, 4 insertions(+), 7 deletions(-) diff --git a/cfg.mk b/cfg.mk index 5d3ef26..c2230b8 100644 --- a/cfg.mk +++ b/cfg.mk @@ -581,9 +581,6 @@ _autogen: $(srcdir)/autogen.sh ./config.status -# Exempt @...@ uses of these symbols. -_makefile_at_at_check_exceptions = ' && !/(SCHEMA|SYSCONF)DIR/' - # regenerate HACKING as part of the syntax-check syntax-check: $(top_srcdir)/HACKING diff --git a/tools/Makefile.am b/tools/Makefile.am index 826674a..7a1a5c6 100644 --- a/tools/Makefile.am +++ b/tools/Makefile.am @@ -23,14 +23,14 @@ dist_man1_MANS = virt-xml-validate.1 virt-pki-validate.1 virsh.1 virt-xml-validate: virt-xml-validate.in Makefile - $(AM_V_GEN)sed -e 's,@SCHEMADIR@,$(pkgdatadir)/schemas,' < $< > $@ \ + $(AM_V_GEN)sed -e 's,::SCHEMADIR::,$(pkgdatadir)/schemas,' < $< > $@ \ || (rm $@ && exit 1) && chmod +x $@ virt-xml-validate.1: virt-xml-validate.in $(AM_V_GEN)$(POD2MAN) $< $(srcdir)/$@ virt-pki-validate: virt-pki-validate.in Makefile - $(AM_V_GEN)sed -e 's,@SYSCONFDIR@,$(sysconfdir),' < $< > $@ \ + $(AM_V_GEN)sed -e 's,::SYSCONFDIR::,$(sysconfdir),' < $< > $@ \ || (rm $@ && exit 1) && chmod +x $@ virt-pki-validate.1: virt-pki-validate.in diff --git a/tools/virt-pki-validate.in b/tools/virt-pki-validate.in index 01825d1..f429d77 100755 --- a/tools/virt-pki-validate.in +++ b/tools/virt-pki-validate.in @@ -25,7 +25,7 @@ echo Found "$CERTOOL" # # Check the directory structure # -SYSCONFDIR="@SYSCONFDIR@" +SYSCONFDIR="::SYSCONFDIR::" PKI="$SYSCONFDIR/pki" if [ ! -d "$PKI" ] then diff --git a/tools/virt-xml-validate.in b/tools/virt-xml-validate.in index 1959261..33f6c0f 100644 --- a/tools/virt-xml-validate.in +++ b/tools/virt-xml-validate.in @@ -57,7 +57,7 @@ if [ -z "$TYPE" ]; then esac fi -SCHEMA="@SCHEMADIR@/${TYPE}.rng" +SCHEMA="::SCHEMADIR::/${TYPE}.rng" if [ ! -f "$SCHEMA" ]; then echo "$0: schema $SCHEMA does not exist" -- 1.7.4.4

2 3

[libvirt] [PATCH] Remove macvtap dependancy on domain configuration
by Daniel P. Berrange 22 Jun '11

22 Jun '11

Files under src/util must not depend on src/conf Solve the macvtap problem by moving the definition of macvtap modes from domain_conf.h into macvtap.h * src/util/macvtap.c, src/util/macvtap.h: Add enum for macvtap modes * src/conf/domain_conf.c, src/conf/domain_conf.h: Remove enum for macvtap modes --- src/conf/domain_conf.c | 17 ++++++----------- src/conf/domain_conf.h | 15 +-------------- src/util/macvtap.c | 40 +++++++++++++++++----------------------- src/util/macvtap.h | 14 +++++++++++++- 4 files changed, 37 insertions(+), 49 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index b5c0f83..c528960 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -447,12 +447,6 @@ VIR_ENUM_IMPL(virDomainSeclabel, VIR_DOMAIN_SECLABEL_LAST, "dynamic", "static") -VIR_ENUM_IMPL(virDomainNetdevMacvtap, VIR_DOMAIN_NETDEV_MACVTAP_MODE_LAST, - "vepa", - "private", - "bridge", - "passthrough") - VIR_ENUM_IMPL(virVirtualPort, VIR_VIRTUALPORT_TYPE_LAST, "none", "802.1Qbg", @@ -1156,7 +1150,7 @@ static virDomainObjPtr virDomainObjNew(virCapsPtr caps) } if (caps->privateDataAllocFunc && - !(domain->privateData = (caps->privateDataAllocFunc)())) { + !(domain->privateData = (caps->privateDataAllocFunc)(caps->privateDataOpaque))) { virReportOOMError(); VIR_FREE(domain); return NULL; @@ -2974,14 +2968,14 @@ virDomainNetDefParseXML(virCapsPtr caps, if (mode != NULL) { int m; - if ((m = virDomainNetdevMacvtapTypeFromString(mode)) < 0) { + if ((m = virMacvtapModeTypeFromString(mode)) < 0) { virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Unkown mode has been specified")); goto error; } def->data.direct.mode = m; } else - def->data.direct.mode = VIR_DOMAIN_NETDEV_MACVTAP_MODE_VEPA; + def->data.direct.mode = VIR_MACVTAP_MODE_VEPA; if (virtPortParsed) def->data.direct.virtPortProfile = virtPort; @@ -6737,12 +6731,13 @@ static virDomainObjPtr virDomainObjParseXML(virCapsPtr caps, char *str = virXMLPropString(nodes[i], "flag"); if (str) { int flag = virDomainTaintTypeFromString(str); - VIR_FREE(str); if (flag < 0) { virDomainReportError(VIR_ERR_INTERNAL_ERROR, _("Unknown taint flag %s"), str); + VIR_FREE(str); goto error; } + VIR_FREE(str); virDomainObjTaint(obj, flag); } } @@ -8634,7 +8629,7 @@ virDomainNetDefFormat(virBufferPtr buf, virBufferEscapeString(buf, " <source dev='%s'", def->data.direct.linkdev); virBufferAsprintf(buf, " mode='%s'", - virDomainNetdevMacvtapTypeToString(def->data.direct.mode)); + virMacvtapModeTypeToString(def->data.direct.mode)); virBufferAddLit(buf, "/>\n"); virVirtualPortProfileFormat(buf, &def->data.direct.virtPortProfile, " "); diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 994ff91..aa25e36 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -348,17 +348,6 @@ enum virDomainNetVirtioTxModeType { VIR_DOMAIN_NET_VIRTIO_TX_MODE_LAST, }; -/* the mode type for macvtap devices */ -enum virDomainNetdevMacvtapType { - VIR_DOMAIN_NETDEV_MACVTAP_MODE_VEPA, - VIR_DOMAIN_NETDEV_MACVTAP_MODE_PRIVATE, - VIR_DOMAIN_NETDEV_MACVTAP_MODE_BRIDGE, - VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU, - - VIR_DOMAIN_NETDEV_MACVTAP_MODE_LAST, -}; - - /* Stores the virtual network interface configuration */ typedef struct _virDomainNetDef virDomainNetDef; typedef virDomainNetDef *virDomainNetDefPtr; @@ -396,7 +385,7 @@ struct _virDomainNetDef { } internal; struct { char *linkdev; - int mode; + int mode; /* enum virMacvtapMode from util/macvtap.h */ virVirtualPortProfileParams virtPortProfile; } direct; } data; @@ -1615,8 +1604,6 @@ int virDomainStateReasonFromString(virDomainState state, const char *reason); VIR_ENUM_DECL(virDomainSeclabel) VIR_ENUM_DECL(virDomainClockOffset) -VIR_ENUM_DECL(virDomainNetdevMacvtap) - VIR_ENUM_DECL(virDomainTimerName) VIR_ENUM_DECL(virDomainTimerTrack) VIR_ENUM_DECL(virDomainTimerTickpolicy) diff --git a/src/util/macvtap.c b/src/util/macvtap.c index 48c7b0e..4e4ed79 100644 --- a/src/util/macvtap.c +++ b/src/util/macvtap.c @@ -55,12 +55,17 @@ #include "util.h" #include "macvtap.h" +VIR_ENUM_IMPL(virMacvtapMode, VIR_MACVTAP_MODE_LAST, + "vepa", + "private", + "bridge", + "passthrough") + #if WITH_MACVTAP || WITH_VIRTUALPORT # include "memory.h" # include "logging.h" # include "interface.h" -# include "conf/domain_conf.h" # include "virterror_internal.h" # include "uuid.h" # include "files.h" @@ -468,26 +473,6 @@ int openTap(const char *ifname, } -static uint32_t -macvtapModeFromInt(enum virDomainNetdevMacvtapType mode) -{ - switch (mode) { - case VIR_DOMAIN_NETDEV_MACVTAP_MODE_PRIVATE: - return MACVLAN_MODE_PRIVATE; - - case VIR_DOMAIN_NETDEV_MACVTAP_MODE_BRIDGE: - return MACVLAN_MODE_BRIDGE; - - case VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU: - return MACVLAN_MODE_PASSTHRU; - - case VIR_DOMAIN_NETDEV_MACVTAP_MODE_VEPA: - default: - return MACVLAN_MODE_VEPA; - } -} - - /** * configMacvtapTap: * @tapfd: file descriptor of the macvtap tap @@ -643,6 +628,13 @@ restoreMacAddress(const char *linkdev, return 0; } +static const uint32_t modeMap[VIR_MACVTAP_MODE_LAST] = { + [VIR_MACVTAP_MODE_VEPA] = MACVLAN_MODE_VEPA, + [VIR_MACVTAP_MODE_PRIVATE] = MACVLAN_MODE_PRIVATE, + [VIR_MACVTAP_MODE_BRIDGE] = MACVLAN_MODE_BRIDGE, + [VIR_MACVTAP_MODE_PASSTHRU] = MACVLAN_MODE_PASSTHRU, +}; + /** * openMacvtapTap: * Create an instance of a macvtap device and open its tap character @@ -667,7 +659,7 @@ int openMacvtapTap(const char *tgifname, const unsigned char *macaddress, const char *linkdev, - int mode, + enum virMacvtapMode mode, int vnet_hdr, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, @@ -679,10 +671,12 @@ openMacvtapTap(const char *tgifname, int c, rc; char ifname[IFNAMSIZ]; int retries, do_retry = 0; - uint32_t macvtapMode = macvtapModeFromInt(mode); + uint32_t macvtapMode; const char *cr_ifname; int ifindex; + macvtapMode = modeMap[mode]; + *res_ifname = NULL; VIR_DEBUG("%s: VM OPERATION: %s", __FUNCTION__, virVMOperationTypeToString(vmOp)); diff --git a/src/util/macvtap.h b/src/util/macvtap.h index 2843596..1b85989 100644 --- a/src/util/macvtap.h +++ b/src/util/macvtap.h @@ -34,6 +34,17 @@ enum virVirtualPortType { VIR_VIRTUALPORT_TYPE_LAST, }; +/* the mode type for macvtap devices */ +enum virMacvtapMode { + VIR_MACVTAP_MODE_VEPA, + VIR_MACVTAP_MODE_PRIVATE, + VIR_MACVTAP_MODE_BRIDGE, + VIR_MACVTAP_MODE_PASSTHRU, + + VIR_MACVTAP_MODE_LAST, +}; + + # ifdef IFLA_VF_PORT_PROFILE_MAX # define LIBVIRT_IFLA_VF_PORT_PROFILE_MAX IFLA_VF_PORT_PROFILE_MAX # else @@ -78,7 +89,7 @@ enum virVMOperationType { int openMacvtapTap(const char *ifname, const unsigned char *macaddress, const char *linkdev, - int mode, + enum virMacvtapMode mode, int vnet_hdr, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, @@ -110,5 +121,6 @@ int vpDisassociatePortProfileId(const char *macvtap_ifname, VIR_ENUM_DECL(virVirtualPort) VIR_ENUM_DECL(virVMOperation) +VIR_ENUM_DECL(virMacvtapMode) #endif /* __UTIL_MACVTAP_H__ */ -- 1.7.4.4

2 3

[libvirt] [PATCH]: set and get the MAC of an interface
by Gerhard Stenzel 22 Jun '11

22 Jun '11

The following patch renames the function to set the MAC of an interface from ifSetInterfaceMac() to brSetInterfaceMac() and makes it available to other components. It also adds brGetInterfaceMac() to retrieve the MAC. Signed-off-by: Gerhard Stenzel <gerhard.stenzel(a)de.ibm.com> Index: libvirt/src/util/bridge.c =================================================================== --- libvirt.orig/src/util/bridge.c +++ libvirt/src/util/bridge.c @@ -286,8 +286,41 @@ brDeleteInterface(brControl *ctl ATTRIBU } # endif + +/** + * brGetInterfaceMac: + * @ctl: bridge control pointer + * @ifname: interface name to set MTU for + * @macaddr: MAC address (VIR_MAC_BUFLEN in size) + * + * This function gets the @macaddr for a given interface @ifname. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +int brGetInterfaceMac(brControl *ctl, const char *ifname, + unsigned char *macaddr) +{ + struct ifreq ifr; + + if (!ctl || !ifname) + return EINVAL; + + memset(&ifr, 0, sizeof(struct ifreq)); + if (virStrcpyStatic(ifr.ifr_name, ifname) == NULL) + return EINVAL; + + if(ioctl(ctl->fd, SIOCGIFHWADDR, (char *)&ifr) != 0){ + return errno; + } + + memcpy(macaddr, ifr.ifr_ifru.ifru_hwaddr.sa_data, VIR_MAC_BUFLEN); + + return 0; +} + + /** - * ifSetInterfaceMac: + * brSetInterfaceMac: * @ctl: bridge control pointer * @ifname: interface name to set MTU for * @macaddr: MAC address (VIR_MAC_BUFLEN in size) @@ -297,7 +330,7 @@ brDeleteInterface(brControl *ctl ATTRIBU * * Returns 0 in case of success or an errno code in case of failure. */ -static int ifSetInterfaceMac(brControl *ctl, const char *ifname, +int brSetInterfaceMac(brControl *ctl, const char *ifname, const unsigned char *macaddr) { struct ifreq ifr; @@ -521,7 +554,7 @@ brAddTap(brControl *ctl, * seeing the kernel allocate random MAC for the TAP * device before we set our static MAC. */ - if ((errno = ifSetInterfaceMac(ctl, ifr.ifr_name, macaddr))) + if ((errno = brSetInterfaceMac(ctl, ifr.ifr_name, macaddr))) goto error; /* We need to set the interface MTU before adding it * to the bridge, because the bridge will have its Index: libvirt/src/util/bridge.h =================================================================== --- libvirt.orig/src/util/bridge.h +++ libvirt/src/util/bridge.h @@ -106,6 +106,14 @@ int brGetEnableSTP (brContr const char *bridge, int *enable); +int brSetInterfaceMac (brControl *ctl, + const char *ifname, + const unsigned char *macaddr); + +int brGetInterfaceMac (brControl *ctl, + const char *ifname, + unsigned char *macaddr); + # endif /* WITH_BRIDGE */ #endif /* __QEMUD_BRIDGE_H__ */ =================================================================== Best regards, Gerhard Stenzel ------------------------------------------------------------------------------------- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Martin Jetter Geschaeftsfuehrung: Dirk Wittkopp Sitz der Gesellschaft: Boeblingen Registergericht: Amtsgericht Stuttgart, HRB 243294

2 2

[libvirt] [RFC] Introduce virDomainBlockCopy API
by Jiri Denemark 22 Jun '11

22 Jun '11

This API starts asynchronous live copy of a block device (specified by target element of the xml argument) into a new source (which must already exist). The process can be controlled in the same way as migration: monitored with virDomainJobInfo() and canceled using virDomainAbortJob(). I don't particularly like the name (but I wasn't able to come up with a better one) since it doesn't reflect the fact that once a block device is switched to use the new source once copying finishes. In other words, the goal of this API is to update the device to use new source (just like virDomainUpdateDeviceFlags) but before doing so all data is copied from the old source into the new one (unlike the UpdateDevice API). --- include/libvirt/libvirt.h.in | 4 +++ src/driver.h | 6 ++++ src/libvirt.c | 54 ++++++++++++++++++++++++++++++++++++++++++ src/libvirt_public.syms | 1 + 4 files changed, 65 insertions(+), 0 deletions(-) diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in index cb9e8ca..41f6b2e 100644 --- a/include/libvirt/libvirt.h.in +++ b/include/libvirt/libvirt.h.in @@ -2983,6 +2983,10 @@ typedef struct _virTypedParameter virMemoryParameter; */ typedef virMemoryParameter *virMemoryParameterPtr; +int virDomainBlockCopy(virDomainPtr domain, + const char *xml, + unsigned int flags); + #ifdef __cplusplus } #endif diff --git a/src/driver.h b/src/driver.h index 62bbc1d..8731d68 100644 --- a/src/driver.h +++ b/src/driver.h @@ -666,6 +666,11 @@ typedef int virDomainBlockPullInfoPtr info, unsigned int flags); +typedef int + (*virDrvDomainBlockCopy)(virDomainPtr domain, + const char *xml, + unsigned int flags); + /** * _virDriver: * @@ -809,6 +814,7 @@ struct _virDriver { virDrvDomainBlockPullAll domainBlockPullAll; virDrvDomainBlockPullAbort domainBlockPullAbort; virDrvDomainGetBlockPullInfo domainGetBlockPullInfo; + virDrvDomainBlockCopy domainBlockCopy; }; typedef int diff --git a/src/libvirt.c b/src/libvirt.c index c57e0c3..ad5da9b 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -15548,3 +15548,57 @@ error: virDispatchError(dom->conn); return -1; } + +/** + * virDomainBlockCopy: + * @domain: pointer to domain object + * @xml: pointer to XML description of the block device to copy + * @flags: currently unused, use 0 + * + * Starts asynchronous live copy of a block device (specified by target element + * of the xml) into a new source (which must already exist). The process can be + * controlled in the same way as migration: monitored with virDomainJobInfo() + * and canceled using virDomainAbortJob(). + * + * Returns -1 in case of failure, 0 when successful. + */ +int +virDomainBlockCopy(virDomainPtr domain, const char *xml, unsigned int flags) +{ + virConnectPtr conn; + + VIR_DOMAIN_DEBUG(domain, "xml=%s, flags=%d", xml, flags); + + virResetLastError(); + + if (!VIR_IS_CONNECTED_DOMAIN(domain)) { + virLibDomainError(VIR_ERR_INVALID_DOMAIN, __FUNCTION__); + virDispatchError(NULL); + return -1; + } + + if (xml == NULL) { + virLibDomainError(VIR_ERR_INVALID_ARG, __FUNCTION__); + goto error; + } + + if (domain->conn->flags & VIR_CONNECT_RO) { + virLibDomainError(VIR_ERR_OPERATION_DENIED, __FUNCTION__); + goto error; + } + conn = domain->conn; + + if (conn->driver->domainBlockCopy) { + int ret; + ret = conn->driver->domainBlockCopy(domain, xml, flags); + if (ret < 0) + goto error; + return ret; + } + + virLibConnError(VIR_ERR_NO_SUPPORT, __FUNCTION__); + +error: + virDispatchError(domain->conn); + return -1; +} diff --git a/src/libvirt_public.syms b/src/libvirt_public.syms index b2915f6..3028e15 100644 --- a/src/libvirt_public.syms +++ b/src/libvirt_public.syms @@ -452,6 +452,7 @@ LIBVIRT_0.9.2 { LIBVIRT_0.9.3 { global: + virDomainBlockCopy; virDomainBlockPull; virDomainBlockPullAbort; virDomainBlockPullAll; -- 1.7.5.3

4 6

[libvirt] qemu error: "fsdev is not supported by this qemu build."
by Richard W.M. Jones 22 Jun '11

22 Jun '11

When trying to add a 9p filesystem to a guest, I get the qemu error "fsdev is not supported by this qemu build." (when starting the guest). I'm still looking at this. It seems to be something to do with the machine type causing the -fsdev option not to be registered inside qemu. I just wanted to record the problem here in case anyone else has any ideas. Full XML and qemu command line is below. Occurs with: libvirt-0.9.2-2.fc16.x86_64 2:qemu-system-x86-0.14.0-8.fc16.x86_64 Also occurs with the latest qemu & libvirt in Fedora 15, but reportedly not with older versions of qemu. Rich. ---------------------------------------------------------------------- <domain type='kvm'> <name>F14x64</name> <uuid>799ae7ca-23a2-f622-c691-48feebcddbe6</uuid> <memory>1048576</memory> <currentMemory>1048576</currentMemory> <vcpu>1</vcpu> <os> <type arch='x86_64' machine='pc-0.14'>hvm</type> <boot dev='hd'/> </os> <features> <acpi/> <apic/> <pae/> </features> <cpu match='exact'> <model>core2duo</model> <vendor>Intel</vendor> </cpu> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <emulator>/usr/bin/qemu-kvm</emulator> <disk type='block' device='disk'> <driver name='qemu' type='raw'/> <source dev='/dev/vg_pin/F14x64'/> <target dev='hda' bus='ide'/> <address type='drive' controller='0' bus='0' unit='0'/> </disk> <controller type='ide' index='0'> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x1'/> </controller> <controller type='virtio-serial' index='0'> <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> </controller> <filesystem type='mount' accessmode='passthrough'> <source dir='/tmp'/> <target dir='org.kernel.tmp'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/> </filesystem> <interface type='network'> <mac address='52:54:00:18:04:63'/> <source network='default'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </interface> <serial type='pty'> <target port='0'/> </serial> <console type='pty'> <target type='serial' port='0'/> </console> <channel type='unix'> <source mode='bind' path='/tmp/f14socket'/> <target type='virtio' name='org.libguestfs.channel.0'/> <address type='virtio-serial' controller='0' bus='0' port='1'/> </channel> <input type='mouse' bus='ps2'/> <graphics type='vnc' port='-1' autoport='yes' keymap='de'/> <video> <model type='cirrus' vram='9216' heads='1'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </video> <watchdog model='ib700' action='reset'/> <memballoon model='virtio'> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </memballoon> </devices> </domain> ---------------------------------------------------------------------- LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin QEMU_AUDIO_DRV=none /usr/bin/qemu-kvm -S -M pc-0.14 -cpu core2duo -enable-kvm -m 1024 -smp 1,sockets=1,cores=1,threads=1 -name F14x64 -uuid 799ae7ca-23a2-f622-c691-48feebcddbe6 -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/F14x64.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=utc -boot c -device virtio-serial-pci,id=virtio-serial0,bus=pci.0,addr=0x5 -drive file=/dev/vg_pin/F14x64,if=none,id=drive-ide0-0-0,format=raw -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0 -fsdev local,security_model=passthrough,id=fsdev-fs0,path=/tmp -device virtio-9p-pci,id=fs0,fsdev=fsdev-fs0,mount_tag=org.kernel.tmp,bus=pci.0,addr=0x6 -netdev tap,fd=23,id=hostnet0 -device rtl8139,netdev=hostnet0,id=net0,mac=52:54:00:18:04:63,bus=pci.0,addr=0x4 -chardev pty,id=charserial0 -device isa-serial,chardev=charserial0,id=serial0 -chardev socket,id=charchannel0,path=/tmp/f14socket,server,nowait -device virtserialport,bus=virtio-serial0.0,nr=1,chardev=charchannel0,id=channel0,name=org.libguestfs.channel.0 -usb -vnc 127.0.0.1:0 -k de -vga cirrus -device ib700,id=watchdog0 -watchdog-action reset -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 qemu-kvm: -fsdev local,security_model=passthrough,id=fsdev-fs0,path=/tmp: there is no option group "fsdev" fsdev is not supported by this qemu build. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/

2 3

[libvirt] [PATCH] util: Correct the error prompt string
by Osier Yang 22 Jun '11

22 Jun '11

virCommandProcessIO: It's reading from stdout or stderr of child, but not reading. --- src/util/command.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/src/util/command.c b/src/util/command.c index cb682fc..81a2345 100644 --- a/src/util/command.c +++ b/src/util/command.c @@ -1504,7 +1504,9 @@ virCommandProcessIO(virCommandPtr cmd) if (errno != EINTR && errno != EAGAIN) { virReportSystemError(errno, "%s", - _("unable to write to child input")); + (fds[i].fd == outfd) ? + _("unable to read child stdout") : + _("unable to read child stderr")); goto cleanup; } } else if (done == 0) { -- 1.7.4

2 2

[libvirt] [PATCH] qemu: Fix one memory leak
by Osier Yang 22 Jun '11

22 Jun '11

* src/qemu/qemu_domain.c: (qemuDomainAppendLog) Free "message" in "cleanup". --- src/qemu/qemu_domain.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 5fe66ac..fab316f 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -947,5 +947,6 @@ cleanup: if (fd != logFD) VIR_FORCE_CLOSE(fd); + VIR_FREE(message); return ret; } -- 1.7.4

2 2

[libvirt] [PATCH 1/2] cleanup: make nlComm commonly available
by Stefan Berger 22 Jun '11

22 Jun '11

In a first cleanup step, make nlComm from macvtap.c commonly available for other code to use. Since nlComm uses Linux-specific structures as parameters it's prototype is only visible on Linux. Signed-off-by: Stefan Berger <stefanb(a)linux.vnet.ibm.com> --- src/Makefile.am | 1 src/libvirt_private.syms | 4 + src/util/macvtap.c | 98 ---------------------------------- src/util/netlink.c | 135 +++++++++++++++++++++++++++++++++++++++++++++++ src/util/netlink.h | 14 ++++ 5 files changed, 155 insertions(+), 97 deletions(-) Index: libvirt-acl/src/Makefile.am =================================================================== --- libvirt-acl.orig/src/Makefile.am +++ libvirt-acl/src/Makefile.am @@ -63,6 +63,7 @@ UTIL_SOURCES = \ util/logging.c util/logging.h \ util/macvtap.c util/macvtap.h \ util/memory.c util/memory.h \ + util/netlink.c util/netlink.h \ util/pci.c util/pci.h \ util/processinfo.c util/processinfo.h \ util/hostusb.c util/hostusb.h \ Index: libvirt-acl/src/util/macvtap.c =================================================================== --- libvirt-acl.orig/src/util/macvtap.c +++ libvirt-acl/src/util/macvtap.c @@ -39,12 +39,8 @@ # include <sys/ioctl.h> # include <linux/if.h> -# include <linux/netlink.h> -# include <linux/rtnetlink.h> # include <linux/if_tun.h> -# include <netlink/msg.h> - /* Older kernels lacked this enum value. */ # if !HAVE_DECL_MACVLAN_MODE_PASSTHRU # define MACVLAN_MODE_PASSTHRU 8 @@ -64,6 +60,7 @@ # include "virterror_internal.h" # include "uuid.h" # include "files.h" +# include "netlink.h" # define VIR_FROM_THIS VIR_FROM_NET @@ -79,8 +76,6 @@ # define NLMSGBUF_SIZE 256 # define RATTBUF_SIZE 64 -# define NETLINK_ACK_TIMEOUT_S 2 - # define STATUS_POLL_TIMEOUT_USEC (10 * MICROSEC_PER_SEC) # define STATUS_POLL_INTERVL_USEC (MICROSEC_PER_SEC / 8) @@ -96,97 +91,6 @@ enum virVirtualPortOp { }; -/** - * nlComm: - * @nlmsg: pointer to netlink message - * @respbuf: pointer to pointer where response buffer will be allocated - * @respbuflen: pointer to integer holding the size of the response buffer - * on return of the function. - * @nl_pid: the pid of the process to talk to, i.e., pid = 0 for kernel - * - * Send the given message to the netlink layer and receive response. - * Returns 0 on success, -1 on error. In case of error, no response - * buffer will be returned. - */ -static -int nlComm(struct nl_msg *nl_msg, - unsigned char **respbuf, unsigned int *respbuflen, - int nl_pid) -{ - int rc = 0; - struct sockaddr_nl nladdr = { - .nl_family = AF_NETLINK, - .nl_pid = nl_pid, - .nl_groups = 0, - }; - ssize_t nbytes; - struct timeval tv = { - .tv_sec = NETLINK_ACK_TIMEOUT_S, - }; - fd_set readfds; - int fd; - int n; - struct nlmsghdr *nlmsg = nlmsg_hdr(nl_msg); - struct nl_handle *nlhandle = nl_handle_alloc(); - - if (!nlhandle) { - virReportSystemError(errno, - "%s", _("cannot allocate nlhandle for netlink")); - return -1; - } - - if (nl_connect(nlhandle, NETLINK_ROUTE) < 0) { - virReportSystemError(errno, - "%s", _("cannot connect to netlink socket")); - rc = -1; - goto err_exit; - } - - nlmsg_set_dst(nl_msg, &nladdr); - - nlmsg->nlmsg_pid = getpid(); - - nbytes = nl_send_auto_complete(nlhandle, nl_msg); - if (nbytes < 0) { - virReportSystemError(errno, - "%s", _("cannot send to netlink socket")); - rc = -1; - goto err_exit; - } - - fd = nl_socket_get_fd(nlhandle); - - FD_ZERO(&readfds); - FD_SET(fd, &readfds); - - n = select(fd + 1, &readfds, NULL, NULL, &tv); - if (n <= 0) { - if (n < 0) - virReportSystemError(errno, "%s", - _("error in select call")); - if (n == 0) - virReportSystemError(ETIMEDOUT, "%s", - _("no valid netlink response was received")); - rc = -1; - goto err_exit; - } - - *respbuflen = nl_recv(nlhandle, &nladdr, respbuf, NULL); - if (*respbuflen <= 0) { - virReportSystemError(errno, - "%s", _("nl_recv failed")); - rc = -1; - } -err_exit: - if (rc == -1) { - VIR_FREE(*respbuf); - *respbuf = NULL; - *respbuflen = 0; - } - - nl_handle_destroy(nlhandle); - return rc; -} # if WITH_MACVTAP Index: libvirt-acl/src/util/netlink.c =================================================================== --- /dev/null +++ libvirt-acl/src/util/netlink.c @@ -0,0 +1,135 @@ +/* + * Copyright (C) 2010-2011 Red Hat, Inc. + * Copyright (C) 2010 IBM Corporation + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + * Authors: + * Stefan Berger <stefanb(a)us.ibm.com> + * + * Notes: + * netlink: http://lovezutto.googlepages.com/netlink.pdf + * iproute2 package + * + */ + +#include <config.h> + +#include <errno.h> +#include <unistd.h> +#include <sys/types.h> + +#include "netlink.h" +#include "memory.h" +#include "virterror_internal.h" + +#define VIR_FROM_THIS VIR_FROM_NET + +#define NETLINK_ACK_TIMEOUT_S 2 + +/** + * nlComm: + * @nlmsg: pointer to netlink message + * @respbuf: pointer to pointer where response buffer will be allocated + * @respbuflen: pointer to integer holding the size of the response buffer + * on return of the function. + * @nl_pid: the pid of the process to talk to, i.e., pid = 0 for kernel + * + * Send the given message to the netlink layer and receive response. + * Returns 0 on success, -1 on error. In case of error, no response + * buffer will be returned. + */ +#if __linux__ +int nlComm(struct nl_msg *nl_msg, + unsigned char **respbuf, unsigned int *respbuflen, + int nl_pid) +{ + int rc = 0; + struct sockaddr_nl nladdr = { + .nl_family = AF_NETLINK, + .nl_pid = nl_pid, + .nl_groups = 0, + }; + ssize_t nbytes; + struct timeval tv = { + .tv_sec = NETLINK_ACK_TIMEOUT_S, + }; + fd_set readfds; + int fd; + int n; + struct nlmsghdr *nlmsg = nlmsg_hdr(nl_msg); + struct nl_handle *nlhandle = nl_handle_alloc(); + + if (!nlhandle) { + virReportSystemError(errno, + "%s", _("cannot allocate nlhandle for netlink")); + return -1; + } + + if (nl_connect(nlhandle, NETLINK_ROUTE) < 0) { + virReportSystemError(errno, + "%s", _("cannot connect to netlink socket")); + rc = -1; + goto err_exit; + } + + nlmsg_set_dst(nl_msg, &nladdr); + + nlmsg->nlmsg_pid = getpid(); + + nbytes = nl_send_auto_complete(nlhandle, nl_msg); + if (nbytes < 0) { + virReportSystemError(errno, + "%s", _("cannot send to netlink socket")); + rc = -1; + goto err_exit; + } + + fd = nl_socket_get_fd(nlhandle); + + FD_ZERO(&readfds); + FD_SET(fd, &readfds); + + n = select(fd + 1, &readfds, NULL, NULL, &tv); + if (n <= 0) { + if (n < 0) + virReportSystemError(errno, "%s", + _("error in select call")); + if (n == 0) + virReportSystemError(ETIMEDOUT, "%s", + _("no valid netlink response was received")); + rc = -1; + goto err_exit; + } + + *respbuflen = nl_recv(nlhandle, &nladdr, respbuf, NULL); + if (*respbuflen <= 0) { + virReportSystemError(errno, + "%s", _("nl_recv failed")); + rc = -1; + } +err_exit: + if (rc == -1) { + VIR_FREE(*respbuf); + *respbuf = NULL; + *respbuflen = 0; + } + + nl_handle_destroy(nlhandle); + return rc; +} + +#endif /* __linux__ */ + Index: libvirt-acl/src/util/netlink.h =================================================================== --- /dev/null +++ libvirt-acl/src/util/netlink.h @@ -0,0 +1,14 @@ +#ifndef __VIR_NETLINK_H__ +# define __VIR_NETLINK_H__ + +# if __linux__ + +# include <netlink/msg.h> + +int nlComm(struct nl_msg *nl_msg, + unsigned char **respbuf, unsigned int *respbuflen, + int nl_pid); + +#endif /* __linux__ */ + +#endif /* __VIR_NETLINK_H__ */ Index: libvirt-acl/src/libvirt_private.syms =================================================================== --- libvirt-acl.orig/src/libvirt_private.syms +++ libvirt-acl/src/libvirt_private.syms @@ -668,6 +668,10 @@ virResizeN; virShrinkN; +#netlink.h +nlComm; + + # network.h virSocketAddrBroadcast; virSocketAddrBroadcastByPrefix;

3 3

[libvirt] minor doc generation glitch
by Eric Blake 22 Jun '11

22 Jun '11

Looking at http://libvirt.org/html/libvirt-libvirt.html#virDomainShutoffReason as an example, I see several places where there are spurious "*" in the generated documentation. It looks like anywhere we have: typedef enum { VALUE = 0; /* Multi-line * comment */ } virName; that the python generator is not stripping out the '^ *\*?' regex for continued comment lines. My python is a bit too weak to quickly determine a fix, so I'm throwing this out for anyone else who wants an easy bug to fix. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

2 5

[libvirt] [PATCH v2] qemu: domain I/O asynchronous handling
by Michal Privoznik 22 Jun '11

22 Jun '11

For virtio disks and interfaces, qemu allows users to enable or disable ioeventfd feature. This means, qemu can execute domain code, while another thread waits for I/O event. Basically, in some cases it is win, in some loss. This feature is available via 'ioeventfd' attribute in disk and interface <driver> element. It accepts 'on' and 'off'. Leaving this attribute out defaults to hypervisor decision. --- diff to v1: -reverted to 'ioeventfd' attribute: https://www.redhat.com/archives/libvir-list/2011-June/msg00712.html docs/formatdomain.html.in | 34 +++++++++++++- docs/schemas/domain.rng | 14 ++++++ src/conf/domain_conf.c | 49 +++++++++++++++++++- src/conf/domain_conf.h | 11 ++++ src/libvirt_private.syms | 2 + src/qemu/qemu_capabilities.c | 3 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 13 +++++ tests/qemuhelptest.c | 3 +- .../qemuxml2argv-disk-ioeventfd.args | 11 ++++ .../qemuxml2argv-disk-ioeventfd.xml | 50 ++++++++++++++++++++ tests/qemuxml2argvtest.c | 4 ++ 12 files changed, 191 insertions(+), 4 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index ab39417..39e1a85 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -785,7 +785,7 @@ </disk> ... <disk type='network'> - <driver name="qemu" type="raw" io="threads"/> + <driver name="qemu" type="raw" io="threads" ioeventfd="on"/> <source protocol="sheepdog" name="image_name"> <host name="hostname" port="7000"/> </source> @@ -869,6 +869,20 @@ policies on I/O; qemu guests support "threads" and "native". Since 0.8.8 </li> + <li> + The optional <code>ioeventfd</code> attribute allows users to + set <a href='https://patchwork.kernel.org/patch/43390/'> + domain I/O asynchronous handling</a> for disk device. + The default is left to the discretion of the hypervisor. + Accepted values are "on" and "off". Enabling this allows + qemu to execute VM while a separate thread handles I/O. + Typically guests experiencing high system CPU utilization + during I/O will benefit from this. On the other hand, + on overloaded host it could increase guest I/O latency. + Since 0.9.3 (QEMU and KVM only) + In general you should leave this option alone, unless you + are very certain you know what you are doing. + </li> </ul> </dd> <dt><code>boot</code></dt> @@ -1649,7 +1663,7 @@ qemu-kvm -net nic,model=? /dev/null <source network='default'/> <target dev='vnet1'/> <model type='virtio'/> - <driver name='vhost' txmode='iothread'/> + <driver name='vhost' txmode='iothread' ioeventfd='on'/> </interface> </devices> ...</pre> @@ -1700,6 +1714,22 @@ qemu-kvm -net nic,model=? /dev/null In general you should leave this option alone, unless you are very certain you know what you are doing. </dd> + <dt><code>ioeventfd</code></dt> + <dd> + This optional attribute allows users to set + <a href='https://patchwork.kernel.org/patch/43390/'> + domain I/O asynchronous handling</a> for interface device. + The default is left to the discretion of the hypervisor. + Accepted values are "on" and "off". Enabling this allows + qemu to execute VM while a separate thread handles I/O. + Typically guests experiencing high system CPU utilization + during I/O will benefit from this. On the other hand, + on overloaded host it could increase guest I/O latency. + Since 0.9.3 (QEMU and KVM only) + + In general you should leave this option alone, unless you + are very certain you know what you are doing. + </dd> </dl> <h5><a name="elementsNICSTargetOverride">Overriding the target element</a></h5> diff --git a/docs/schemas/domain.rng b/docs/schemas/domain.rng index 6de024e..891662d 100644 --- a/docs/schemas/domain.rng +++ b/docs/schemas/domain.rng @@ -778,6 +778,9 @@ <optional> <ref name="driverIO"/> </optional> + <optional> + <ref name="ioeventfd"/> + </optional> <empty/> </element> </define> @@ -817,6 +820,14 @@ </choice> </attribute> </define> + <define name="ioeventfd"> + <attribute name="ioeventfd"> + <choice> + <value>on</value> + <value>off</value> + </choice> + </attribute> + </define> <define name="controller"> <element name="controller"> <choice> @@ -1117,6 +1128,9 @@ </choice> </attribute> </optional> + <optional> + <ref name="ioeventfd"/> + </optional> <empty/> </element> </optional> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 5360863..2234857 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -163,6 +163,11 @@ VIR_ENUM_IMPL(virDomainDiskIo, VIR_DOMAIN_DISK_IO_LAST, "default", "native", "threads") +VIR_ENUM_IMPL(virDomainIoEventFd, VIR_DOMAIN_IO_EVENT_FD_LAST, + "default", + "on", + "off") + VIR_ENUM_IMPL(virDomainController, VIR_DOMAIN_CONTROLLER_TYPE_LAST, "ide", @@ -2013,6 +2018,7 @@ virDomainDiskDefParseXML(virCapsPtr caps, char *cachetag = NULL; char *error_policy = NULL; char *iotag = NULL; + char *ioeventfd = NULL; char *devaddr = NULL; virStorageEncryptionPtr encryption = NULL; char *serial = NULL; @@ -2128,6 +2134,7 @@ virDomainDiskDefParseXML(virCapsPtr caps, cachetag = virXMLPropString(cur, "cache"); error_policy = virXMLPropString(cur, "error_policy"); iotag = virXMLPropString(cur, "io"); + ioeventfd = virXMLPropString(cur, "ioeventfd"); } else if (xmlStrEqual(cur->name, BAD_CAST "readonly")) { def->readonly = 1; } else if (xmlStrEqual(cur->name, BAD_CAST "shareable")) { @@ -2264,6 +2271,24 @@ virDomainDiskDefParseXML(virCapsPtr caps, } } + if (ioeventfd) { + if (def->bus != VIR_DOMAIN_DISK_BUS_VIRTIO) { + virDomainReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("disk ioeventfd mode supported " + "only for virtio bus")); + goto error; + } + + int i; + if ((i = virDomainIoEventFdTypeFromString(ioeventfd)) <= 0) { + virDomainReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("unknown disk ioeventfd mode '%s'"), + ioeventfd); + goto error; + } + def->ioeventfd=i; + } + if (devaddr) { if (virDomainParseLegacyDeviceAddress(devaddr, &def->info.addr.pci) < 0) { @@ -2326,6 +2351,7 @@ cleanup: VIR_FREE(cachetag); VIR_FREE(error_policy); VIR_FREE(iotag); + VIR_FREE(ioeventfd); VIR_FREE(devaddr); VIR_FREE(serial); virStorageEncryptionFree(encryption); @@ -2713,6 +2739,7 @@ virDomainNetDefParseXML(virCapsPtr caps, char *model = NULL; char *backend = NULL; char *txmode = NULL; + char *ioeventfd = NULL; char *filter = NULL; char *internal = NULL; char *devaddr = NULL; @@ -2802,6 +2829,7 @@ virDomainNetDefParseXML(virCapsPtr caps, } else if (xmlStrEqual (cur->name, BAD_CAST "driver")) { backend = virXMLPropString(cur, "name"); txmode = virXMLPropString(cur, "txmode"); + ioeventfd = virXMLPropString(cur, "ioeventfd"); } else if (xmlStrEqual (cur->name, BAD_CAST "filterref")) { filter = virXMLPropString(cur, "filter"); VIR_FREE(filterparams); @@ -3018,6 +3046,16 @@ virDomainNetDefParseXML(virCapsPtr caps, } def->driver.virtio.txmode = m; } + if (ioeventfd) { + int i; + if ((i = virDomainIoEventFdTypeFromString(ioeventfd)) <= 0) { + virDomainReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("unknown interface ioeventfd mode '%s'"), + ioeventfd); + goto error; + } + def->driver.virtio.ioeventfd = i; + } } if (filter != NULL) { @@ -3057,6 +3095,7 @@ cleanup: VIR_FREE(model); VIR_FREE(backend); VIR_FREE(txmode); + VIR_FREE(ioeventfd); VIR_FREE(filter); VIR_FREE(type); VIR_FREE(internal); @@ -8291,6 +8330,7 @@ virDomainDiskDefFormat(virBufferPtr buf, const char *cachemode = virDomainDiskCacheTypeToString(def->cachemode); const char *error_policy = virDomainDiskErrorPolicyTypeToString(def->error_policy); const char *iomode = virDomainDiskIoTypeToString(def->iomode); + const char *ioeventfd = virDomainIoEventFdTypeToString(def->ioeventfd); if (!type) { virDomainReportError(VIR_ERR_INTERNAL_ERROR, @@ -8322,7 +8362,8 @@ virDomainDiskDefFormat(virBufferPtr buf, " <disk type='%s' device='%s'>\n", type, device); - if (def->driverName || def->driverType || def->cachemode) { + if (def->driverName || def->driverType || def->cachemode || + def->ioeventfd) { virBufferAsprintf(buf, " <driver"); if (def->driverName) virBufferAsprintf(buf, " name='%s'", def->driverName); @@ -8334,6 +8375,8 @@ virDomainDiskDefFormat(virBufferPtr buf, virBufferAsprintf(buf, " error_policy='%s'", error_policy); if (def->iomode) virBufferAsprintf(buf, " io='%s'", iomode); + if (def->ioeventfd) + virBufferAsprintf(buf, " ioeventfd='%s'", ioeventfd); virBufferAsprintf(buf, "/>\n"); } @@ -8624,6 +8667,10 @@ virDomainNetDefFormat(virBufferPtr buf, virBufferAsprintf(buf, " txmode='%s'", virDomainNetVirtioTxModeTypeToString(def->driver.virtio.txmode)); } + if (def->driver.virtio.ioeventfd) { + virBufferAsprintf(buf, " ioeventfd='%s'", + virDomainIoEventFdTypeToString(def->driver.virtio.ioeventfd)); + } virBufferAddLit(buf, "/>\n"); } } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index ff5c28d..994ff91 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -206,6 +206,14 @@ enum virDomainDiskIo { VIR_DOMAIN_DISK_IO_LAST }; +enum virDomainIoEventFd { + VIR_DOMAIN_IO_EVENT_FD_DEFAULT = 0, + VIR_DOMAIN_IO_EVENT_FD_ON, + VIR_DOMAIN_IO_EVENT_FD_OFF, + + VIR_DOMAIN_IO_EVENT_FD_LAST +}; + /* Stores the virtual disk configuration */ typedef struct _virDomainDiskDef virDomainDiskDef; typedef virDomainDiskDef *virDomainDiskDefPtr; @@ -225,6 +233,7 @@ struct _virDomainDiskDef { int error_policy; int bootIndex; int iomode; + int ioeventfd; unsigned int readonly : 1; unsigned int shared : 1; virDomainDeviceInfo info; @@ -361,6 +370,7 @@ struct _virDomainNetDef { struct { enum virDomainNetBackendType name; /* which driver backend to use */ enum virDomainNetVirtioTxModeType txmode; + enum virDomainIoEventFd ioeventfd; } virtio; } driver; union { @@ -1554,6 +1564,7 @@ VIR_ENUM_DECL(virDomainDiskCache) VIR_ENUM_DECL(virDomainDiskErrorPolicy) VIR_ENUM_DECL(virDomainDiskProtocol) VIR_ENUM_DECL(virDomainDiskIo) +VIR_ENUM_DECL(virDomainIoEventFd) VIR_ENUM_DECL(virDomainController) VIR_ENUM_DECL(virDomainControllerModel) VIR_ENUM_DECL(virDomainFS) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 03d2ddb..5c8d272 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -292,6 +292,8 @@ virDomainHostdevDefFree; virDomainHostdevModeTypeToString; virDomainHostdevSubsysTypeToString; virDomainInputDefFree; +virDomainIoEventFdTypeFromString; +virDomainIoEventFdTypeToString; virDomainLeaseIndex; virDomainLeaseInsert; virDomainLeaseInsertPreAlloc; diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 28c89b5..ad62a07 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -121,6 +121,7 @@ VIR_ENUM_IMPL(qemuCaps, QEMU_CAPS_LAST, "device-qxl-vga", "pci-multifunction", /* 60 */ + "virtio-blk-pci.ioeventfd", ); struct qemu_feature_flags { @@ -1207,6 +1208,8 @@ qemuCapsParseDeviceStr(const char *str, virBitmapPtr flags) qemuCapsSet(flags, QEMU_CAPS_VIRTIO_TX_ALG); if (strstr(str, "name \"qxl-vga\"")) qemuCapsSet(flags, QEMU_CAPS_DEVICE_QXL_VGA); + if (strstr(str, "virtio-blk-pci.ioeventfd")) + qemuCapsSet(flags, QEMU_CAPS_VIRTIO_IOEVENTFD); return 0; } diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index e6d2fa3..0b9c8be 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -96,6 +96,7 @@ enum qemuCapsFlags { QEMU_CAPS_VIRTIO_TX_ALG = 58, /* -device virtio-net-pci,tx=string */ QEMU_CAPS_DEVICE_QXL_VGA = 59, /* Is the primary and vga campatible qxl device named qxl-vga? */ QEMU_CAPS_PCI_MULTIFUNCTION = 60, /* -device multifunction=on|off */ + QEMU_CAPS_VIRTIO_IOEVENTFD = 61, /* IOeventFD feature: virtio-{net|blk}-pci.ioeventfd=on/off */ QEMU_CAPS_LAST, /* this must always be the last item */ }; diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 6346243..f7c06f8 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -1287,6 +1287,16 @@ qemuBuildDeviceAddressStr(virBufferPtr buf, return 0; } +static int +qemuBuildIoEventFdStr(virBufferPtr buf, + enum virDomainIoEventFd use, + virBitmapPtr qemuCaps) +{ + if (use && qemuCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_IOEVENTFD)) + virBufferAsprintf(buf, ",ioeventfd=%s", + virDomainIoEventFdTypeToString(use)); + return 0; +} #define QEMU_SERIAL_PARAM_ACCEPTED_CHARS \ "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789-_" @@ -1554,6 +1564,7 @@ qemuBuildDriveDevStr(virDomainDiskDefPtr disk, break; case VIR_DOMAIN_DISK_BUS_VIRTIO: virBufferAddLit(&opt, "virtio-blk-pci"); + qemuBuildIoEventFdStr(&opt, disk->ioeventfd, qemuCaps); qemuBuildDeviceAddressStr(&opt, &disk->info, qemuCaps); break; case VIR_DOMAIN_DISK_BUS_USB: @@ -1777,6 +1788,8 @@ qemuBuildNicDevStr(virDomainNetDefPtr net, goto error; } } + if (usingVirtio) + qemuBuildIoEventFdStr(&buf, net->driver.virtio.ioeventfd, qemuCaps); if (vlan == -1) virBufferAsprintf(&buf, ",netdev=host%s", net->info.alias); else diff --git a/tests/qemuhelptest.c b/tests/qemuhelptest.c index 327a0c7..119e771 100644 --- a/tests/qemuhelptest.c +++ b/tests/qemuhelptest.c @@ -475,7 +475,8 @@ mymain(void) QEMU_CAPS_CCID_PASSTHRU, QEMU_CAPS_CHARDEV_SPICEVMC, QEMU_CAPS_DEVICE_QXL_VGA, - QEMU_CAPS_VIRTIO_TX_ALG); + QEMU_CAPS_VIRTIO_TX_ALG, + QEMU_CAPS_VIRTIO_IOEVENTFD); return ret == 0 ? EXIT_SUCCESS : EXIT_FAILURE; } diff --git a/tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.args b/tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.args new file mode 100644 index 0000000..c512f15 --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.args @@ -0,0 +1,11 @@ +LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test QEMU_AUDIO_DRV=none \ +/usr/bin/qemu -S -M pc-0.13 -m 1024 -smp 1 -nodefaults \ +-monitor unix:/tmp/test-monitor,server,nowait -no-acpi \ +-boot dc -device virtio-serial-pci,id=virtio-serial0,bus=pci.0,addr=0x6 \ +-drive file=/var/lib/libvirt/images/f14.img,if=none,id=drive-virtio-disk0 \ +-device virtio-blk-pci,ioeventfd=on,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0 \ +-drive file=/var/lib/libvirt/Fedora-14-x86_64-Live-KDE.iso,if=none,media=cdrom,id=drive-ide0-1-0 \ +-device ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0 \ +-device virtio-net-pci,tx=bh,ioeventfd=off,vlan=0,id=net0,mac=52:54:00:e5:48:58,bus=pci.0,addr=0x3 \ +-net user,vlan=0,name=hostnet0 -serial pty -usb -vnc 127.0.0.1:-809 -std-vga \ +-device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x5 diff --git a/tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.xml b/tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.xml new file mode 100644 index 0000000..c565c9f --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-disk-ioeventfd.xml @@ -0,0 +1,50 @@ +<domain type='qemu'> + <name>test</name> + <memory>1048576</memory> + <vcpu>1</vcpu> + <os> + <type arch='x86_64' machine='pc-0.13'>hvm</type> + <boot dev='cdrom'/> + <boot dev='hd'/> + <bootmenu enable='yes'/> + </os> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>restart</on_crash> + <devices> + <emulator>/usr/bin/qemu</emulator> + <disk type='file' device='disk'> + <driver name='qemu' type='qcow2' ioeventfd='on'/> + <source file='/var/lib/libvirt/images/f14.img'/> + <target dev='vda' bus='virtio'/> + <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> + </disk> + <disk type='file' device='cdrom'> + <driver name='qemu' type='raw'/> + <source file='/var/lib/libvirt/Fedora-14-x86_64-Live-KDE.iso'/> + <target dev='hdc' bus='ide'/> + <readonly/> + <address type='drive' controller='0' bus='1' unit='0'/> + </disk> + <interface type='user'> + <mac address='52:54:00:e5:48:58'/> + <model type='virtio'/> + <driver name='vhost' txmode='iothread' ioeventfd='off'/> + </interface> + <controller type='virtio-serial' index='0'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/> + </controller> + <serial type='pty'> + <target port='0'/> + </serial> + <console type='pty'> + <target type='serial' port='0'/> + </console> + <graphics type='vnc' port='5091' autoport='no' listen='127.0.0.1'/> + <video> + <model type='vga' vram='9216' heads='1'/> + <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> + </video> + </devices> +</domain> diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index bd07efa..782664a 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -350,6 +350,10 @@ mymain(void) DO_TEST("disk-aio", false, QEMU_CAPS_DRIVE, QEMU_CAPS_DRIVE_AIO, QEMU_CAPS_DRIVE_CACHE_V2, QEMU_CAPS_DRIVE_FORMAT); + DO_TEST("disk-ioeventfd", false, + QEMU_CAPS_DRIVE, QEMU_CAPS_VIRTIO_IOEVENTFD, + QEMU_CAPS_VIRTIO_TX_ALG, QEMU_CAPS_DEVICE); + DO_TEST("graphics-vnc", false, NONE); DO_TEST("graphics-vnc-socket", false, NONE); -- 1.7.5.rc3

2 2

[libvirt] [PATCH v4]: set and restore MAC address of a NIC when using PASSTHROUGH mode
by Gerhard Stenzel 22 Jun '11

22 Jun '11

Next try .. The following patch addresses the problem that when a PASSTHROUGH mode DIRECT NIC connection is made the MAC address of the NIC is not automatically set and reset to the configured VM MAC and back again. The attached patch fixes this problem by setting and resetting the MAC while remembering the previous setting while the VM is running. This also works if libvirtd is restarted while the VM is running. the patch passes make syntax-check Signed-off-by: Dirk Herrendoerfer <d.herrendoerfer at herrendoerfer.name> Signed-off-by: Gerhard Stenzel <gerhard.stenzel(a)de.ibm.com> --- Index: libvirt/src/qemu/qemu_command.c =================================================================== --- libvirt.orig/src/qemu/qemu_command.c +++ libvirt/src/qemu/qemu_command.c @@ -128,7 +128,7 @@ qemuPhysIfaceConnect(virDomainDefPtr def rc = openMacvtapTap(net->ifname, net->mac, net->data.direct.linkdev, net->data.direct.mode, vnet_hdr, def->uuid, &net->data.direct.virtPortProfile, &res_ifname, - vmop); + vmop, driver->stateDir); if (rc >= 0) { qemuAuditNetDevice(def, net, res_ifname, true); VIR_FREE(net->ifname); @@ -149,7 +149,9 @@ qemuPhysIfaceConnect(virDomainDefPtr def if (err) { VIR_FORCE_CLOSE(rc); delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, - &net->data.direct.virtPortProfile); + net->data.direct.mode, + &net->data.direct.virtPortProfile, + driver->stateDir); VIR_FREE(net->ifname); } } Index: libvirt/src/qemu/qemu_process.c =================================================================== --- libvirt.orig/src/qemu/qemu_process.c +++ libvirt/src/qemu/qemu_process.c @@ -2876,7 +2876,8 @@ void qemuProcessStop(struct qemud_driver virDomainNetDefPtr net = def->nets[i]; if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, - &net->data.direct.virtPortProfile); + net->data.direct.mode, + &net->data.direct.virtPortProfile, driver->stateDir); VIR_FREE(net->ifname); } } Index: libvirt/src/util/macvtap.c =================================================================== --- libvirt.orig/src/util/macvtap.c +++ libvirt/src/util/macvtap.c @@ -545,6 +545,104 @@ configMacvtapTap(int tapfd, int vnet_hdr return 0; } +/** + * replaceMacAdress: + * @macaddress: new MAC address for interface + * @linkdev: name of interface + * @stateDir: directory to store old MAC address + * + * Returns 0 on success, -1 in case of fatal error, error code otherwise. + * + */ +static int +replaceMacAdress(const unsigned char *macaddress, + const char *linkdev, + char *stateDir) +{ + unsigned char oldmac[6]; + int rc; + + rc = ifaceGetMacaddr(linkdev, oldmac); + + if (rc) { + virReportSystemError(rc, + _("Getting MAC address from '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]); + } else { + char *path = NULL; + char macstr[VIR_MAC_STRING_BUFLEN]; + + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + return errno; + } + virFormatMacAddr(oldmac, macstr); + if (virFileWriteStr(path, macstr, O_CREAT|O_TRUNC|O_WRONLY) < 0) { + virReportSystemError(errno, _("Unable to preserve mac for %s"), + linkdev); + return errno; + } + } + + rc = ifaceSetMacaddr(linkdev, macaddress); + if (rc) { + virReportSystemError(errno, + _("Setting MAC address on '%s' to " + "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + macaddress[0], macaddress[1], macaddress[2], + macaddress[3], macaddress[4], macaddress[5]); + } + return rc; +} + +/** + * restoreMacAddress: + * @linkdev: name of interface + * @stateDir: directory containing old MAC address + * + * Returns 0 on success, -1 in case of fatal error, error code otherwise. + * + */ +static int +restoreMacAddress(const char *linkdev, + char *stateDir) +{ + int ret; + char *oldmacname = NULL; + char *macstr = NULL; + char *path = NULL; + unsigned char oldmac[6]; + + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + return -1; + } + + if (virFileReadAll(path, VIR_MAC_STRING_BUFLEN, &macstr) < 0) { + return errno; + } + + if (virParseMacAddr(macstr, &oldmac[0]) != 0) { + macvtapError(VIR_ERR_INTERNAL_ERROR, + _("Cannot parse MAC address from '%s'"), + oldmacname); + return -1; + } + + /*reset mac and remove file-ignore results*/ + ignore_value(ifaceSetMacaddr(linkdev, oldmac)); + ignore_value(unlink(path)); + VIR_FREE(macstr); + return 0; +} /** * openMacvtapTap: @@ -575,7 +673,8 @@ openMacvtapTap(const char *tgifname, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, char **res_ifname, - enum virVMOperationType vmOp) + enum virVMOperationType vmOp, + char *stateDir) { const char *type = "macvtap"; int c, rc; @@ -589,6 +688,19 @@ openMacvtapTap(const char *tgifname, VIR_DEBUG("%s: VM OPERATION: %s", __FUNCTION__, virVMOperationTypeToString(vmOp)); + /** Note: When using PASSTHROUGH mode with MACVTAP devices the link + * device's MAC address must be set to the VMs MAC address. In + * order to not confuse the first switch or bridge in line this MAC + * address must be reset when the VM is shut down. + * This is especially important when using SRIOV capable cards that + * emulate their switch in firmware. + */ + if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + if (replaceMacAdress(macaddress, linkdev, stateDir) != 0) { + return -1; + } + } + if (tgifname) { if(ifaceGetIndex(false, tgifname, &ifindex) == 0) { if (STRPREFIX(tgifname, @@ -684,8 +796,14 @@ void delMacvtap(const char *ifname, const unsigned char *macaddr, const char *linkdev, - virVirtualPortProfileParamsPtr virtPortProfile) + int mode, + virVirtualPortProfileParamsPtr virtPortProfile, + char *stateDir) { + if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + restoreMacAddress(linkdev, stateDir); + } + if (ifname) { vpDisassociatePortProfileId(ifname, macaddr, linkdev, Index: libvirt/src/util/macvtap.h =================================================================== --- libvirt.orig/src/util/macvtap.h +++ libvirt/src/util/macvtap.h @@ -83,12 +83,15 @@ int openMacvtapTap(const char *ifname, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, char **res_ifname, - enum virVMOperationType vmop); + enum virVMOperationType vmop, + char *stateDir); void delMacvtap(const char *ifname, const unsigned char *macaddress, const char *linkdev, - virVirtualPortProfileParamsPtr virtPortProfile); + int mode, + virVirtualPortProfileParamsPtr virtPortProfile, + char *stateDir); int vpAssociatePortProfileId(const char *macvtap_ifname, const unsigned char *macvtap_macaddr, Index: libvirt/src/qemu/qemu_hotplug.c =================================================================== --- libvirt.orig/src/qemu/qemu_hotplug.c +++ libvirt/src/qemu/qemu_hotplug.c @@ -1610,7 +1610,9 @@ int qemuDomainDetachNetDevice(struct qem #if WITH_MACVTAP if (detach->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(detach->ifname, detach->mac, detach->data.direct.linkdev, - &detach->data.direct.virtPortProfile); + detach->data.direct.mode, + &detach->data.direct.virtPortProfile, + driver->stateDir); VIR_FREE(detach->ifname); } #endif Index: libvirt/src/util/interface.c =================================================================== --- libvirt.orig/src/util/interface.c +++ libvirt/src/util/interface.c @@ -390,3 +390,99 @@ ifaceGetVlanID(const char *vlanifname AT return ENOSYS; } #endif /* __linux__ */ + +/** + * ifaceGetMacaddr: + * @ifname: interface name to set MTU for + * @macaddr: MAC address (VIR_MAC_BUFLEN in size) + * + * This function gets the @macaddr for a given interface @ifname. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +#ifdef __linux__ +int +ifaceGetMacaddr(const char *ifname ATTRIBUTE_UNUSED, + unsigned char *macaddr ATTRIBUTE_UNUSED) +{ + struct ifreq ifr; + int fd; + + if (!ifname) + return EINVAL; + + fd = socket(AF_INET, SOCK_STREAM, 0); + if (fd < 0) + return errno; + + memset(&ifr, 0, sizeof(struct ifreq)); + if (virStrcpyStatic(ifr.ifr_name, ifname) == NULL) + return EINVAL; + + if (ioctl(fd, SIOCGIFHWADDR, (char *)&ifr) != 0) + return errno; + + memcpy(macaddr, ifr.ifr_ifru.ifru_hwaddr.sa_data, VIR_MAC_BUFLEN); + + return 0; +} + +#else + +int +ifaceGetMacaddr(const char *ifname ATTRIBUTE_UNUSED, + unsigned char *macaddr ATTRIBUTE_UNUSED) +{ + return ENOSYS; +} + +#endif /* __linux__ */ + +/** + * ifaceSetMacaddr: + * @ifname: interface name to set MTU for + * @macaddr: MAC address (VIR_MAC_BUFLEN in size) + * + * This function sets the @macaddr for a given interface @ifname. This + * gets rid of the kernel's automatically assigned random MAC. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +#ifdef __linux__ +int +ifaceSetMacaddr(const char *ifname ATTRIBUTE_UNUSED, + const unsigned char *macaddr ATTRIBUTE_UNUSED) +{ + struct ifreq ifr; + int fd; + + if (!ifname) + return EINVAL; + + fd = socket(AF_INET, SOCK_STREAM, 0); + if (fd < 0) + return errno; + + memset(&ifr, 0, sizeof(struct ifreq)); + if (virStrcpyStatic(ifr.ifr_name, ifname) == NULL) + return EINVAL; + + /* To fill ifr.ifr_hdaddr.sa_family field */ + if (ioctl(fd, SIOCGIFHWADDR, &ifr) != 0) + return errno; + + memcpy(ifr.ifr_hwaddr.sa_data, macaddr, VIR_MAC_BUFLEN); + + return ioctl(fd, SIOCSIFHWADDR, &ifr) == 0 ? 0 : errno; +} + +#else + +int +ifaceSetMacaddr(const char *ifname ATTRIBUTE_UNUSED, + const unsigned char *macaddr ATTRIBUTE_UNUSED) +{ + return ENOSYS; +} + +#endif /* __linux__ */ Index: libvirt/src/util/interface.h =================================================================== --- libvirt.orig/src/util/interface.h +++ libvirt/src/util/interface.h @@ -32,4 +32,8 @@ int ifaceGetIndex(bool reportError, cons int ifaceGetVlanID(const char *vlanifname, int *vlanid); +int ifaceSetMacaddr(const char *ifname, const unsigned char *macaddr); + +int ifaceGetMacaddr(const char *ifname, unsigned char *macaddr); + #endif /* __VIR_INTERFACE_H__ */ Index: libvirt/src/libvirt_private.syms =================================================================== --- libvirt.orig/src/libvirt_private.syms +++ libvirt/src/libvirt_private.syms @@ -507,7 +507,9 @@ ifaceCheck; ifaceCtrl; ifaceGetFlags; ifaceGetIndex; +ifaceGetMacaddr; ifaceGetVlanID; +ifaceSetMacaddr; ifaceIsUp; =================================================================== Best regards, Gerhard Stenzel ------------------------------------------------------------------------------------- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Martin Jetter Geschaeftsfuehrung: Dirk Wittkopp Sitz der Gesellschaft: Boeblingen Registergericht: Amtsgericht Stuttgart, HRB 243294

4 7

[libvirt] [PATCH 2/2] cleanup: make several interface functions commonly available
by Stefan Berger 21 Jun '11

21 Jun '11

In a second cleanup step this patch makes several interface functions from macvtap.c commonly available by moving them into interface.c and prefixing their names with 'iface'. Those functions taking Linux-specific structures as parameters are only visible on Linux. ifaceRestoreMacAddress returns the return code from the ifaceSetMacAddr call and display an error message if setting the MAC address did not work. The caller is unchanged and still ignores the return code (which is ok). Signed-off-by: Stefan Berger <stefanb(a)linux.vnet.ibm.com> --- src/libvirt_private.syms | 8 src/util/interface.c | 590 +++++++++++++++++++++++++++++++++++++++++++++++ src/util/interface.h | 33 ++ src/util/macvtap.c | 495 --------------------------------------- 4 files changed, 640 insertions(+), 486 deletions(-) Index: libvirt-acl/src/util/interface.c =================================================================== --- libvirt-acl.orig/src/util/interface.c +++ libvirt-acl/src/util/interface.c @@ -27,6 +27,7 @@ #include <sys/socket.h> #include <sys/ioctl.h> +#include <fcntl.h> #ifdef __linux__ # include <linux/if.h> @@ -40,6 +41,10 @@ #include "interface.h" #include "virterror_internal.h" #include "files.h" +#include "memory.h" +#include "netlink.h" + +#define VIR_FROM_THIS VIR_FROM_NET #define ifaceError(code, ...) \ virReportErrorHelper(VIR_FROM_NET, code, __FILE__, \ @@ -486,3 +491,588 @@ ifaceSetMacaddr(const char *ifname ATTRI } #endif /* __linux__ */ + + +/** + * ifaceLinkAdd + * + * @type: The type of device, i.e., "macvtap" + * @macaddress: The MAC address of the device + * @macaddrsize: The size of the MAC address, typically '6' + * @ifname: The name the interface is supposed to have; optional parameter + * @srcdev: The name of the 'link' device + * @macvlan_mode: The macvlan mode to use + * @retry: Pointer to integer that will be '1' upon return if an interface + * with the same name already exists and it is worth to try + * again with a different name + * + * Create a macvtap device with the given properties. + * + * Returns 0 on success, -1 on fatal error. + */ +#if __linux__ +int +ifaceMacvtapLinkAdd(const char *type, + const unsigned char *macaddress, int macaddrsize, + const char *ifname, + const char *srcdev, + uint32_t macvlan_mode, + int *retry) +{ + int rc = 0; + struct nlmsghdr *resp; + struct nlmsgerr *err; + struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC }; + int ifindex; + unsigned char *recvbuf = NULL; + unsigned int recvbuflen; + struct nl_msg *nl_msg; + struct nlattr *linkinfo, *info_data; + + if (ifaceGetIndex(true, srcdev, &ifindex) != 0) + return -1; + + *retry = 0; + + nl_msg = nlmsg_alloc_simple(RTM_NEWLINK, + NLM_F_REQUEST | NLM_F_CREATE | NLM_F_EXCL); + if (!nl_msg) { + virReportOOMError(); + return -1; + } + + if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) + goto buffer_too_small; + + if (nla_put_u32(nl_msg, IFLA_LINK, ifindex) < 0) + goto buffer_too_small; + + if (nla_put(nl_msg, IFLA_ADDRESS, macaddrsize, macaddress) < 0) + goto buffer_too_small; + + if (ifname && + nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0) + goto buffer_too_small; + + if (!(linkinfo = nla_nest_start(nl_msg, IFLA_LINKINFO))) + goto buffer_too_small; + + if (nla_put(nl_msg, IFLA_INFO_KIND, strlen(type), type) < 0) + goto buffer_too_small; + + if (macvlan_mode > 0) { + if (!(info_data = nla_nest_start(nl_msg, IFLA_INFO_DATA))) + goto buffer_too_small; + + if (nla_put(nl_msg, IFLA_MACVLAN_MODE, sizeof(macvlan_mode), + &macvlan_mode) < 0) + goto buffer_too_small; + + nla_nest_end(nl_msg, info_data); + } + + nla_nest_end(nl_msg, linkinfo); + + if (nlComm(nl_msg, &recvbuf, &recvbuflen, 0) < 0) { + rc = -1; + goto err_exit; + } + + if (recvbuflen < NLMSG_LENGTH(0) || recvbuf == NULL) + goto malformed_resp; + + resp = (struct nlmsghdr *)recvbuf; + + switch (resp->nlmsg_type) { + case NLMSG_ERROR: + err = (struct nlmsgerr *)NLMSG_DATA(resp); + if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) + goto malformed_resp; + + switch (err->error) { + + case 0: + break; + + case -EEXIST: + *retry = 1; + rc = -1; + break; + + default: + virReportSystemError(-err->error, + _("error creating %s type of interface"), + type); + rc = -1; + } + break; + + case NLMSG_DONE: + break; + + default: + goto malformed_resp; + } + +err_exit: + nlmsg_free(nl_msg); + + VIR_FREE(recvbuf); + + return rc; + +malformed_resp: + nlmsg_free(nl_msg); + + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("malformed netlink response message")); + VIR_FREE(recvbuf); + return -1; + +buffer_too_small: + nlmsg_free(nl_msg); + + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("allocated netlink buffer is too small")); + return -1; +} + +#else + +int +ifaceMacvtapLinkAdd(const char *type ATTRIBUTE_UNUSED, + const unsigned char *macaddress ATTRIBUTE_UNUSED, + int macaddrsize ATTRIBUTE_UNUSED, + const char *ifname ATTRIBUTE_UNUSED, + const char *srcdev ATTRIBUTE_UNUSED, + uint32_t macvlan_mode ATTRIBUTE_UNUSED, + int *retry ATTRIBUTE_UNUSED) +{ + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("ifaceMacvtapLinkAdd is not supported on non-linux platforms")); + return -1; +} + +#endif + + +/** + * ifaceLinkDel + * + * @ifname: Name of the interface + * + * Tear down an interface with the given name. + * + * Returns 0 on success, -1 on fatal error. + */ +#if __linux__ +int +ifaceLinkDel(const char *ifname) +{ + int rc = 0; + struct nlmsghdr *resp; + struct nlmsgerr *err; + struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC }; + unsigned char *recvbuf = NULL; + unsigned int recvbuflen; + struct nl_msg *nl_msg; + + nl_msg = nlmsg_alloc_simple(RTM_DELLINK, + NLM_F_REQUEST | NLM_F_CREATE | NLM_F_EXCL); + if (!nl_msg) { + virReportOOMError(); + return -1; + } + + if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) + goto buffer_too_small; + + if (nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0) + goto buffer_too_small; + + if (nlComm(nl_msg, &recvbuf, &recvbuflen, 0) < 0) { + rc = -1; + goto err_exit; + } + + if (recvbuflen < NLMSG_LENGTH(0) || recvbuf == NULL) + goto malformed_resp; + + resp = (struct nlmsghdr *)recvbuf; + + switch (resp->nlmsg_type) { + case NLMSG_ERROR: + err = (struct nlmsgerr *)NLMSG_DATA(resp); + if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) + goto malformed_resp; + + if (err->error) { + virReportSystemError(-err->error, + _("error destroying %s interface"), + ifname); + rc = -1; + } + break; + + case NLMSG_DONE: + break; + + default: + goto malformed_resp; + } + +err_exit: + nlmsg_free(nl_msg); + + VIR_FREE(recvbuf); + + return rc; + +malformed_resp: + nlmsg_free(nl_msg); + + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("malformed netlink response message")); + VIR_FREE(recvbuf); + return -1; + +buffer_too_small: + nlmsg_free(nl_msg); + + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("allocated netlink buffer is too small")); + return -1; +} + +#else + +int +ifaceLinkDel(const char *ifname ATTRIBUTE_UNUSED) +{ + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("ifaceLinkDel is not supported on non-linux platforms")); + return -1; +} + +#endif + + +#if __linux__ +static struct nla_policy ifla_policy[IFLA_MAX + 1] = +{ + [IFLA_VF_PORTS] = { .type = NLA_NESTED }, +}; + +/** + * ifaceMacvtapLinkDump + * + * @nltarget_kernel: whether to send the message to the kernel or another + * process + * @ifname: The name of the interface; only use if ifindex < 0 + * @ifindex: The interface index; may be < 0 if ifname is given + * @nlattr: pointer to a pointer of netlink attributes that will contain + * the results + * @recvbuf: Pointer to the buffer holding the returned netlink response + * message; free it, once not needed anymore + * @getPidFunc: Pointer to a function that will be invoked if the kernel + * is not the target of the netlink message but it is to be + * sent to another process. + * + * Get information about an interface given its name or index. + * + * Returns 0 on success, -1 on fatal error. + */ +int +ifaceMacvtapLinkDump(bool nltarget_kernel, const char *ifname, int ifindex, + struct nlattr **tb, unsigned char **recvbuf, + uint32_t (*getPidFunc)(void)) +{ + int rc = 0; + struct nlmsghdr *resp; + struct nlmsgerr *err; + struct ifinfomsg ifinfo = { + .ifi_family = AF_UNSPEC, + .ifi_index = ifindex + }; + unsigned int recvbuflen; + uint32_t pid = 0; + struct nl_msg *nl_msg; + + *recvbuf = NULL; + + nl_msg = nlmsg_alloc_simple(RTM_GETLINK, NLM_F_REQUEST); + if (!nl_msg) { + virReportOOMError(); + return -1; + } + + if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) + goto buffer_too_small; + + if (ifindex < 0 && ifname) { + if (nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0) + goto buffer_too_small; + } + + if (!nltarget_kernel) { + pid = getPidFunc(); + if (pid == 0) { + rc = -1; + goto err_exit; + } + } + + if (nlComm(nl_msg, recvbuf, &recvbuflen, pid) < 0) { + rc = -1; + goto err_exit; + } + + if (recvbuflen < NLMSG_LENGTH(0) || *recvbuf == NULL) + goto malformed_resp; + + resp = (struct nlmsghdr *)*recvbuf; + + switch (resp->nlmsg_type) { + case NLMSG_ERROR: + err = (struct nlmsgerr *)NLMSG_DATA(resp); + if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) + goto malformed_resp; + + if (err->error) { + virReportSystemError(-err->error, + _("error dumping %s (%d) interface"), + ifname, ifindex); + rc = -1; + } + break; + + case GENL_ID_CTRL: + case NLMSG_DONE: + if (nlmsg_parse(resp, sizeof(struct ifinfomsg), + tb, IFLA_MAX, ifla_policy)) { + goto malformed_resp; + } + break; + + default: + goto malformed_resp; + } + + if (rc != 0) + VIR_FREE(*recvbuf); + +err_exit: + nlmsg_free(nl_msg); + + return rc; + +malformed_resp: + nlmsg_free(nl_msg); + + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("malformed netlink response message")); + VIR_FREE(*recvbuf); + return -1; + +buffer_too_small: + nlmsg_free(nl_msg); + + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("allocated netlink buffer is too small")); + return -1; +} + +#endif + + +/** + * ifaceGetNthParent + * + * @ifindex : the index of the interface or -1 if ifname is given + * @ifname : the name of the interface; ignored if ifindex is valid + * @nthParent : the nth parent interface to get + * @parent_ifindex : pointer to int + * @parent_ifname : pointer to buffer of size IFNAMSIZ + * @nth : the nth parent that is actually returned; if for example eth0.100 + * was given and the 100th parent is to be returned, then eth0 will + * most likely be returned with nth set to 1 since the chain does + * not have more interfaces + * + * Get the nth parent interface of the given interface. 0 is the interface + * itself. + * + * Return 0 on success, != 0 otherwise + */ +#if __linux__ +int +ifaceGetNthParent(int ifindex, const char *ifname, unsigned int nthParent, + int *parent_ifindex, char *parent_ifname, + unsigned int *nth) +{ + int rc; + struct nlattr *tb[IFLA_MAX + 1] = { NULL, }; + unsigned char *recvbuf = NULL; + bool end = false; + unsigned int i = 0; + + *nth = 0; + + if (ifindex <= 0 && ifaceGetIndex(true, ifname, &ifindex) != 0) + return 1; + + while (!end && i <= nthParent) { + rc = ifaceMacvtapLinkDump(true, ifname, ifindex, tb, &recvbuf, NULL); + if (rc) + break; + + if (tb[IFLA_IFNAME]) { + if (!virStrcpy(parent_ifname, (char*)RTA_DATA(tb[IFLA_IFNAME]), + IFNAMSIZ)) { + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("buffer for root interface name is too small")); + VIR_FREE(recvbuf); + return 1; + } + *parent_ifindex = ifindex; + } + + if (tb[IFLA_LINK]) { + ifindex = *(int *)RTA_DATA(tb[IFLA_LINK]); + ifname = NULL; + } else + end = true; + + VIR_FREE(recvbuf); + + i++; + } + + if (nth) + *nth = i - 1; + + return rc; +} + +#else + +int +ifaceGetNthParent(int ifindex ATTRIBUTE_UNUSED, + const char *ifname ATTRIBUTE_UNUSED, + unsigned int nthParent ATTRIBUTE_UNUSED, + int *parent_ifindex ATTRIBUTE_UNUSED, + char *parent_ifname ATTRIBUTE_UNUSED, + unsigned int *nth ATTRIBUTE_UNUSED) +{ + ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", + _("ifaceGetNthParent is not supported on non-linux platforms")); + return -1; +} + +#endif + +/** + * ifaceReplaceMacAddress: + * @macaddress: new MAC address for interface + * @linkdev: name of interface + * @stateDir: directory to store old MAC address + * + * Returns 0 on success, -1 in case of fatal error, error code otherwise. + * + */ +int +ifaceReplaceMacAddress(const unsigned char *macaddress, + const char *linkdev, + const char *stateDir) +{ + unsigned char oldmac[6]; + int rc; + + rc = ifaceGetMacaddr(linkdev, oldmac); + + if (rc) { + virReportSystemError(rc, + _("Getting MAC address from '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]); + } else { + char *path = NULL; + char macstr[VIR_MAC_STRING_BUFLEN]; + + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + return errno; + } + virFormatMacAddr(oldmac, macstr); + if (virFileWriteStr(path, macstr, O_CREAT|O_TRUNC|O_WRONLY) < 0) { + virReportSystemError(errno, _("Unable to preserve mac for %s"), + linkdev); + return errno; + } + } + + rc = ifaceSetMacaddr(linkdev, macaddress); + if (rc) { + virReportSystemError(rc, + _("Setting MAC address on '%s' to " + "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + macaddress[0], macaddress[1], macaddress[2], + macaddress[3], macaddress[4], macaddress[5]); + } + return rc; +} + +/** + * ifaceRestoreMacAddress: + * @linkdev: name of interface + * @stateDir: directory containing old MAC address + * + * Returns 0 on success, -1 in case of fatal error, error code otherwise. + * + */ +int +ifaceRestoreMacAddress(const char *linkdev, + const char *stateDir) +{ + int rc; + char *oldmacname = NULL; + char *macstr = NULL; + char *path = NULL; + unsigned char oldmac[6]; + + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + return -1; + } + + if (virFileReadAll(path, VIR_MAC_STRING_BUFLEN, &macstr) < 0) { + return errno; + } + + if (virParseMacAddr(macstr, &oldmac[0]) != 0) { + ifaceError(VIR_ERR_INTERNAL_ERROR, + _("Cannot parse MAC address from '%s'"), + oldmacname); + return -1; + } + + /*reset mac and remove file-ignore results*/ + rc = ifaceSetMacaddr(linkdev, oldmac); + if (rc) { + virReportSystemError(rc, + _("Setting MAC address on '%s' to " + "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]); + } + ignore_value(unlink(path)); + VIR_FREE(macstr); + return rc; +} Index: libvirt-acl/src/util/interface.h =================================================================== --- libvirt-acl.orig/src/util/interface.h +++ libvirt-acl/src/util/interface.h @@ -10,6 +10,13 @@ #ifndef __VIR_INTERFACE_H__ # define __VIR_INTERFACE_H__ +# include <stdint.h> + +# if __linux__ +# include <sys/socket.h> +# include <linux/netlink.h> +# endif + # include "datatypes.h" int ifaceGetFlags(const char *name, short *flags); @@ -36,4 +43,30 @@ int ifaceSetMacaddr(const char *ifname, int ifaceGetMacaddr(const char *ifname, unsigned char *macaddr); +int ifaceMacvtapLinkAdd(const char *type, + const unsigned char *macaddress, int macaddrsize, + const char *ifname, + const char *srcdev, + uint32_t macvlan_mode, + int *retry); + +int ifaceLinkDel(const char *ifname); + +# if __linux__ +int ifaceMacvtapLinkDump(bool nltarget_kernel, const char *ifname, int ifindex, + struct nlattr **tb, unsigned char **recvbuf, + uint32_t (*getPidFunc)(void)); +# endif + +int ifaceGetNthParent(int ifindex, const char *ifname, unsigned int nthParent, + int *parent_ifindex, char *parent_ifname, + unsigned int *nth); + +int ifaceReplaceMacAddress(const unsigned char *macaddress, + const char *linkdev, + const char *stateDir); + +int ifaceRestoreMacAddress(const char *linkdev, + const char *stateDir); + #endif /* __VIR_INTERFACE_H__ */ Index: libvirt-acl/src/util/macvtap.c =================================================================== --- libvirt-acl.orig/src/util/macvtap.c +++ libvirt-acl/src/util/macvtap.c @@ -95,212 +95,6 @@ enum virVirtualPortOp { # if WITH_MACVTAP -static int -link_add(const char *type, - const unsigned char *macaddress, int macaddrsize, - const char *ifname, - const char *srcdev, - uint32_t macvlan_mode, - int *retry) -{ - int rc = 0; - struct nlmsghdr *resp; - struct nlmsgerr *err; - struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC }; - int ifindex; - unsigned char *recvbuf = NULL; - unsigned int recvbuflen; - struct nl_msg *nl_msg; - struct nlattr *linkinfo, *info_data; - - if (ifaceGetIndex(true, srcdev, &ifindex) != 0) - return -1; - - *retry = 0; - - nl_msg = nlmsg_alloc_simple(RTM_NEWLINK, - NLM_F_REQUEST | NLM_F_CREATE | NLM_F_EXCL); - if (!nl_msg) { - virReportOOMError(); - return -1; - } - - if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) - goto buffer_too_small; - - if (nla_put_u32(nl_msg, IFLA_LINK, ifindex) < 0) - goto buffer_too_small; - - if (nla_put(nl_msg, IFLA_ADDRESS, macaddrsize, macaddress) < 0) - goto buffer_too_small; - - if (ifname && - nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0) - goto buffer_too_small; - - if (!(linkinfo = nla_nest_start(nl_msg, IFLA_LINKINFO))) - goto buffer_too_small; - - if (nla_put(nl_msg, IFLA_INFO_KIND, strlen(type), type) < 0) - goto buffer_too_small; - - if (macvlan_mode > 0) { - if (!(info_data = nla_nest_start(nl_msg, IFLA_INFO_DATA))) - goto buffer_too_small; - - if (nla_put(nl_msg, IFLA_MACVLAN_MODE, sizeof(macvlan_mode), - &macvlan_mode) < 0) - goto buffer_too_small; - - nla_nest_end(nl_msg, info_data); - } - - nla_nest_end(nl_msg, linkinfo); - - if (nlComm(nl_msg, &recvbuf, &recvbuflen, 0) < 0) { - rc = -1; - goto err_exit; - } - - if (recvbuflen < NLMSG_LENGTH(0) || recvbuf == NULL) - goto malformed_resp; - - resp = (struct nlmsghdr *)recvbuf; - - switch (resp->nlmsg_type) { - case NLMSG_ERROR: - err = (struct nlmsgerr *)NLMSG_DATA(resp); - if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) - goto malformed_resp; - - switch (err->error) { - - case 0: - break; - - case -EEXIST: - *retry = 1; - rc = -1; - break; - - default: - virReportSystemError(-err->error, - _("error creating %s type of interface"), - type); - rc = -1; - } - break; - - case NLMSG_DONE: - break; - - default: - goto malformed_resp; - } - -err_exit: - nlmsg_free(nl_msg); - - VIR_FREE(recvbuf); - - return rc; - -malformed_resp: - nlmsg_free(nl_msg); - - macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", - _("malformed netlink response message")); - VIR_FREE(recvbuf); - return -1; - -buffer_too_small: - nlmsg_free(nl_msg); - - macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", - _("allocated netlink buffer is too small")); - return -1; -} - - -static int -link_del(const char *ifname) -{ - int rc = 0; - struct nlmsghdr *resp; - struct nlmsgerr *err; - struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC }; - unsigned char *recvbuf = NULL; - unsigned int recvbuflen; - struct nl_msg *nl_msg; - - nl_msg = nlmsg_alloc_simple(RTM_DELLINK, - NLM_F_REQUEST | NLM_F_CREATE | NLM_F_EXCL); - if (!nl_msg) { - virReportOOMError(); - return -1; - } - - if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) - goto buffer_too_small; - - if (nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0) - goto buffer_too_small; - - if (nlComm(nl_msg, &recvbuf, &recvbuflen, 0) < 0) { - rc = -1; - goto err_exit; - } - - if (recvbuflen < NLMSG_LENGTH(0) || recvbuf == NULL) - goto malformed_resp; - - resp = (struct nlmsghdr *)recvbuf; - - switch (resp->nlmsg_type) { - case NLMSG_ERROR: - err = (struct nlmsgerr *)NLMSG_DATA(resp); - if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) - goto malformed_resp; - - if (err->error) { - virReportSystemError(-err->error, - _("error destroying %s interface"), - ifname); - rc = -1; - } - break; - - case NLMSG_DONE: - break; - - default: - goto malformed_resp; - } - -err_exit: - nlmsg_free(nl_msg); - - VIR_FREE(recvbuf); - - return rc; - -malformed_resp: - nlmsg_free(nl_msg); - - macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", - _("malformed netlink response message")); - VIR_FREE(recvbuf); - return -1; - -buffer_too_small: - nlmsg_free(nl_msg); - - macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", - _("allocated netlink buffer is too small")); - return -1; -} - - /* Open the macvtap's tap device. * @ifname: Name of the macvtap interface * @retries : Number of retries in case udev for example may need to be @@ -450,104 +244,6 @@ configMacvtapTap(int tapfd, int vnet_hdr } /** - * replaceMacAdress: - * @macaddress: new MAC address for interface - * @linkdev: name of interface - * @stateDir: directory to store old MAC address - * - * Returns 0 on success, -1 in case of fatal error, error code otherwise. - * - */ -static int -replaceMacAdress(const unsigned char *macaddress, - const char *linkdev, - char *stateDir) -{ - unsigned char oldmac[6]; - int rc; - - rc = ifaceGetMacaddr(linkdev, oldmac); - - if (rc) { - virReportSystemError(rc, - _("Getting MAC address from '%s' " - "to '%02x:%02x:%02x:%02x:%02x:%02x' failed."), - linkdev, - oldmac[0], oldmac[1], oldmac[2], - oldmac[3], oldmac[4], oldmac[5]); - } else { - char *path = NULL; - char macstr[VIR_MAC_STRING_BUFLEN]; - - if (virAsprintf(&path, "%s/%s", - stateDir, - linkdev) < 0) { - virReportOOMError(); - return errno; - } - virFormatMacAddr(oldmac, macstr); - if (virFileWriteStr(path, macstr, O_CREAT|O_TRUNC|O_WRONLY) < 0) { - virReportSystemError(errno, _("Unable to preserve mac for %s"), - linkdev); - return errno; - } - } - - rc = ifaceSetMacaddr(linkdev, macaddress); - if (rc) { - virReportSystemError(errno, - _("Setting MAC address on '%s' to " - "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), - linkdev, - macaddress[0], macaddress[1], macaddress[2], - macaddress[3], macaddress[4], macaddress[5]); - } - return rc; -} - -/** - * restoreMacAddress: - * @linkdev: name of interface - * @stateDir: directory containing old MAC address - * - * Returns 0 on success, -1 in case of fatal error, error code otherwise. - * - */ -static int -restoreMacAddress(const char *linkdev, - char *stateDir) -{ - char *oldmacname = NULL; - char *macstr = NULL; - char *path = NULL; - unsigned char oldmac[6]; - - if (virAsprintf(&path, "%s/%s", - stateDir, - linkdev) < 0) { - virReportOOMError(); - return -1; - } - - if (virFileReadAll(path, VIR_MAC_STRING_BUFLEN, &macstr) < 0) { - return errno; - } - - if (virParseMacAddr(macstr, &oldmac[0]) != 0) { - macvtapError(VIR_ERR_INTERNAL_ERROR, - _("Cannot parse MAC address from '%s'"), - oldmacname); - return -1; - } - - /*reset mac and remove file-ignore results*/ - ignore_value(ifaceSetMacaddr(linkdev, oldmac)); - ignore_value(unlink(path)); - VIR_FREE(macstr); - return 0; -} - -/** * openMacvtapTap: * Create an instance of a macvtap device and open its tap character * device. @@ -599,7 +295,7 @@ openMacvtapTap(const char *tgifname, * emulate their switch in firmware. */ if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { - if (replaceMacAdress(macaddress, linkdev, stateDir) != 0) { + if (ifaceReplaceMacAddress(macaddress, linkdev, stateDir) != 0) { return -1; } } @@ -615,8 +311,8 @@ openMacvtapTap(const char *tgifname, return -1; } cr_ifname = tgifname; - rc = link_add(type, macaddress, 6, tgifname, linkdev, - macvtapMode, &do_retry); + rc = ifaceMacvtapLinkAdd(type, macaddress, 6, tgifname, linkdev, + macvtapMode, &do_retry); if (rc) return -1; } else { @@ -625,8 +321,8 @@ create_name: for (c = 0; c < 8192; c++) { snprintf(ifname, sizeof(ifname), MACVTAP_NAME_PATTERN, c); if (ifaceGetIndex(false, ifname, &ifindex) == ENODEV) { - rc = link_add(type, macaddress, 6, ifname, linkdev, - macvtapMode, &do_retry); + rc = ifaceMacvtapLinkAdd(type, macaddress, 6, ifname, linkdev, + macvtapMode, &do_retry); if (rc == 0) break; @@ -679,7 +375,7 @@ disassociate_exit: vmOp); link_del_exit: - link_del(cr_ifname); + ifaceLinkDel(cr_ifname); return rc; } @@ -704,7 +400,7 @@ delMacvtap(const char *ifname, char *stateDir) { if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { - restoreMacAddress(linkdev, stateDir); + ifaceRestoreMacAddress(linkdev, stateDir); } if (ifname) { @@ -712,7 +408,7 @@ delMacvtap(const char *ifname, linkdev, virtPortProfile, VIR_VM_OP_DESTROY); - link_del(ifname); + ifaceLinkDel(ifname); } } @@ -720,11 +416,6 @@ delMacvtap(const char *ifname, # ifdef IFLA_PORT_MAX -static struct nla_policy ifla_policy[IFLA_MAX + 1] = -{ - [IFLA_VF_PORTS] = { .type = NLA_NESTED }, -}; - static struct nla_policy ifla_port_policy[IFLA_PORT_MAX + 1] = { [IFLA_PORT_RESPONSE] = { .type = NLA_U16 }, @@ -764,173 +455,6 @@ getLldpadPid(void) { } -static int -link_dump(bool nltarget_kernel, const char *ifname, int ifindex, - struct nlattr **tb, unsigned char **recvbuf) -{ - int rc = 0; - struct nlmsghdr *resp; - struct nlmsgerr *err; - struct ifinfomsg ifinfo = { - .ifi_family = AF_UNSPEC, - .ifi_index = ifindex - }; - unsigned int recvbuflen; - uint32_t pid = 0; - struct nl_msg *nl_msg; - - *recvbuf = NULL; - - nl_msg = nlmsg_alloc_simple(RTM_GETLINK, NLM_F_REQUEST); - if (!nl_msg) { - virReportOOMError(); - return -1; - } - - if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) - goto buffer_too_small; - - if (ifindex < 0 && ifname) { - if (nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0) - goto buffer_too_small; - } - - if (!nltarget_kernel) { - pid = getLldpadPid(); - if (pid == 0) { - rc = -1; - goto err_exit; - } - } - - if (nlComm(nl_msg, recvbuf, &recvbuflen, pid) < 0) { - rc = -1; - goto err_exit; - } - - if (recvbuflen < NLMSG_LENGTH(0) || *recvbuf == NULL) - goto malformed_resp; - - resp = (struct nlmsghdr *)*recvbuf; - - switch (resp->nlmsg_type) { - case NLMSG_ERROR: - err = (struct nlmsgerr *)NLMSG_DATA(resp); - if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) - goto malformed_resp; - - if (err->error) { - virReportSystemError(-err->error, - _("error dumping %s (%d) interface"), - ifname, ifindex); - rc = -1; - } - break; - - case GENL_ID_CTRL: - case NLMSG_DONE: - if (nlmsg_parse(resp, sizeof(struct ifinfomsg), - tb, IFLA_MAX, ifla_policy)) { - goto malformed_resp; - } - break; - - default: - goto malformed_resp; - } - - if (rc != 0) - VIR_FREE(*recvbuf); - -err_exit: - nlmsg_free(nl_msg); - - return rc; - -malformed_resp: - nlmsg_free(nl_msg); - - macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", - _("malformed netlink response message")); - VIR_FREE(*recvbuf); - return -1; - -buffer_too_small: - nlmsg_free(nl_msg); - - macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", - _("allocated netlink buffer is too small")); - return -1; -} - - -/** - * ifaceGetNthParent - * - * @ifindex : the index of the interface or -1 if ifname is given - * @ifname : the name of the interface; ignored if ifindex is valid - * @nthParent : the nth parent interface to get - * @parent_ifindex : pointer to int - * @parent_ifname : pointer to buffer of size IFNAMSIZ - * @nth : the nth parent that is actually returned; if for example eth0.100 - * was given and the 100th parent is to be returned, then eth0 will - * most likely be returned with nth set to 1 since the chain does - * not have more interfaces - * - * Get the nth parent interface of the given interface. 0 is the interface - * itself. - * - * Return 0 on success, != 0 otherwise - */ -static int -ifaceGetNthParent(int ifindex, const char *ifname, unsigned int nthParent, - int *parent_ifindex, char *parent_ifname, - unsigned int *nth) -{ - int rc; - struct nlattr *tb[IFLA_MAX + 1] = { NULL, }; - unsigned char *recvbuf = NULL; - bool end = false; - unsigned int i = 0; - - *nth = 0; - - if (ifindex <= 0 && ifaceGetIndex(true, ifname, &ifindex) != 0) - return 1; - - while (!end && i <= nthParent) { - rc = link_dump(true, ifname, ifindex, tb, &recvbuf); - if (rc) - break; - - if (tb[IFLA_IFNAME]) { - if (!virStrcpy(parent_ifname, (char*)RTA_DATA(tb[IFLA_IFNAME]), - IFNAMSIZ)) { - macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", - _("buffer for root interface name is too small")); - VIR_FREE(recvbuf); - return 1; - } - *parent_ifindex = ifindex; - } - - if (tb[IFLA_LINK]) { - ifindex = *(int *)RTA_DATA(tb[IFLA_LINK]); - ifname = NULL; - } else - end = true; - - VIR_FREE(recvbuf); - - i++; - } - - if (nth) - *nth = i - 1; - - return rc; -} - /** * getPortProfileStatus * @@ -1258,7 +782,8 @@ doPortProfileOpCommon(bool nltarget_kern } while (--repeats >= 0) { - rc = link_dump(nltarget_kernel, NULL, ifindex, tb, &recvbuf); + rc = ifaceMacvtapLinkDump(nltarget_kernel, NULL, ifindex, tb, + &recvbuf, getLldpadPid); if (rc) goto err_exit; rc = getPortProfileStatus(tb, vf, instanceId, nltarget_kernel, Index: libvirt-acl/src/libvirt_private.syms =================================================================== --- libvirt-acl.orig/src/libvirt_private.syms +++ libvirt-acl/src/libvirt_private.syms @@ -499,9 +499,15 @@ ifaceCtrl; ifaceGetFlags; ifaceGetIndex; ifaceGetMacaddr; +ifaceGetNthParent; ifaceGetVlanID; -ifaceSetMacaddr; ifaceIsUp; +ifaceLinkDel; +ifaceMacvtapLinkAdd; +ifaceMacvtapLinkDump; +ifaceReplaceMacAddress; +ifaceRestoreMacaddress; +ifaceSetMacaddr; # interface_conf.h

2 1

[libvirt] [PATCH] build: require newer netcf when it is available
by Eric Blake 21 Jun '11

21 Jun '11

When building for newer Fedora or RHEL, take advantage of the newer netcf packaging to guarantee interface snapshot support. * libvirt.spec.in (BuildRequires): Bump minimum version on platforms that support netcf 0.1.8. --- I don't think we need an explicit 'Requires: netcf-libs >= 0.1.8' line for the runtime dependency, since netcf uses versioned symbols. I'm not sure whether F16 was the right target for requiring 0.1.8, or whether we should also try to include F15 (but 0.1.8 has not yet been built for F15). libvirt.spec.in | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 75b145a..11b8591 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -424,8 +424,12 @@ BuildRequires: libcap-ng-devel >= 0.5.0 BuildRequires: libssh2-devel %endif %if %{with_netcf} +%if 0%{?fedora} >= 16 || 0%{?rhel} >= 6 +BuildRequires: netcf-devel >= 0.1.8 +%else BuildRequires: netcf-devel >= 0.1.4 %endif +%endif %if %{with_esx} %if 0%{?fedora} >= 9 || 0%{?rhel} >= 6 BuildRequires: libcurl-devel -- 1.7.4.4

2 2

[libvirt] [PATCH] virsh: avoid bogus description
by Eric Blake 21 Jun '11

21 Jun '11

https://bugzilla.redhat.com/show_bug.cgi?id=682121 Gettext reserves the empty string for internal use, and it must not be passed through _(). We were violating this for commands that (for whatever reason) used "" for their description. * tools/virsh.c (vshCmddefHelp): Don't translate empty string. Reported by Tatsuo Kawasaki. --- tools/virsh.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index a315f05..bdd5005 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -12113,7 +12113,8 @@ vshCmddefHelp(vshControl *ctl, const char *cmdname) vshError(ctl, _("command '%s' doesn't exist"), cmdname); return false; } else { - const char *desc = _(vshCmddefGetInfo(def, "desc")); + /* Don't translate desc until after we know it isn't "". */ + const char *desc = vshCmddefGetInfo(def, "desc"); const char *help = _(vshCmddefGetInfo(def, "help")); char buf[256]; uint32_t opts_need_arg; @@ -12167,7 +12168,7 @@ vshCmddefHelp(vshControl *ctl, const char *cmdname) if (desc[0]) { /* Print the description only if it's not empty. */ fputs(_("\n DESCRIPTION\n"), stdout); - fprintf(stdout, " %s\n", desc); + fprintf(stdout, " %s\n", _(desc)); } if (def->opts) { -- 1.7.4.4

3 3

[libvirt] [libvirt-php] Announcement: Release 0.4.2 available
by Michal Novotny 21 Jun '11

21 Jun '11

Hello everybody. it's been a rather long time since the first release of libvirt-php bindings however there are many new features so I've decided to release a new version of libvirt-php, release 0.4.2. New features: - Storage pool / volume related functions (by Lyre) - VNC client support to get screenshot using gvnccapture (if present) - experimental - sending keys and client pointer events over VNC protocol - API to change domain VCPU count, memory amount and guest boot order - libvirt resources tracking - can be obtained by libvirt_print_binding_resources() API function - debug logging using libvirt_logfile_set() API function (default max log file size can be obtained in PHPInfo output) - functions to add/remove domain disks/NICs in an easy way (using a simple PHP API functions) For this release thanks goes to those guys: - Lyre <liyong(a)skybility.com> - storage-related functions as well as many bug fixes - Yukihiro Kawada <warp.kawada(a)gmail.com> - fix for the memory leak in libvirt_list_domains() - Daniel P. Berrange <berrange(a)redhat.com> - bug fixes/improvements for the build system - Radek Hladik <r.hladik(a)cybersales.cz> - patch to add API to update device - Tiziano Mueller <dev-zero(a)gentoo.org> - several bug fixes for threaded php and libvirt_connect/version failures without Xen I hope I didn't forget to mention anybody ;-) The new release can be found at: http://libvirt.org/sources/php/libvirt-php-0.4.2.tar.gz Michal -- Michal Novotny <minovotn(a)redhat.com>, RHCE, Red Hat Virtualization Team (xen userspace) | libvirt-php bindings

1 0

[libvirt] [PATCH 00/10] python: Implement missing stream/event ops
by Cole Robinson 21 Jun '11

21 Jun '11

Currently it isn't possible to implement the equivalent of 'virsh console' with the python bindings. This series implements the missing functionality required to do so. Cole Robinson (10): python: libvirt-override: use simpler debug python: generator: Don't print warning if nothing to warn about python: Implement bindings for virStreamEventAddCallback python: Implement virStreamSend/Recv python: Implement virStreamSend/RecvAll helpers Promote virEvent*Handle/Timeout to public API events: Correct virEventAddTimeout docs python: Add bindings for virEvent*Handle/Timeout python: events: Fix C->Python handle callback prototype python: Mark event callback wrappers as private daemon/libvirtd.c | 1 - daemon/mdns.c | 1 - examples/domain-events/events-python/event-test.py | 6 +- python/generator.py | 27 +- python/libvirt-override-virConnect.py | 102 ++--- python/libvirt-override-virStream.py | 123 +++++- python/libvirt-override.c | 492 ++++++++++++++----- python/libvirt-override.py | 80 +++- python/typewrappers.c | 14 + python/typewrappers.h | 2 +- src/conf/domain_event.c | 1 - src/conf/domain_event.h | 1 - src/fdstream.c | 1 - src/libvirt_private.syms | 9 - src/libvirt_public.syms | 6 + src/libxl/libxl_driver.c | 1 - src/lxc/lxc_driver.c | 1 - src/network/bridge_driver.c | 1 - src/node_device/node_device_hal.c | 1 - src/node_device/node_device_udev.c | 1 - src/openvz/openvz_driver.c | 1 - src/qemu/qemu_domain.c | 1 - src/qemu/qemu_driver.c | 1 - src/qemu/qemu_monitor.c | 1 - src/remote/remote_driver.c | 1 - src/test/test_driver.c | 1 - src/uml/uml_driver.c | 1 - src/util/event.c | 56 +++ src/util/event.h | 73 --- src/util/util.c | 1 - src/vbox/vbox_tmpl.c | 1 - src/xen/xen_inotify.c | 1 - src/xen/xs_internal.c | 1 - tools/console.c | 1 - 34 files changed, 692 insertions(+), 320 deletions(-) -- 1.7.4.4

3 31

Re: [libvirt] Network device abstraction aka virtual switch - V3
by Neil Wilson 21 Jun '11

21 Jun '11

> The current modes are: > > <forward layer='network' mode='route|nat'/> > > (in addition to not listing any mode, which equates to "isolated") > > Here are suggested new modes: Has anybody considered the migration requirements of networks in this new layout? If you move a machine attached to a 'route|nat' network to another Host you need to extend the network to the new Host and eventually you need to move the supporting processes that hand out the IP addresses/DNS addresses etc (if you're decommissioning a Host server for example). If you try to start a copy of the network on the new host, then it generally clashes with the pre-existing routing It would be useful to be able to extend a network onto another Host somehow and be able to migrate the supporting processes (DNSMasq, IP addressing, etc) between Hosts. (And obviously remove extensions as required as well). Perhaps this suggests that the address and route management systems need separating from the definition of the bridge network. (There's no real reason why the address management systems couldn't be attached to macvtap and vepa networks as well as basic bridges). Rgs Neil Wilson

2 1

[libvirt] [PATCH] Support reboots with the QEMU driver
by Daniel P. Berrange 21 Jun '11

21 Jun '11

For controlled shutdown we issue a 'system_powerdown' command to the QEMU monitor. This triggers an ACPI event which (most) guest OS wire up to a controlled shutdown. There is no equiv ACPI event to trigger a controlled reboot. This patch attempts to fake a reboot. - In qemuDomainObjPrivatePtr we have a bool fakeReboot flag. - The virDomainReboot method sets this flag and then triggers a normal 'system_powerdown'. - The QEMU process is started with '-no-shutdown' so that the guest CPUs pause when it powers off the guest - When we receive the 'POWEROFF' event from QEMU JSON monitor if fakeReboot is not set we invoke the qemuProcessKill command and shutdown continues normally - If fakeReboot was set, we spawn a background thread which issues 'system_reset' to perform a warm reboot of the guest hardware. Then it issues 'cont' to start the CPUs again * src/qemu/qemu_command.c: Add -no-shutdown flag if we have JSON support * src/qemu/qemu_domain.h: Add 'fakeReboot' flag to qemuDomainObjPrivate struct * src/qemu/qemu_driver.c: Fake reboot using the system_powerdown command if JSON support is available * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c, src/qemu/qemu_monitor_json.h, src/qemu/qemu_monitor_text.c, src/qemu/qemu_monitor_text.h: Add binding for system_reset command * src/qemu/qemu_process.c: Reset the guest & start CPUs if fakeReboot is set --- src/qemu/qemu_command.c | 7 +++ src/qemu/qemu_domain.h | 2 + src/qemu/qemu_driver.c | 65 ++++++++++++++++++++++++++++- src/qemu/qemu_monitor.c | 19 ++++++++ src/qemu/qemu_monitor.h | 1 + src/qemu/qemu_monitor_json.c | 19 ++++++++ src/qemu/qemu_monitor_json.h | 1 + src/qemu/qemu_monitor_text.c | 13 ++++++ src/qemu/qemu_monitor_text.h | 1 + src/qemu/qemu_process.c | 95 +++++++++++++++++++++++++++++++++++++++++- 10 files changed, 220 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 6346243..9ae1409 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3207,6 +3207,13 @@ qemuBuildCommandLine(virConnectPtr conn, def->onReboot != VIR_DOMAIN_LIFECYCLE_RESTART) virCommandAddArg(cmd, "-no-reboot"); + /* If JSON monitor is enabled, we can receive an event + * when QEMU stops. If we use no-shutdown, then we can + * watch for this event and do a soft/warm reboot. + */ + if (monitor_json) + virCommandAddArg(cmd, "-no-shutdown"); + if (!(def->features & (1 << VIR_DOMAIN_FEATURE_ACPI))) virCommandAddArg(cmd, "-no-acpi"); diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h index bacf5b5..b8469b6 100644 --- a/src/qemu/qemu_domain.h +++ b/src/qemu/qemu_domain.h @@ -89,6 +89,8 @@ struct _qemuDomainObjPrivate { virBitmapPtr qemuCaps; char *lockState; + + bool fakeReboot; }; struct qemuDomainWatchdogEvent diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 853c84c..4c827cf 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -1429,7 +1429,7 @@ cleanup: } -static int qemudDomainShutdown(virDomainPtr dom) { +static int qemuDomainShutdown(virDomainPtr dom) { struct qemud_driver *driver = dom->conn->privateData; virDomainObjPtr vm; int ret = -1; @@ -1461,6 +1461,8 @@ static int qemudDomainShutdown(virDomainPtr dom) { ret = qemuMonitorSystemPowerdown(priv->mon); qemuDomainObjExitMonitor(vm); + priv->fakeReboot = false; + endjob: if (qemuDomainObjEndJob(vm) == 0) vm = NULL; @@ -1472,11 +1474,66 @@ cleanup: } +static int qemuDomainReboot(virDomainPtr dom, unsigned int flags ATTRIBUTE_UNUSED) { + struct qemud_driver *driver = dom->conn->privateData; + virDomainObjPtr vm; + int ret = -1; + qemuDomainObjPrivatePtr priv; + + qemuDriverLock(driver); + vm = virDomainFindByUUID(&driver->domains, dom->uuid); + qemuDriverUnlock(driver); + + if (!vm) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(dom->uuid, uuidstr); + qemuReportError(VIR_ERR_NO_DOMAIN, + _("no domain with matching uuid '%s'"), uuidstr); + goto cleanup; + } + priv = vm->privateData; + +#if HAVE_YAJL + if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MONITOR_JSON)) { + if (qemuDomainObjBeginJob(vm) < 0) + goto cleanup; + + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto endjob; + } + + qemuDomainObjEnterMonitor(vm); + ret = qemuMonitorSystemPowerdown(priv->mon); + qemuDomainObjExitMonitor(vm); + + priv->fakeReboot = true; + + endjob: + if (qemuDomainObjEndJob(vm) == 0) + vm = NULL; + } else { +#endif + qemuReportError(VIR_ERR_NO_SUPPORT, "%s", + _("Reboot is not supported without the JSON monitor")); +#if HAVE_YAJL + } +#endif + +cleanup: + if (vm) + virDomainObjUnlock(vm); + return ret; +} + + static int qemudDomainDestroy(virDomainPtr dom) { struct qemud_driver *driver = dom->conn->privateData; virDomainObjPtr vm; int ret = -1; virDomainEventPtr event = NULL; + qemuDomainObjPrivatePtr priv; qemuDriverLock(driver); vm = virDomainFindByUUID(&driver->domains, dom->uuid); @@ -1488,6 +1545,9 @@ static int qemudDomainDestroy(virDomainPtr dom) { goto cleanup; } + priv = vm->privateData; + priv->fakeReboot = false; + /* Although qemuProcessStop does this already, there may * be an outstanding job active. We want to make sure we * can kill the process even if a job is active. Killing @@ -8146,7 +8206,8 @@ static virDriver qemuDriver = { .domainLookupByName = qemudDomainLookupByName, /* 0.2.0 */ .domainSuspend = qemudDomainSuspend, /* 0.2.0 */ .domainResume = qemudDomainResume, /* 0.2.0 */ - .domainShutdown = qemudDomainShutdown, /* 0.2.0 */ + .domainShutdown = qemuDomainShutdown, /* 0.2.0 */ + .domainReboot = qemuDomainReboot, /* 0.9.3 */ .domainDestroy = qemudDomainDestroy, /* 0.2.0 */ .domainGetOSType = qemudDomainGetOSType, /* 0.2.2 */ .domainGetMaxMemory = qemudDomainGetMaxMemory, /* 0.4.2 */ diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index 1428921..1dafc4b 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -1096,6 +1096,25 @@ int qemuMonitorSystemPowerdown(qemuMonitorPtr mon) } +int qemuMonitorSystemReset(qemuMonitorPtr mon) +{ + int ret; + VIR_DEBUG("mon=%p", mon); + + if (!mon) { + qemuReportError(VIR_ERR_INVALID_ARG, "%s", + _("monitor must not be NULL")); + return -1; + } + + if (mon->json) + ret = qemuMonitorJSONSystemReset(mon); + else + ret = qemuMonitorTextSystemReset(mon); + return ret; +} + + int qemuMonitorGetCPUInfo(qemuMonitorPtr mon, int **pids) { diff --git a/src/qemu/qemu_monitor.h b/src/qemu/qemu_monitor.h index 3bb0269..ae74a3c 100644 --- a/src/qemu/qemu_monitor.h +++ b/src/qemu/qemu_monitor.h @@ -194,6 +194,7 @@ int qemuMonitorStartCPUs(qemuMonitorPtr mon, int qemuMonitorStopCPUs(qemuMonitorPtr mon); int qemuMonitorGetStatus(qemuMonitorPtr mon, bool *running); +int qemuMonitorSystemReset(qemuMonitorPtr mon); int qemuMonitorSystemPowerdown(qemuMonitorPtr mon); int qemuMonitorGetCPUInfo(qemuMonitorPtr mon, diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 56ec65b..6cc6ea7 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -947,6 +947,25 @@ int qemuMonitorJSONSystemPowerdown(qemuMonitorPtr mon) } +int qemuMonitorJSONSystemReset(qemuMonitorPtr mon) +{ + int ret; + virJSONValuePtr cmd = qemuMonitorJSONMakeCommand("system_reset", NULL); + virJSONValuePtr reply = NULL; + if (!cmd) + return -1; + + ret = qemuMonitorJSONCommand(mon, cmd, &reply); + + if (ret == 0) + ret = qemuMonitorJSONCheckError(cmd, reply); + + virJSONValueFree(cmd); + virJSONValueFree(reply); + return ret; +} + + /* * [ { "CPU": 0, "current": true, "halted": false, "pc": 3227107138 }, * { "CPU": 1, "current": false, "halted": true, "pc": 7108165 } ] diff --git a/src/qemu/qemu_monitor_json.h b/src/qemu/qemu_monitor_json.h index 393d8fc..a72bf7c 100644 --- a/src/qemu/qemu_monitor_json.h +++ b/src/qemu/qemu_monitor_json.h @@ -49,6 +49,7 @@ int qemuMonitorJSONStopCPUs(qemuMonitorPtr mon); int qemuMonitorJSONGetStatus(qemuMonitorPtr mon, bool *running); int qemuMonitorJSONSystemPowerdown(qemuMonitorPtr mon); +int qemuMonitorJSONSystemReset(qemuMonitorPtr mon); int qemuMonitorJSONGetCPUInfo(qemuMonitorPtr mon, int **pids); diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index a16ea91..67333aa 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -417,6 +417,19 @@ int qemuMonitorTextSystemPowerdown(qemuMonitorPtr mon) { } +int qemuMonitorTextSystemReset(qemuMonitorPtr mon) { + char *info; + + if (qemuMonitorHMPCommand(mon, "system_reset", &info) < 0) { + qemuReportError(VIR_ERR_OPERATION_FAILED, + "%s", _("system reset operation failed")); + return -1; + } + VIR_FREE(info); + return 0; +} + + int qemuMonitorTextGetCPUInfo(qemuMonitorPtr mon, int **pids) { diff --git a/src/qemu/qemu_monitor_text.h b/src/qemu/qemu_monitor_text.h index 4fa5064..7536557 100644 --- a/src/qemu/qemu_monitor_text.h +++ b/src/qemu/qemu_monitor_text.h @@ -46,6 +46,7 @@ int qemuMonitorTextStopCPUs(qemuMonitorPtr mon); int qemuMonitorTextGetStatus(qemuMonitorPtr mon, bool *running); int qemuMonitorTextSystemPowerdown(qemuMonitorPtr mon); +int qemuMonitorTextSystemReset(qemuMonitorPtr mon); int qemuMonitorTextGetCPUInfo(qemuMonitorPtr mon, int **pids); diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 6838417..694d8a9 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -348,12 +348,99 @@ qemuProcessHandleReset(qemuMonitorPtr mon ATTRIBUTE_UNUSED, } +/* + * Since we have the '-no-shutdown' flag set, the + * QEMU process will currently have guest OS shutdown + * and the CPUS stopped. To fake the reboot, we thus + * want todo a reset of the virtual hardware, followed + * by restart of the CPUs. This should result in the + * guest OS booting up again + */ +static void +qemuProcessFakeReboot(void *opaque) +{ + struct qemud_driver *driver = qemu_driver; + virDomainObjPtr vm = opaque; + qemuDomainObjPrivatePtr priv = vm->privateData; + virDomainEventPtr event = NULL; + int ret = -1; + VIR_DEBUG("vm=%p", vm); + qemuDriverLock(driver); + virDomainObjLock(vm); + if (qemuDomainObjBeginJob(vm) < 0) + goto cleanup; + + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("guest unexpectedly quit")); + goto endjob; + } + + qemuDomainObjEnterMonitorWithDriver(driver, vm); + if (qemuMonitorSystemReset(priv->mon) < 0) { + qemuDomainObjExitMonitorWithDriver(driver, vm); + goto endjob; + } + qemuDomainObjExitMonitorWithDriver(driver, vm); + + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("guest unexpectedly quit")); + goto endjob; + } + + if (qemuProcessStartCPUs(driver, vm, NULL, + VIR_DOMAIN_RUNNING_BOOTED) < 0) { + if (virGetLastError() == NULL) + qemuReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("resume operation failed")); + goto endjob; + } + event = virDomainEventNewFromObj(vm, + VIR_DOMAIN_EVENT_RESUMED, + VIR_DOMAIN_EVENT_RESUMED_UNPAUSED); + + ret = 0; + +endjob: + if (qemuDomainObjEndJob(vm) == 0) + vm = NULL; + +cleanup: + if (vm) { + if (ret == -1) + qemuProcessKill(vm); + if (virDomainObjUnref(vm) > 0) + virDomainObjUnlock(vm); + } + if (event) + qemuDomainEventQueue(driver, event); + qemuDriverUnlock(driver); +} + + static int qemuProcessHandleShutdown(qemuMonitorPtr mon ATTRIBUTE_UNUSED, virDomainObjPtr vm) { + qemuDomainObjPrivatePtr priv = vm->privateData; + VIR_DEBUG("vm=%p", vm); + virDomainObjLock(vm); - ((qemuDomainObjPrivatePtr) vm->privateData)->gotShutdown = true; + priv->gotShutdown = true; + if (priv->fakeReboot) { + virDomainObjRef(vm); + virThread th; + if (virThreadCreate(&th, + false, + qemuProcessFakeReboot, + vm) < 0) { + VIR_ERROR("Failed to create reboot thread, killing domain"); + qemuProcessKill(vm); + } + } else { + qemuProcessKill(vm); + } virDomainObjUnlock(vm); return 0; @@ -1920,6 +2007,11 @@ qemuProcessPrepareMonitorChr(struct qemud_driver *driver, } +/* + * Precondition: Both driver and vm must be locked, + * and a job must be active. This method will call + * {Enter,Exit}MonitorWithDriver + */ int qemuProcessStartCPUs(struct qemud_driver *driver, virDomainObjPtr vm, virConnectPtr conn, virDomainRunningReason reason) @@ -2182,6 +2274,7 @@ int qemuProcessStart(virConnectPtr conn, goto cleanup; vm->def->id = driver->nextvmid++; + priv->fakeReboot = false; /* Run an early hook to set-up missing devices */ if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) { -- 1.7.4.4

2 4

[libvirt] FreeBSD portability question, --network question
by Jason Helfman 21 Jun '11

21 Jun '11

Hey Everyone, I've ported libvirt to FreeBSD, however it has the networking disabled for the build, however I was curious what would need to be changed for this to work. Currently, we do have virtualbox-ose in the FreeBSD portstree, and I am curious if the networking would need to be reworked to support libvirt with networking in FreeBSD in the same manner in which virtualbox creates interfaces and uses the networking stack? Basically, I am seeking to learn what is missing, and how to either find resources to get this implemented, or implement it myself. The latter of implementing it myself is unlikely, as I really don't have very much source control experience at a networking stack layer. Thanks very much, Jason -- Jason Helfman System Administrator experts-exchange.com http://www.experts-exchange.com/M_4830110.html E4AD 7CF1 1396 27F6 79DD 4342 5E92 AD66 8C8C FBA5

2 1

[libvirt] [PATCH v3]: set and restore MAC address of a NIC when using PASSTHROUGH mode
by Gerhard Stenzel 21 Jun '11

21 Jun '11

This is another rework of the patch from Dirk addressing all comments received so far. The following patch addresses the problem that when a PASSTHROUGH mode DIRECT NIC connection is made the MAC address of the NIC is not automatically set and reset to the configured VM MAC and back again. The attached patch fixes this problem by setting and resetting the MAC while remembering the previous setting while the VM is running. This also works if libvirtd is restarted while the VM is running. the patch passes make syntax-check Signed-off-by: Dirk Herrendoerfer <d.herrendoerfer at herrendoerfer.name> Signed-off-by: Gerhard Stenzel <gerhard.stenzel(a)de.ibm.com> --- Index: libvirt/src/qemu/qemu_command.c =================================================================== --- libvirt.orig/src/qemu/qemu_command.c +++ libvirt/src/qemu/qemu_command.c @@ -128,7 +128,7 @@ qemuPhysIfaceConnect(virDomainDefPtr def rc = openMacvtapTap(net->ifname, net->mac, net->data.direct.linkdev, net->data.direct.mode, vnet_hdr, def->uuid, &net->data.direct.virtPortProfile, &res_ifname, - vmop); + vmop, driver->stateDir); if (rc >= 0) { qemuAuditNetDevice(def, net, res_ifname, true); VIR_FREE(net->ifname); @@ -149,7 +149,9 @@ qemuPhysIfaceConnect(virDomainDefPtr def if (err) { VIR_FORCE_CLOSE(rc); delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, - &net->data.direct.virtPortProfile); + net->data.direct.mode, + &net->data.direct.virtPortProfile, + driver->stateDir); VIR_FREE(net->ifname); } } Index: libvirt/src/qemu/qemu_process.c =================================================================== --- libvirt.orig/src/qemu/qemu_process.c +++ libvirt/src/qemu/qemu_process.c @@ -2876,7 +2876,8 @@ void qemuProcessStop(struct qemud_driver virDomainNetDefPtr net = def->nets[i]; if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, - &net->data.direct.virtPortProfile); + net->data.direct.mode, + &net->data.direct.virtPortProfile, driver->stateDir); VIR_FREE(net->ifname); } } Index: libvirt/src/util/macvtap.c =================================================================== --- libvirt.orig/src/util/macvtap.c +++ libvirt/src/util/macvtap.c @@ -545,6 +545,106 @@ configMacvtapTap(int tapfd, int vnet_hdr return 0; } +/** + * replaceMacAdress: + * @macaddress: new MAC address for interface + * @linkdev: name of interface + * @stateDir: directory to store old MAC address + * + * Returns 0 on success, -1 in case of fatal error, error code otherwise. + * + */ +static int +replaceMacAdress(const unsigned char *macaddress, + const char *linkdev, + char *stateDir) +{ + unsigned char oldmac[6]; + int rc; + + rc = ifaceGetMacaddr(linkdev, oldmac); + + if (rc) { + virReportSystemError(rc, + _("Getting MAC address from '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]); + } else { + char *path = NULL; + + char macstr[VIR_MAC_STRING_BUFLEN]; + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + return errno; + } + virFormatMacAddr(oldmac, macstr); + if (virFileWriteStr(path, macstr, O_CREAT|O_TRUNC|O_WRONLY) < 0) { + virReportSystemError(errno, _("Unable to preserve mac for %s"), + linkdev); + return errno; + } + } + + rc = ifaceSetMacaddr(linkdev, macaddress); + if (rc) { + virReportSystemError(errno, + _("Setting MAC address on '%s' to " + "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + macaddress[0], macaddress[1], macaddress[2], + macaddress[3], macaddress[4], macaddress[5]); + } + return rc; +} + +/** + * restoreMacAddress: + * @linkdev: name of interface + * @stateDir: directory containing old MAC address + * + * Returns 0 on success, -1 in case of fatal error, error code otherwise. + * + */ +static int +restoreMacAddress(const char *linkdev, + char *stateDir) +{ + int ret; + char *oldmacname = NULL; + char *macstr = NULL; + char *path = NULL; + unsigned char oldmac[6]; + + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + return -1; + } + + if (virFileReadAll(path, VIR_MAC_STRING_BUFLEN, &macstr) < 0) { + virReportSystemError(errno, _("Unable to preserve mac for %s"), + linkdev); + return errno; + } + + if (virParseMacAddr(macstr, &oldmac[0]) != 0) { + virReportSystemError(VIR_ERR_INTERNAL_ERROR, + _("Cannot parse MAC address from '%s'"), + oldmacname); + return -1; + } + + /*reset mac and remove file-ignore results*/ + ret = ifaceSetMacaddr(linkdev, oldmac); + ret = unlink(path); + VIR_FREE(macstr); + return 0; +} /** * openMacvtapTap: @@ -575,7 +675,8 @@ openMacvtapTap(const char *tgifname, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, char **res_ifname, - enum virVMOperationType vmOp) + enum virVMOperationType vmOp, + char *stateDir) { const char *type = "macvtap"; int c, rc; @@ -589,6 +690,21 @@ openMacvtapTap(const char *tgifname, VIR_DEBUG("%s: VM OPERATION: %s", __FUNCTION__, virVMOperationTypeToString(vmOp)); + /** Note: When using PASSTHROUGH mode with MACVTAP devices the link + * device's MAC address must be set to the VMs MAC address. In + * order to not confuse the first switch or bridge in line this MAC + * address must be reset when the VM is shut down. + * This is especially important when using SRIOV capable cards that + * emulate their switch in firmware. + */ + if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + if (replaceMacAdress(macaddress, linkdev, stateDir) != 0) { + virReportSystemError(errno, + _("cannot replace MAC address on $s"), + linkdev); + } + } + if (tgifname) { if(ifaceGetIndex(false, tgifname, &ifindex) == 0) { if (STRPREFIX(tgifname, @@ -684,8 +800,18 @@ void delMacvtap(const char *ifname, const unsigned char *macaddr, const char *linkdev, - virVirtualPortProfileParamsPtr virtPortProfile) + int mode, + virVirtualPortProfileParamsPtr virtPortProfile, + char *stateDir) { + if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + if (restoreMacAddress(linkdev, stateDir) != 0) { + virReportSystemError(errno, + _("cannot restore MAC address on $s"), + linkdev); + } + } + if (ifname) { vpDisassociatePortProfileId(ifname, macaddr, linkdev, Index: libvirt/src/util/macvtap.h =================================================================== --- libvirt.orig/src/util/macvtap.h +++ libvirt/src/util/macvtap.h @@ -83,12 +83,15 @@ int openMacvtapTap(const char *ifname, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, char **res_ifname, - enum virVMOperationType vmop); + enum virVMOperationType vmop, + char *stateDir); void delMacvtap(const char *ifname, const unsigned char *macaddress, const char *linkdev, - virVirtualPortProfileParamsPtr virtPortProfile); + int mode, + virVirtualPortProfileParamsPtr virtPortProfile, + char *stateDir); int vpAssociatePortProfileId(const char *macvtap_ifname, const unsigned char *macvtap_macaddr, Index: libvirt/src/qemu/qemu_hotplug.c =================================================================== --- libvirt.orig/src/qemu/qemu_hotplug.c +++ libvirt/src/qemu/qemu_hotplug.c @@ -1610,7 +1610,9 @@ int qemuDomainDetachNetDevice(struct qem #if WITH_MACVTAP if (detach->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(detach->ifname, detach->mac, detach->data.direct.linkdev, - &detach->data.direct.virtPortProfile); + detach->data.direct.mode, + &detach->data.direct.virtPortProfile, + driver->stateDir); VIR_FREE(detach->ifname); } #endif Index: libvirt/src/util/interface.c =================================================================== --- libvirt.orig/src/util/interface.c +++ libvirt/src/util/interface.c @@ -390,3 +390,102 @@ ifaceGetVlanID(const char *vlanifname AT return ENOSYS; } #endif /* __linux__ */ + +/** + * ifaceGetMacaddr: + * @ifname: interface name to set MTU for + * @macaddr: MAC address (VIR_MAC_BUFLEN in size) + * + * This function gets the @macaddr for a given interface @ifname. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +#ifdef __linux__ +int +ifaceGetMacaddr(const char *ifname, + unsigned char *macaddr) +{ + struct ifreq ifr; + int fd; + + if (!ifname) + return EINVAL; + + fd = socket(AF_INET, SOCK_STREAM, 0); + if(fd < 0){ + return errno; + } + + memset(&ifr, 0, sizeof(struct ifreq)); + if (virStrcpyStatic(ifr.ifr_name, ifname) == NULL) + return EINVAL; + + if(ioctl(fd, SIOCGIFHWADDR, (char *)&ifr) != 0){ + return errno; + } + + memcpy(macaddr, ifr.ifr_ifru.ifru_hwaddr.sa_data, VIR_MAC_BUFLEN); + + return 0; +} + +#else + +int +ifaceGetMacaddr(const char *ifname, + unsigned char *macaddr) +{ + return ENOSYS; +} + +#endif /* __linux__ */ + +/** + * ifaceSetMacaddr: + * @ifname: interface name to set MTU for + * @macaddr: MAC address (VIR_MAC_BUFLEN in size) + * + * This function sets the @macaddr for a given interface @ifname. This + * gets rid of the kernel's automatically assigned random MAC. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +#ifdef __linux__ +int +ifaceSetMacaddr(const char *ifname, + const unsigned char *macaddr) +{ + struct ifreq ifr; + int fd; + + if (!ifname) + return EINVAL; + + fd = socket(AF_INET, SOCK_STREAM, 0); + if(fd < 0){ + return errno; + } + + memset(&ifr, 0, sizeof(struct ifreq)); + if (virStrcpyStatic(ifr.ifr_name, ifname) == NULL) + return EINVAL; + + /* To fill ifr.ifr_hdaddr.sa_family field */ + if (ioctl(fd, SIOCGIFHWADDR, &ifr) != 0) + return errno; + + memcpy(ifr.ifr_hwaddr.sa_data, macaddr, VIR_MAC_BUFLEN); + + return ioctl(fd, SIOCSIFHWADDR, &ifr) == 0 ? 0 : errno; +} + +#else + +int +ifaceSetMacaddr(const char *ifname, + const unsigned char *macaddr) +{ + return ENOSYS; +} + +#endif /* __linux__ */ Index: libvirt/src/util/interface.h =================================================================== --- libvirt.orig/src/util/interface.h +++ libvirt/src/util/interface.h @@ -32,4 +32,8 @@ int ifaceGetIndex(bool reportError, cons int ifaceGetVlanID(const char *vlanifname, int *vlanid); +int ifaceSetMacaddr(const char *ifname, const unsigned char *macaddr); + +int ifaceGetMacaddr(const char *ifname, unsigned char *macaddr); + #endif /* __VIR_INTERFACE_H__ */ Index: libvirt/src/libvirt_private.syms =================================================================== --- libvirt.orig/src/libvirt_private.syms +++ libvirt/src/libvirt_private.syms @@ -507,7 +507,9 @@ ifaceCheck; ifaceCtrl; ifaceGetFlags; ifaceGetIndex; +ifaceGetMacaddr; ifaceGetVlanID; +ifaceSetMacaddr; ifaceIsUp; =================================================================== Best regards, Gerhard Stenzel ------------------------------------------------------------------------------------- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Martin Jetter Geschaeftsfuehrung: Dirk Wittkopp Sitz der Gesellschaft: Boeblingen Registergericht: Amtsgericht Stuttgart, HRB 243294

3 2

[libvirt] [PATCH 0/4] RFC: vcpupin: some improvements of virsh vcpupin command
by Taku Izumi 21 Jun '11

21 Jun '11

Hi all, I'll resend this patchset because what sent yesterday was broken. Sorry to be a nuisance. --- This patchset improves virsh vcpupin command like the following: (i) introduce special markup "-" and "^" which are similar to XML schema of "cpuset" attribute # virsh vcpupin VM 0 0-8,^4 --config # virsh dumpxml VM ... <vcpu>4</vcpu> <cputune> <vcpupin vcpu='0' cpuset='0-3,5-8'/> </cputune> ... (ii) allow to receive the special keyword 'r' as a cpulist parameter when resetting vcpupin setting. # virsh vcpupin VM 0 r --config (iii) when resetting, that is, pinning the specified virtual cpu to all host physical cpus, vcpupin setting will be deleted # virsh dumpxml VM ... <vcpu>4</vcpu> <cputune> </cputune> ... The following my patchset is the prerequisite of this: http://www.redhat.com/archives/libvir-list/2011-June/msg00441.html *[PATCH 1/4] vcpupin: improve vcpupin definition of virsh vcpupin *[PATCH 2/4] vcpupin: add reset option to virsh vcpupin command *[PATCH 3/4] vcpupin: add virDomainVcpupinDel function *[PATCH 4/4] vcpupin: add vcpupin resetting feature to qemu driver Best regards, Taku Izumi

4 12

[libvirt] [PATCH] conf: Make full use of goto lable
by Osier Yang 21 Jun '11

21 Jun '11

* virDomainDefParse: There is a goto label "no_memory", which reports OOM error, and it then fallthrough lable "error". This patch changes things like following: virReportOOMError(); goto error; into: goto no_memory; --- src/conf/domain_conf.c | 21 +++++++-------------- 1 files changed, 7 insertions(+), 14 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 5360863..f80ab43 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -5832,8 +5832,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, char *set = tmp; def->cpumasklen = VIR_DOMAIN_CPUMASK_LEN; if (VIR_ALLOC_N(def->cpumask, def->cpumasklen) < 0) { - virReportOOMError(); - goto error; + goto no_memory; } if (virDomainCpuSetParse((const char **)&set, 0, def->cpumask, @@ -5894,8 +5893,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, int nodemasklen = VIR_DOMAIN_CPUMASK_LEN; if (VIR_ALLOC_N(def->numatune.memory.nodemask, nodemasklen) < 0) { - virReportOOMError(); - goto error; + goto no_memory; } /* "nodeset" leads same syntax with "cpuset". */ @@ -6010,8 +6008,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, if (def->os.bootloader) { def->os.type = strdup("xen"); if (!def->os.type) { - virReportOOMError(); - goto error; + goto no_memory; } } else { virDomainReportError(VIR_ERR_OS_TYPE, @@ -6028,8 +6025,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, def->virtType == VIR_DOMAIN_VIRT_XEN) { VIR_FREE(def->os.type); if (!(def->os.type = strdup("xen"))) { - virReportOOMError(); - goto error; + goto no_memory; } } @@ -6056,8 +6052,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, goto error; } if (!(def->os.arch = strdup(defaultArch))) { - virReportOOMError(); - goto error; + goto no_memory; } } @@ -6069,8 +6064,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, virDomainVirtTypeToString(def->virtType)); if (defaultMachine != NULL) { if (!(def->os.machine = strdup(defaultMachine))) { - virReportOOMError(); - goto error; + goto no_memory; } } } @@ -6403,8 +6397,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, virDomainInputDefPtr input; if (VIR_ALLOC(input) < 0) { - virReportOOMError(); - goto error; + goto no_memory; } if (STREQ(def->os.type, "hvm")) { input->type = VIR_DOMAIN_INPUT_TYPE_MOUSE; -- 1.7.4

2 2

[libvirt] [PATCH v2] Add support for fd: protocol
by Corey Bryant 21 Jun '11

21 Jun '11

sVirt provides SELinux MAC isolation for Qemu guest processes and their corresponding resources (image files). sVirt provides this support by labeling guests and resources with security labels that are stored in file system extended attributes. Some file systems, such as NFS, do not support the extended attribute security namespace, which is needed for image file isolation when using the sVirt SELinux security driver in libvirt. The proposed solution entails a combination of Qemu, libvirt, and SELinux patches that work together to isolate multiple guests' images when they're stored in the same NFS mount. This results in an environment where sVirt isolation and NFS image file isolation can both be provided. This patch contains the Qemu code to support this solution. I would like to solicit input from the libvirt community prior to starting the libvirt patch. Currently, Qemu opens an image file in addition to performing the necessary read and write operations. The proposed solution will move the open out of Qemu and into libvirt. Once libvirt opens an image file for the guest, it will pass the file descriptor to Qemu via a new fd: protocol. If the image file resides in an NFS mount, the following SELinux policy changes will provide image isolation: - A new SELinux boolean is created (e.g. virt_read_write_nfs) to allow Qemu (svirt_t) to only have SELinux read and write permissions on nfs_t files - Qemu (svirt_t) also gets SELinux use permissions on libvirt (virtd_t) file descriptors Following is a sample invocation of Qemu using the fd: protocol on the command line: qemu -drive file=fd:4,format=qcow2 The fd: protocol is also supported by the drive_add monitor command. This requires that the specified file descriptor is passed to the monitor alongside a prior getfd monitor command. There are some additional features provided by certain image types where Qemu reopens the image file. All of these scenarios will be unsupported for the fd: protocol, at least for this patch: - The -snapshot command line option - The savevm monitor command - The snapshot_blkdev monitor command - Starting Qemu with a backing file The thought is that this support can be added in the future, but is not required for the initial fd: support. This patch was tested with the following formats: raw, cow, qcow, qcow2, qed, and vmdk, using the fd: protocol from the command line and the monitor. Tests were also run to verify existing file name support and qemu-img were not regressed. Non-valid file descriptors, fd: without format, snapshot and backing files were also tested. Signed-off-by: Corey Bryant <coreyb(a)linux.vnet.ibm.com> --- block.c | 16 ++++++++++ block.h | 1 + block/cow.c | 5 +++ block/qcow.c | 5 +++ block/qcow2.c | 5 +++ block/qed.c | 4 ++ block/raw-posix.c | 81 +++++++++++++++++++++++++++++++++++++++++++++++------ block/vmdk.c | 5 +++ block_int.h | 1 + blockdev.c | 10 ++++++ monitor.c | 5 +++ monitor.h | 1 + qemu-options.hx | 8 +++-- qemu-tool.c | 5 +++ 14 files changed, 140 insertions(+), 12 deletions(-) diff --git a/block.c b/block.c index 24a25d5..500db84 100644 --- a/block.c +++ b/block.c @@ -536,6 +536,10 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags, char tmp_filename[PATH_MAX]; char backing_filename[PATH_MAX]; + if (bdrv_is_fd_protocol(bs)) { + return -ENOTSUP; + } + /* if snapshot, we create a temporary backing file and open it instead of opening 'filename' directly */ @@ -585,6 +589,10 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags, /* Find the right image format driver */ if (!drv) { + /* format must be specified for fd: protocol */ + if (bdrv_is_fd_protocol(bs)) { + return -ENOTSUP; + } ret = find_image_format(filename, &drv); } @@ -1460,6 +1468,11 @@ int bdrv_enable_write_cache(BlockDriverState *bs) return bs->enable_write_cache; } +int bdrv_is_fd_protocol(BlockDriverState *bs) +{ + return bs->fd_protocol; +} + /* XXX: no longer used */ void bdrv_set_change_cb(BlockDriverState *bs, void (*change_cb)(void *opaque, int reason), @@ -1964,6 +1977,9 @@ int bdrv_snapshot_create(BlockDriverState *bs, BlockDriver *drv = bs->drv; if (!drv) return -ENOMEDIUM; + if (bdrv_is_fd_protocol(bs)) { + return -ENOTSUP; + } if (drv->bdrv_snapshot_create) return drv->bdrv_snapshot_create(bs, sn_info); if (bs->file) diff --git a/block.h b/block.h index da7d39c..dd46d52 100644 --- a/block.h +++ b/block.h @@ -182,6 +182,7 @@ int bdrv_is_removable(BlockDriverState *bs); int bdrv_is_read_only(BlockDriverState *bs); int bdrv_is_sg(BlockDriverState *bs); int bdrv_enable_write_cache(BlockDriverState *bs); +int bdrv_is_fd_protocol(BlockDriverState *bs); int bdrv_is_inserted(BlockDriverState *bs); int bdrv_media_changed(BlockDriverState *bs); int bdrv_is_locked(BlockDriverState *bs); diff --git a/block/cow.c b/block/cow.c index 4cf543c..e17f8e7 100644 --- a/block/cow.c +++ b/block/cow.c @@ -82,6 +82,11 @@ static int cow_open(BlockDriverState *bs, int flags) pstrcpy(bs->backing_file, sizeof(bs->backing_file), cow_header.backing_file); + if (bs->backing_file[0] != '\0' && bdrv_is_fd_protocol(bs)) { + /* backing file currently not supported by fd: protocol */ + goto fail; + } + bitmap_size = ((bs->total_sectors + 7) >> 3) + sizeof(cow_header); s->cow_sectors_offset = (bitmap_size + 511) & ~511; return 0; diff --git a/block/qcow.c b/block/qcow.c index a26c886..1e46bdb 100644 --- a/block/qcow.c +++ b/block/qcow.c @@ -157,6 +157,11 @@ static int qcow_open(BlockDriverState *bs, int flags) if (bdrv_pread(bs->file, header.backing_file_offset, bs->backing_file, len) != len) goto fail; bs->backing_file[len] = '\0'; + + if (bs->backing_file[0] != '\0' && bdrv_is_fd_protocol(bs)) { + /* backing file currently not supported by fd: protocol */ + goto fail; + } } return 0; diff --git a/block/qcow2.c b/block/qcow2.c index 8451ded..8b9c160 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -270,6 +270,11 @@ static int qcow2_open(BlockDriverState *bs, int flags) goto fail; } bs->backing_file[len] = '\0'; + + if (bs->backing_file[0] != '\0' && bdrv_is_fd_protocol(bs)) { + ret = -ENOTSUP; + goto fail; + } } if (qcow2_read_snapshots(bs) < 0) { ret = -EINVAL; diff --git a/block/qed.c b/block/qed.c index 3970379..5028897 100644 --- a/block/qed.c +++ b/block/qed.c @@ -446,6 +446,10 @@ static int bdrv_qed_open(BlockDriverState *bs, int flags) return ret; } + if (bs->backing_file[0] != '\0' && bdrv_is_fd_protocol(bs)) { + return -ENOTSUP; + } + if (s->header.features & QED_F_BACKING_FORMAT_NO_PROBE) { pstrcpy(bs->backing_format, sizeof(bs->backing_format), "raw"); } diff --git a/block/raw-posix.c b/block/raw-posix.c index 4cd7d7a..c72de3d 100644 --- a/block/raw-posix.c +++ b/block/raw-posix.c @@ -28,6 +28,7 @@ #include "block_int.h" #include "module.h" #include "block/raw-posix-aio.h" +#include "monitor.h" #ifdef CONFIG_COCOA #include <paths.h> @@ -183,7 +184,8 @@ static int raw_open_common(BlockDriverState *bs, const char *filename, int bdrv_flags, int open_flags) { BDRVRawState *s = bs->opaque; - int fd, ret; + int fd = -1; + int ret; ret = raw_normalize_devicepath(&filename); if (ret != 0) { @@ -205,15 +207,17 @@ static int raw_open_common(BlockDriverState *bs, const char *filename, if (!(bdrv_flags & BDRV_O_CACHE_WB)) s->open_flags |= O_DSYNC; - s->fd = -1; - fd = qemu_open(filename, s->open_flags, 0644); - if (fd < 0) { - ret = -errno; - if (ret == -EROFS) - ret = -EACCES; - return ret; + if (s->fd == -1) { + fd = qemu_open(filename, s->open_flags, 0644); + if (fd < 0) { + ret = -errno; + if (ret == -EROFS) { + ret = -EACCES; + } + return ret; + } + s->fd = fd; } - s->fd = fd; s->aligned_buf = NULL; if ((bdrv_flags & BDRV_O_NOCACHE)) { @@ -270,6 +274,7 @@ static int raw_open(BlockDriverState *bs, const char *filename, int flags) { BDRVRawState *s = bs->opaque; + s->fd = -1; s->type = FTYPE_FILE; return raw_open_common(bs, filename, flags, 0); } @@ -890,6 +895,60 @@ static BlockDriver bdrv_file = { .create_options = raw_create_options, }; +static int raw_open_fd(BlockDriverState *bs, const char *filename, int flags) +{ + BDRVRawState *s = bs->opaque; + const char *fd_str; + int fd; + + /* extract the file descriptor - fail if it's not fd: */ + if (!strstart(filename, "fd:", &fd_str)) { + return -EINVAL; + } + + if (!qemu_isdigit(fd_str[0])) { + /* get fd from monitor */ + fd = qemu_get_fd(fd_str); + if (fd == -1) { + return -EBADF; + } + } else { + char *endptr = NULL; + + fd = strtol(fd_str, &endptr, 10); + if (*endptr || (fd == 0 && fd_str == endptr)) { + return -EBADF; + } + } + + s->fd = fd; + s->type = FTYPE_FILE; + + return raw_open_common(bs, filename, flags, 0); +} + +static BlockDriver bdrv_file_fd = { + .format_name = "file", + .protocol_name = "fd", + .instance_size = sizeof(BDRVRawState), + .bdrv_probe = NULL, /* no probe for protocols */ + .bdrv_file_open = raw_open_fd, + .bdrv_read = raw_read, + .bdrv_write = raw_write, + .bdrv_close = raw_close, + .bdrv_flush = raw_flush, + .bdrv_discard = raw_discard, + + .bdrv_aio_readv = raw_aio_readv, + .bdrv_aio_writev = raw_aio_writev, + .bdrv_aio_flush = raw_aio_flush, + + .bdrv_truncate = raw_truncate, + .bdrv_getlength = raw_getlength, + + .create_options = raw_create_options, +}; + /***********************************************/ /* host device */ @@ -998,6 +1057,7 @@ static int hdev_open(BlockDriverState *bs, const char *filename, int flags) } #endif + s->fd = -1; s->type = FTYPE_FILE; #if defined(__linux__) { @@ -1168,6 +1228,7 @@ static int floppy_open(BlockDriverState *bs, const char *filename, int flags) BDRVRawState *s = bs->opaque; int ret; + s->fd = -1; s->type = FTYPE_FD; /* open will not fail even if no floppy is inserted, so add O_NONBLOCK */ @@ -1280,6 +1341,7 @@ static int cdrom_open(BlockDriverState *bs, const char *filename, int flags) { BDRVRawState *s = bs->opaque; + s->fd = -1; s->type = FTYPE_CD; /* open will not fail even if no CD is inserted, so add O_NONBLOCK */ @@ -1503,6 +1565,7 @@ static void bdrv_file_init(void) * Register all the drivers. Note that order is important, the driver * registered last will get probed first. */ + bdrv_register(&bdrv_file_fd); bdrv_register(&bdrv_file); bdrv_register(&bdrv_host_device); #ifdef __linux__ diff --git a/block/vmdk.c b/block/vmdk.c index 922b23d..2ea808e 100644 --- a/block/vmdk.c +++ b/block/vmdk.c @@ -353,6 +353,11 @@ static int vmdk_parent_open(BlockDriverState *bs) return -1; pstrcpy(bs->backing_file, end_name - p_name + 1, p_name); + + if (bs->backing_file[0] != '\0' && bdrv_is_fd_protocol(bs)) { + /* backing file currently not supported by fd: protocol */ + return -1; + } } return 0; diff --git a/block_int.h b/block_int.h index fa91337..a305ee2 100644 --- a/block_int.h +++ b/block_int.h @@ -152,6 +152,7 @@ struct BlockDriverState { int encrypted; /* if true, the media is encrypted */ int valid_key; /* if true, a valid encryption key has been set */ int sg; /* if true, the device is a /dev/sg* */ + int fd_protocol; /* if true, the fd: protocol was specified */ /* event callback when inserting/removing */ void (*change_cb)(void *opaque, int reason); void *change_opaque; diff --git a/blockdev.c b/blockdev.c index e81e0ab..a536c20 100644 --- a/blockdev.c +++ b/blockdev.c @@ -546,6 +546,10 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi) bdrv_flags |= ro ? 0 : BDRV_O_RDWR; + if (strncmp(file, "fd:", 3) == 0) { + dinfo->bdrv->fd_protocol = 1; + } + ret = bdrv_open(dinfo->bdrv, file, bdrv_flags, drv); if (ret < 0) { error_report("could not open disk image %s: %s", @@ -606,6 +610,12 @@ int do_snapshot_blkdev(Monitor *mon, const QDict *qdict, QObject **ret_data) goto out; } + if (bdrv_is_fd_protocol(bs)) { + qerror_report(QERR_UNSUPPORTED); + ret = -1; + goto out; + } + pstrcpy(old_filename, sizeof(old_filename), bs->filename); old_drv = bs->drv; diff --git a/monitor.c b/monitor.c index 59a3e76..ea60be2 100644 --- a/monitor.c +++ b/monitor.c @@ -2832,6 +2832,11 @@ int monitor_get_fd(Monitor *mon, const char *fdname) return -1; } +int qemu_get_fd(const char *fdname) +{ + return cur_mon ? monitor_get_fd(cur_mon, fdname) : -1; +} + static const mon_cmd_t mon_cmds[] = { #include "hmp-commands.h" { NULL, NULL, }, diff --git a/monitor.h b/monitor.h index 4f2d328..de5b987 100644 --- a/monitor.h +++ b/monitor.h @@ -51,6 +51,7 @@ int monitor_read_bdrv_key_start(Monitor *mon, BlockDriverState *bs, void *opaque); int monitor_get_fd(Monitor *mon, const char *fdname); +int qemu_get_fd(const char *fdname); void monitor_vprintf(Monitor *mon, const char *fmt, va_list ap) GCC_FMT_ATTR(2, 0); diff --git a/qemu-options.hx b/qemu-options.hx index 1d5ad8b..f9b66f4 100644 --- a/qemu-options.hx +++ b/qemu-options.hx @@ -116,7 +116,7 @@ using @file{/dev/cdrom} as filename (@pxref{host_drives}). ETEXI DEF("drive", HAS_ARG, QEMU_OPTION_drive, - "-drive [file=file][,if=type][,bus=n][,unit=m][,media=d][,index=i]\n" + "-drive [file=[fd:]file][,if=type][,bus=n][,unit=m][,media=d][,index=i]\n" " [,cyls=c,heads=h,secs=s[,trans=t]][,snapshot=on|off]\n" " [,cache=writethrough|writeback|none|unsafe][,format=f]\n" " [,serial=s][,addr=A][,id=name][,aio=threads|native]\n" @@ -129,10 +129,12 @@ STEXI Define a new drive. Valid options are: @table @option -@item file=@var{file} +@item file=[fd:]@var{file} This option defines which disk image (@pxref{disk_images}) to use with this drive. If the filename contains comma, you must double it -(for instance, "file=my,,file" to use file "my,file"). +(for instance, "file=my,,file" to use file "my,file"). @option{fd:}@var{file} +specifies the file descriptor of an already open disk image. +@option{format=}@var{format} is required by @option{fd:}@var{file}. @item if=@var{interface} This option defines on which type on interface the drive is connected. Available types are: ide, scsi, sd, mtd, floppy, pflash, virtio. diff --git a/qemu-tool.c b/qemu-tool.c index 41e5c41..8fe6b8c 100644 --- a/qemu-tool.c +++ b/qemu-tool.c @@ -96,3 +96,8 @@ int64_t qemu_get_clock_ns(QEMUClock *clock) { return 0; } + +int qemu_get_fd(const char *fdname) +{ + return -1; +} -- 1.7.1

6 14

[libvirt] [PATCH] conf: Fix one memory leak
by Osier Yang 21 Jun '11

21 Jun '11

Free def->numatune.memory.nodemask in virDomainDefFree. --- src/conf/domain_conf.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 5360863..38a8587 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -1092,6 +1092,8 @@ void virDomainDefFree(virDomainDefPtr def) virDomainVcpupinDefFree(def->cputune.vcpupin, def->cputune.nvcpupin); + VIR_FREE(def->numatune.memory.nodemask); + virSysinfoDefFree(def->sysinfo); if (def->namespaceData && def->ns.free) -- 1.7.4

2 2

[libvirt] [PATCH 0/7 v2] python: Implement missing stream/event ops
by Cole Robinson 20 Jun '11

20 Jun '11

Currently it isn't possible to implement the equivalent of 'virsh console' with the python bindings. This series implements the missing functionality required to do so. v2: Pushed first 3 patches Actually export virEvent APIs in libvirt.h DV had comments on patch 1 and patch 6 that I didn't address, if I haven't managed to convince him (or others) with my latest mails I can do another rev of the series. Cole Robinson (7): python: Implement virStreamSend/Recv python: Implement virStreamSend/RecvAll helpers Promote virEvent*Handle/Timeout to public API events: Correct virEventAddTimeout docs python: Add bindings for virEvent*Handle/Timeout python: events: Fix C->Python handle callback prototype python: Mark event callback wrappers as private daemon/libvirtd.c | 1 - daemon/mdns.c | 1 - examples/domain-events/events-python/event-test.py | 6 +- include/libvirt/libvirt.h.in | 14 + python/generator.py | 12 +- python/libvirt-override-virConnect.py | 102 +++---- python/libvirt-override-virStream.py | 101 +++++++- python/libvirt-override.c | 274 +++++++++++++++++--- python/libvirt-override.py | 80 ++++++- python/typewrappers.c | 14 + python/typewrappers.h | 1 + src/conf/domain_event.c | 1 - src/conf/domain_event.h | 1 - src/fdstream.c | 1 - src/libvirt_private.syms | 9 - src/libvirt_public.syms | 6 + src/libxl/libxl_driver.c | 1 - src/lxc/lxc_driver.c | 1 - src/network/bridge_driver.c | 1 - src/node_device/node_device_hal.c | 1 - src/node_device/node_device_udev.c | 1 - src/openvz/openvz_driver.c | 1 - src/qemu/qemu_domain.c | 1 - src/qemu/qemu_driver.c | 1 - src/qemu/qemu_monitor.c | 1 - src/remote/remote_driver.c | 1 - src/test/test_driver.c | 1 - src/uml/uml_driver.c | 1 - src/util/event.c | 56 ++++ src/util/event.h | 73 ------ src/util/util.c | 1 - src/vbox/vbox_tmpl.c | 1 - src/xen/xen_inotify.c | 1 - src/xen/xs_internal.c | 1 - tools/console.c | 1 - 35 files changed, 559 insertions(+), 211 deletions(-) -- 1.7.4.4

2 9

[libvirt] virt-dmesg and virt-uname
by Richard W.M. Jones 20 Jun '11

20 Jun '11

I just pushed the virt-dmesg package to Fedora Rawhide[1]. http://koji.fedoraproject.org/koji/buildinfo?buildID=248890 This tool uses libvirt to let you list the kernel messages from a Linux KVM guest without needing to log in or run any special agent in the guest. It's useful when a guest crashes or gets stuck. Usage is pretty simple: # virt-dmesg F14x64 | head <6>[ 0.000000] Initializing cgroup subsys cpuset <6>[ 0.000000] Initializing cgroup subsys cpu <5>[ 0.000000] Linux version 2.6.35.13-91.fc14.x86_64 (mockbuild(a)x86-15.phx2.fedoraproject.org) (gcc version 4.5.1 20100924 (Red Hat 4.5.1-4) (GCC) ) #1 SMP Tue May 3 13:23:06 UTC 2011 <6>[ 0.000000] Command line: ro root=/dev/mapper/vg_f13x64-lv_root rd_LVM_LV=vg_f13x64/lv_root rd_LVM_LV=vg_f13x64/lv_swap rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=uk rhgb quiet <6>[ 0.000000] BIOS-provided physical RAM map: <6>[ 0.000000] BIOS-e820: 0000000000000000 - 000000000009f400 (usable) <6>[ 0.000000] BIOS-e820: 000000000009f400 - 00000000000a0000 (reserved) <6>[ 0.000000] BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved) <6>[ 0.000000] BIOS-e820: 0000000000100000 - 000000003fffd000 (usable) <6>[ 0.000000] BIOS-e820: 000000003fffd000 - 0000000040000000 (reserved) In some cases virt-dmesg won't be able to find the symbol table, in which case you can fall back to dumping out the kernel and grepping it, eg: # virt-dmesg --dump-kernel F14x64 | strings | grep '^<[0-9]>' | less There's an accompanying tool called virt-uname which does what you would think: # virt-uname F14x64 Linux f14x64.home.annexia.org 2.6.35.13-91.fc14.x86_64 #1 SMP Tue May 3 13:23:06 UTC 2011 x86_64 (none) Please let me know if you find (Linux) guests for which virt-dmesg doesn't work properly (particularly the --dump-kernel option which ought to work for any Linux guest). Describe the guest as accurately as you can and provide the output of 'virt-dmesg -v Guest'. Rich. [1] Sadly because of some regressions in libvirt, it *won't* work in Fedora 15 or earlier unless you are able to upgrade to libvirt >= 0.9.1-3. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top

1 1

[libvirt] [PATCH 1/3] virsh: clarify snapshot vs. save
by Eric Blake 20 Jun '11

20 Jun '11

* tools/virsh.c (info_snapshot_create, info_save): Clarify description. * tools/virsh.pod (save): Likewise. --- tools/virsh.c | 6 +++--- tools/virsh.pod | 8 +++++++- 2 files changed, 10 insertions(+), 4 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 1c0b1dc..6b5f3d9 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -1469,7 +1469,7 @@ cmdStart(vshControl *ctl, const vshCmd *cmd) */ static const vshCmdInfo info_save[] = { {"help", N_("save a domain state to a file")}, - {"desc", N_("Save a running domain.")}, + {"desc", N_("Save the RAM state of a running domain.")}, {NULL, NULL} }; @@ -10947,8 +10947,8 @@ cmdQuit(vshControl *ctl, const vshCmd *cmd ATTRIBUTE_UNUSED) * "snapshot-create" command */ static const vshCmdInfo info_snapshot_create[] = { - {"help", N_("Create a snapshot")}, - {"desc", N_("Snapshot create")}, + {"help", N_("Create a snapshot from XML")}, + {"desc", N_("Create a snapshot (disk and RAM) from XML")}, {NULL, NULL} }; diff --git a/tools/virsh.pod b/tools/virsh.pod index c6879cb..33f72ed 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -594,7 +594,8 @@ the state file was created. =item B<save> I<domain-id> I<state-file> -Saves a running domain to a state file so that it can be restored +Saves a running domain (RAM, but not disk state) to a state file so that +it can be restored later. Once saved, the domain will no longer be running on the system, thus the memory allocated for the domain will be free for other domains to use. B<virsh restore> restores from this state file. @@ -603,6 +604,11 @@ This is roughly equivalent to doing a hibernate on a running computer, with all the same limitations. Open network connections may be severed upon restore, as TCP timeouts may have expired. +Domain saved state files assume that disk images will be unchanged +between the creation and restore point. For a more complete system +restore point, where the disk state is saved alongside the memory +state, see the B<snapshot> family of commands. + =item B<schedinfo> optional I<--set> B<parameter=value> I<domain-id> I<--config> I<--live> I<--current> -- 1.7.1

3 10

[libvirt] [PATCH v2]: set and restore MAC address of a NIC when using PASSTHROUGH mode
by Gerhard Stenzel 20 Jun '11

20 Jun '11

This is a rework of the patch Dirk sent out last week. The attached patch addresses the problem that when a PASSTHROUGH mode DIRECT NIC connection is made the MAC address of the NIC is not automatically set and reset to the configured VM MAC and back again. The patch fixes this problem by setting and resetting the MAC while remembering the previous setting while the VM is running. This also works if libvirtd is restarted while the VM is running. The patch passes make syntax-check Signed-off-by: Dirk Herrendoerfer <d.herrendoerfer at herrendoerfer.name> Signed-off-by: Gerhard Stenzel <gerhard.stenzel(a)de.ibm.com> --- src/libvirt_macvtap.syms | 2 + src/qemu/qemu_command.c | 48 +++++++++++++++ src/qemu/qemu_process.c | 45 ++++++++++++++ src/util/macvtap.c | 149 ++++++++++++++++++++++++++++++++++++++++++++++ src/util/macvtap.h | 6 ++ 5 files changed, 250 insertions(+), 0 deletions(-) Index: libvirt/src/libvirt_macvtap.syms =================================================================== --- libvirt.orig/src/libvirt_macvtap.syms +++ libvirt/src/libvirt_macvtap.syms @@ -5,6 +5,8 @@ # macvtap.h delMacvtap; +getMacaddr; openMacvtapTap; +setMacaddr; vpAssociatePortProfileId; vpDisassociatePortProfileId; Index: libvirt/src/qemu/qemu_command.c =================================================================== --- libvirt.orig/src/qemu/qemu_command.c +++ libvirt/src/qemu/qemu_command.c @@ -125,6 +125,77 @@ qemuPhysIfaceConnect(virDomainDefPtr def net->model && STREQ(net->model, "virtio")) vnet_hdr = 1; + /** Note: When using PASSTHROUGH mode with MACVTAP devices the link + * device's MAC address must be set to the VMs MAC address. In + * order to not confuse the first switch or bridge in line this MAC + * address must be reset when the VM is shut down. + * This is especially important when using SRIOV capable cards that + * emulate their switch in firmware. + */ + if (net->data.direct.mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + unsigned char oldmac[6]; + rc = getMacaddr(&oldmac, net->data.direct.linkdev); + if (rc) { + virReportSystemError(rc, + _("Getting MAC address from '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x' failed."), + net->data.direct.linkdev, + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]); + } else { + char *oldmacpath = NULL; + char *oldmacname = NULL; + char *newmacname = NULL; + char *basedir = "/var/run/libvirt/network"; + + if (virAsprintf(&oldmacname,"%02x:%02x:%02x:%02x:%02x:%02x", + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]) < 0) { + virReportOOMError(); + } + if (virAsprintf(&oldmacpath,"%s/%s", basedir, oldmacname) < 0) { + virReportOOMError(); + } + + if (virAsprintf(&newmacname, "%s/%s@%02x:%02x:%02x:%02x:%02x:%02x", + basedir, + net->data.direct.linkdev, + net->mac[0],net->mac[1],net->mac[2], + net->mac[3],net->mac[4],net->mac[5]) < 0) { + virReportOOMError(); + } + + rc = symlink (oldmacname, newmacname); + if (rc) { + virReportSystemError(errno, + _("MAC link file creation failed for %s."), + net->data.direct.linkdev); + } + rc = creat(oldmacpath, S_IWUSR); + if (rc) { + virReportSystemError(errno, + _("MAC link file creation failed for %s."), + oldmacpath); + } + rc = VIR_CLOSE(rc); + if (rc) { + virReportSystemError(errno, + _("MAC link file closing failed for %s."), + oldmacpath); + } + } + + rc = setMacaddr(net->mac, net->data.direct.linkdev); + if (rc) { + virReportSystemError(errno, + _("Setting MAC address on '%s' to " + "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), + net->data.direct.linkdev, + net->mac[0], net->mac[1], net->mac[2], + net->mac[3], net->mac[4], net->mac[5]); + } + } + rc = openMacvtapTap(net->ifname, net->mac, net->data.direct.linkdev, net->data.direct.mode, vnet_hdr, def->uuid, &net->data.direct.virtPortProfile, &res_ifname, Index: libvirt/src/qemu/qemu_process.c =================================================================== --- libvirt.orig/src/qemu/qemu_process.c +++ libvirt/src/qemu/qemu_process.c @@ -2708,6 +2708,48 @@ void qemuProcessStop(struct qemud_driver if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, &net->data.direct.virtPortProfile); + + if (net->data.direct.mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + char *newmacname = NULL; + char *oldmacname = NULL; + char *oldmacpath = NULL; + char *basedir = "/var/run/libvirt/network"; + + if (virAsprintf(&newmacname, "%s/%s@%02x:%02x:%02x:%02x:%02x:%02x", + basedir, + net->data.direct.linkdev, + net->mac[0],net->mac[1],net->mac[2], + net->mac[3],net->mac[4],net->mac[5]) < 0) { + virReportOOMError(); + } + + ret = virFileResolveLink(newmacname, &oldmacpath); + if (ret < 0) { + virReportSystemError(errno, + _("Reading MAC address from '%s' failed !"), + newmacname); + } else { + unsigned char oldmac[6]; + + oldmacname = basename(oldmacpath); + + if (virParseMacAddr(oldmacname, &oldmac[0]) != 0) { + virReportSystemError(VIR_ERR_INTERNAL_ERROR, + _("Cannot parse MAC address from '%s'"), + oldmacname); + } + + VIR_WARN(("Resetting MAC address on '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x'."), + net->data.direct.linkdev, + oldmac[0],oldmac[1],oldmac[2], + oldmac[3],oldmac[4],oldmac[5]); + /*reset mac and remove link file-ignore results*/ + ret = setMacaddr(oldmac, net->data.direct.linkdev); + ret = unlink(oldmacpath); + ret = unlink(newmacname); + } + } VIR_FREE(net->ifname); } } Index: libvirt/src/util/macvtap.c =================================================================== --- libvirt.orig/src/util/macvtap.c +++ libvirt/src/util/macvtap.c @@ -87,6 +87,7 @@ # define LLDPAD_PID_FILE "/var/run/lldpad.pid" +#define MACADDRSIZE 6 enum virVirtualPortOp { ASSOCIATE = 0x1, @@ -191,6 +192,149 @@ err_exit: # if WITH_MACVTAP +/** + * getMacaddr: + * Get the MAC address of a network device + * + * @macaddress: Pointer where the MAC address will be stored + * @srcdev: The interface name of the NIC to get the MAC from + * + * Returns zero in case of success, + * negative value otherwise with error reported. + * + */ +int +getMacaddr(const unsigned char *macaddress, const char *srcdev ) +{ + int sockfd; + int io; + struct ifreq ifr; + + strcpy(ifr.ifr_name, srcdev); + + sockfd = socket(AF_INET, SOCK_STREAM, 0); + if(sockfd < 0){ + return -1; + } + + io = ioctl(sockfd, SIOCGIFHWADDR, (char *)&ifr); + if(io < 0){ + return -1; + } + + memcpy(macaddress, ifr.ifr_ifru.ifru_hwaddr.sa_data, MACADDRSIZE); + + return 0; +} + +/** + * setMacaddr: + * Set the MAC address of a network device + * + * @macaddress: MAC address to assign to the NIC + * @srcdev: The interface name of the NIC + * + * Returns zero in case of success, + * negative value otherwise with error reported. + * + */ +int +setMacaddr(const unsigned char *macaddress, const char *srcdev ) +{ + int rc = 0; + struct nlmsghdr *resp; + struct nlmsgerr *err; + struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC }; + int ifindex; + unsigned char *recvbuf = NULL; + unsigned int recvbuflen; + struct nl_msg *nl_msg; + + if (ifaceGetIndex(true, srcdev, &ifindex) != 0) + return -1; + + nl_msg = nlmsg_alloc_simple(RTM_SETLINK, NLM_F_REQUEST); + + if (!nl_msg) { + virReportOOMError(); + return -1; + } + + if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) + goto buffer_too_small; + + if (nla_put_u32(nl_msg, IFLA_LINK, ifindex) < 0) + goto buffer_too_small; + + if (nla_put(nl_msg, IFLA_ADDRESS, MACADDRSIZE, macaddress) < 0) + goto buffer_too_small; + + if (srcdev && + nla_put(nl_msg, IFLA_IFNAME, strlen(srcdev)+1, srcdev) < 0) + goto buffer_too_small; + + if (nlComm(nl_msg, &recvbuf, &recvbuflen, 0) < 0) { + rc = -1; + goto err_exit; + } + + if (recvbuflen < NLMSG_LENGTH(0) || recvbuf == NULL) + goto malformed_resp; + + resp = (struct nlmsghdr *)recvbuf; + + switch (resp->nlmsg_type) { + case NLMSG_ERROR: + err = (struct nlmsgerr *)NLMSG_DATA(resp); + if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) + goto malformed_resp; + + switch (err->error) { + + case 0: + break; + + case -EEXIST: + rc = -1; + break; + + default: + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("error setting device mac address")); + rc = -1; + } + break; + + case NLMSG_DONE: + break; + + default: + goto malformed_resp; + } + +err_exit: + nlmsg_free(nl_msg); + + VIR_FREE(recvbuf); + + return rc; + +malformed_resp: + nlmsg_free(nl_msg); + + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("malformed netlink response message")); + VIR_FREE(recvbuf); + return -1; + +buffer_too_small: + nlmsg_free(nl_msg); + + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("allocated netlink buffer is too small")); + return -1; +} + static int link_add(const char *type, const unsigned char *macaddress, int macaddrsize, Index: libvirt/src/util/macvtap.h =================================================================== --- libvirt.orig/src/util/macvtap.h +++ libvirt/src/util/macvtap.h @@ -75,6 +75,10 @@ enum virVMOperationType { # include "internal.h" +int getMacaddr(const unsigned char *macaddress, const char *srcdev ); + +int setMacaddr(const unsigned char *macaddress, const char *srcdev ); + int openMacvtapTap(const char *ifname, const unsigned char *macaddress, const char *linkdev, =================================================================== Best regards, Gerhard Stenzel ------------------------------------------------------------------------------------- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Martin Jetter Geschaeftsfuehrung: Dirk Wittkopp Sitz der Gesellschaft: Boeblingen Registergericht: Amtsgericht Stuttgart, HRB 243294

3 3

[libvirt] [PATCH] virsh: Fix compile error due to const mismatch in cmdVcpupin
by Matthias Bolte 20 Jun '11

20 Jun '11

Broken by commit c4a8ca71b10eb adding the reset option. Instead of altering a const string move the while loop into an else clause. --- tools/virsh.c | 110 ++++++++++++++++++++++++++++----------------------------- 1 files changed, 54 insertions(+), 56 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index e04c9fc..0e9efad 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -3000,7 +3000,7 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) int cpumaplen; int i, cpu, lastcpu, maxcpu; bool unuse = false; - char *cur; + const char *cur; int config = vshCommandOptBool(cmd, "config"); int live = vshCommandOptBool(cmd, "live"); int current = vshCommandOptBool(cmd, "current"); @@ -3061,72 +3061,70 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) cpumap = vshCalloc(ctl, 0, cpumaplen); /* Parse cpulist */ - cur = cpulist; - if (*cur == 0) { + if (*cpulist == '\0') { goto parse_error; - } else if (*cur == 'r') { + } else if (*cpulist == 'r') { for (cpu = 0; cpu < maxcpu; cpu++) VIR_USE_CPU(cpumap, cpu); - *cur = 0; - } - - while (*cur != 0) { - - /* the char '^' denotes exclusive */ - if (*cur == '^') { - cur++; - unuse = true; - } - - /* parse physical CPU number */ - if (!c_isdigit(*cur)) - goto parse_error; - cpu = virParseNumber(&cur); - if (cpu < 0) { - goto parse_error; - } - if (cpu >= maxcpu) { - vshError(ctl, _("Physical CPU %d doesn't exist."), cpu); - goto parse_error; - } - virSkipSpaces(&cur); - - if ((*cur == ',') || (*cur == 0)) { - if (unuse) { - VIR_UNUSE_CPU(cpumap, cpu); - } else { - VIR_USE_CPU(cpumap, cpu); + } else { + cur = cpulist; + while (*cur != '\0') { + /* the char '^' denotes exclusive */ + if (*cur == '^') { + cur++; + unuse = true; } - } else if (*cur == '-') { - /* the char '-' denotes range */ - if (unuse) { + + /* parse physical CPU number */ + if (!c_isdigit(*cur)) goto parse_error; - } - cur++; - virSkipSpaces(&cur); - /* parse the end of range */ - lastcpu = virParseNumber(&cur); - if (lastcpu < cpu) { + cpu = virParseNumber(&cur); + if (cpu < 0) { goto parse_error; } - if (lastcpu >= maxcpu) { - vshError(ctl, _("Physical CPU %d doesn't exist."), maxcpu); + if (cpu >= maxcpu) { + vshError(ctl, _("Physical CPU %d doesn't exist."), cpu); goto parse_error; } - for (i = cpu; i <= lastcpu; i++) { - VIR_USE_CPU(cpumap, i); - } virSkipSpaces(&cur); - } - if (*cur == ',') { - cur++; - virSkipSpaces(&cur); - unuse = false; - } else if (*cur == 0) { - break; - } else { - goto parse_error; + if ((*cur == ',') || (*cur == '\0')) { + if (unuse) { + VIR_UNUSE_CPU(cpumap, cpu); + } else { + VIR_USE_CPU(cpumap, cpu); + } + } else if (*cur == '-') { + /* the char '-' denotes range */ + if (unuse) { + goto parse_error; + } + cur++; + virSkipSpaces(&cur); + /* parse the end of range */ + lastcpu = virParseNumber(&cur); + if (lastcpu < cpu) { + goto parse_error; + } + if (lastcpu >= maxcpu) { + vshError(ctl, _("Physical CPU %d doesn't exist."), maxcpu); + goto parse_error; + } + for (i = cpu; i <= lastcpu; i++) { + VIR_USE_CPU(cpumap, i); + } + virSkipSpaces(&cur); + } + + if (*cur == ',') { + cur++; + virSkipSpaces(&cur); + unuse = false; + } else if (*cur == '\0') { + break; + } else { + goto parse_error; + } } } -- 1.7.0.4

1 2

[libvirt] [PATCH] Fix compile warnings in virsh vcpupin
by Daniel P. Berrange 20 Jun '11

20 Jun '11

The 'char *cur' variable was being assigned from a 'const char *' string, thus discarding constness. As well as causing a compile warning, it masked a piece of code which attempts to assign to the previously const string. * tools/virsh.c: Fix const-ness of 'cur' variable in vcpupin --- tools/virsh.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 3df0683..ed26b93 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -3000,7 +3000,7 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) int cpumaplen; int i, cpu, lastcpu, maxcpu; bool unuse = false; - char *cur; + const char *cur; int config = vshCommandOptBool(cmd, "config"); int live = vshCommandOptBool(cmd, "live"); int current = vshCommandOptBool(cmd, "current"); @@ -3067,7 +3067,7 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) } else if (*cur == 'r') { for (cpu = 0; cpu < maxcpu; cpu++) VIR_USE_CPU(cpumap, cpu); - *cur = 0; + cur = ""; } while (*cur != 0) { -- 1.7.4.4

2 1

[libvirt] nwfilter: limit VM traffic to specific MAC
by Shahar Havivi 20 Jun '11

20 Jun '11

Hi, I am trying to add custom filter to block VM traffic to other VMs by limiting the traffic only to the gateways MAC address. The filter XML: <filter name='rhev' chain='root'> <uuid>cd4e5890-ccc9-1b0f-303f-e7fe7123646d</uuid> <filterref filter='allow-dhcp'/> <rule action='drop' direction='out' priority='500'> <mac match='no' dstmacaddr='$MAC'/> </rule> </filter> The MAC is not the interface MAC address it's the gateways MAC that pass as a parameter (I use the gateway address hardcoded as well). The VM is getting DHCP ip but cannot get any traffic, I notice that when I edit (comment and uncomment) the drop rule, the filter is working fine, ie no traffic other then the gateway. 1. Am I doing something wrong? 1. What is the table name that libvirt use for ebtables? Shahar.

2 3

[libvirt] Availability of patchchecker
by Daniel Veillard 20 Jun '11

20 Jun '11

As some may have noticed, I'm often behind on patch review and it happens I notice patches way too late i.e. just before a release. For some time I though that having something automatic to remind me about unreviewed/unapplied patches would be a good way to fight this. I know that Eric, Matthias, Dan ... scans mostly everything but well are all getting quite busy and sometime things get forgotten and that's bad for the person who spent time on it and it's bad for the community. Asking to resend old patches tend to inflate the problem it's not a solution and humanly it's poor taste. So I spent most part of the week writing patchchecker, it's an automated tool, working out of the public mail archives and a git checkout, and was tested only with libvirt but should work also with other projects using MHonarc for mail archives and git (or could be made more flexible, yadda yadda ...) http://www.libvirt.org/git/?p=patchchecker.git git clone git://libvirt.org/patchchecker.git Check the README, there is an indexer to run regulary to extract from the mail archives, and a command to try to show intereting things, right now it reports only patches which didn't get any comments or acks. It detects patchset but doesn't fully use them yet. It's crude but it's automatic ! Next step is to so a bit of XML output and XSLT to export the output as a web page on libvirt.org , I will also need to add support for patch versioning i.e. being able to obsolete an older version with a newer mail (or patch set). I guess discussion and patches should be held and sent to this list, which already raises a limitation of this tool: it currently support only one git checkout, and if you send a patch for it it will complain if it doesn't get applied after being ACK'ed :-) BTW one of the results is that we should try to use the formal ACK keyword all the time (capital) when accepting and commiting a patch and I may occasionally send reply mail to threads with ACK or something like COMMITED if the algorithm of the checker get stuck on one patch. I really expect the amount of manual intervention to be limited to this I want the tool to help me, not give me/us more work :-) Hopefully it may prove useful to a few of us, and possibly be adapted to other projects (as usual I take patches :-) 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/

3 3

[libvirt] [PATCH] remote: Generate virDomainGetBlockPullInfo
by Matthias Bolte 20 Jun '11

20 Jun '11

It was already generatable but skipped. --- daemon/remote.c | 36 ------------------------------------ src/remote/remote_driver.c | 32 -------------------------------- src/remote/remote_protocol.x | 4 ++-- 3 files changed, 2 insertions(+), 70 deletions(-) diff --git a/daemon/remote.c b/daemon/remote.c index 1914b25..8e0f47d 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -1737,42 +1737,6 @@ cleanup: return rv; } -static int -remoteDispatchDomainGetBlockPullInfo(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client ATTRIBUTE_UNUSED, - virConnectPtr conn, - remote_message_header *hdr ATTRIBUTE_UNUSED, - remote_error * rerr, - remote_domain_get_block_pull_info_args *args, - remote_domain_get_block_pull_info_ret *ret) -{ - virDomainPtr dom = NULL; - virDomainBlockPullInfo tmp; - int rv = -1; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - if (!(dom = get_nonnull_domain(conn, args->dom))) - goto cleanup; - - if (virDomainGetBlockPullInfo(dom, args->path, &tmp, args->flags) < 0) - goto cleanup; - ret->cur = tmp.cur; - ret->end = tmp.end; - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - if (dom) - virDomainFree(dom); - return rv; -} - - /*-------------------------------------------------------------*/ static int diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index f4b43e0..5db1064 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -2671,38 +2671,6 @@ done: return rv; } -static int remoteDomainGetBlockPullInfo(virDomainPtr domain, - const char *path, - virDomainBlockPullInfoPtr info, - unsigned int flags) -{ - int rv = -1; - remote_domain_get_block_pull_info_args args; - remote_domain_get_block_pull_info_ret ret; - struct private_data *priv = domain->conn->privateData; - - remoteDriverLock(priv); - - make_nonnull_domain(&args.dom, domain); - args.path = (char *)path; - args.flags = flags; - - if (call(domain->conn, priv, 0, REMOTE_PROC_DOMAIN_GET_BLOCK_PULL_INFO, - (xdrproc_t)xdr_remote_domain_get_block_pull_info_args, - (char *)&args, - (xdrproc_t)xdr_remote_domain_get_block_pull_info_ret, - (char *)&ret) == -1) - goto done; - - info->cur = ret.cur; - info->end = ret.end; - rv = 0; - -done: - remoteDriverUnlock(priv); - return rv; -} - /*----------------------------------------------------------------------*/ static virDrvOpenStatus ATTRIBUTE_NONNULL (1) diff --git a/src/remote/remote_protocol.x b/src/remote/remote_protocol.x index 2b9784b..b136e41 100644 --- a/src/remote/remote_protocol.x +++ b/src/remote/remote_protocol.x @@ -1003,7 +1003,7 @@ struct remote_domain_get_block_pull_info_args { unsigned int flags; }; -struct remote_domain_get_block_pull_info_ret { +struct remote_domain_get_block_pull_info_ret { /* insert@2 */ unsigned hyper cur; unsigned hyper end; }; @@ -2409,7 +2409,7 @@ enum remote_procedure { REMOTE_PROC_DOMAIN_BLOCK_PULL_ALL = 230, /* autogen autogen */ REMOTE_PROC_DOMAIN_BLOCK_PULL_ABORT = 231, /* autogen autogen */ - REMOTE_PROC_DOMAIN_GET_BLOCK_PULL_INFO = 232, /* skipgen skipgen */ + REMOTE_PROC_DOMAIN_GET_BLOCK_PULL_INFO = 232, /* autogen autogen */ REMOTE_PROC_DOMAIN_EVENT_BLOCK_PULL = 233 /* skipgen skipgen */ /* -- 1.7.0.4

2 2

[libvirt] [PATCH] remote: Handle functions thar return optional strings
by Matthias Bolte 20 Jun '11

20 Jun '11

--- daemon/remote.c | 64 ------------------------------------------ daemon/remote_generator.pl | 28 ++++++++++++++++++ src/remote/remote_protocol.x | 2 +- 3 files changed, 29 insertions(+), 65 deletions(-) diff --git a/daemon/remote.c b/daemon/remote.c index 8e0f47d..f2e193d 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -1485,70 +1485,6 @@ cleanup: } static int -remoteDispatchDomainScreenshot(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client, - virConnectPtr conn, - remote_message_header *hdr, - remote_error *rerr, - remote_domain_screenshot_args *args, - remote_domain_screenshot_ret *ret) -{ - int rv = -1; - struct qemud_client_stream *stream = NULL; - virDomainPtr dom = NULL; - char *mime, **mime_p; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - ret->mime = NULL; - - if (!(dom = get_nonnull_domain (conn, args->dom))) - goto cleanup; - - if (!(stream = remoteCreateClientStream(conn, hdr))) - goto cleanup; - - if (!(mime = virDomainScreenshot(dom, stream->st, args->screen, args->flags))) - goto cleanup; - - if (remoteAddClientStream(client, stream, 1) < 0) { - virStreamAbort(stream->st); - goto cleanup; - } - - if (VIR_ALLOC(mime_p) < 0) { - virReportOOMError(); - goto cleanup; - } - - *mime_p = strdup(mime); - if (*mime_p == NULL) { - virReportOOMError(); - VIR_FREE(mime_p); - goto cleanup; - } - - ret->mime = mime_p; - - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - VIR_FREE(mime); - if (dom) - virDomainFree(dom); - if (stream && rv != 0) { - virStreamAbort(stream->st); - remoteFreeClientStream(client, stream); - } - return rv; -} - -static int remoteDispatchNodeGetCPUStats (struct qemud_server *server ATTRIBUTE_UNUSED, struct qemud_client *client ATTRIBUTE_UNUSED, virConnectPtr conn, diff --git a/daemon/remote_generator.pl b/daemon/remote_generator.pl index bcf5fd4..d9d9ded 100755 --- a/daemon/remote_generator.pl +++ b/daemon/remote_generator.pl @@ -351,6 +351,7 @@ elsif ($opt_b) { my @optionals_list = (); my @getters_list = (); my @args_list = (); + my @prepare_ret_list = (); my @ret_list = (); my @free_list = (); my @free_list_on_error = ("remoteDispatchError(rerr);"); @@ -551,6 +552,27 @@ elsif ($opt_b) { $single_ret_var = $1; $single_ret_by_ref = 0; $single_ret_check = " == NULL"; + } elsif ($ret_member =~ m/^remote_string (\S+);/) { + push(@vars_list, "char *$1 = NULL"); + push(@vars_list, "char **$1_p = NULL"); + push(@ret_list, "ret->$1 = $1_p;"); + push(@free_list, " VIR_FREE($1);"); + push(@free_list_on_error, "VIR_FREE($1_p);"); + push(@prepare_ret_list, + "if (VIR_ALLOC($1_p) < 0) {\n" . + " virReportOOMError();\n" . + " goto cleanup;\n" . + " }\n" . + " \n" . + " *$1_p = strdup($1);\n" . + " if (*$1_p == NULL) {\n" . + " virReportOOMError();\n" . + " goto cleanup;\n" . + " }\n"); + + $single_ret_var = $1; + $single_ret_by_ref = 0; + $single_ret_check = " == NULL"; } elsif ($ret_member =~ m/^remote_nonnull_(domain|network|storage_pool|storage_vol|interface|node_device|secret|nwfilter|domain_snapshot) (\S+);/) { my $type_name = name_to_ProcName($1); @@ -864,6 +886,12 @@ elsif ($opt_b) { print "\n"; } + if (@prepare_ret_list) { + print " "; + print join("\n ", @prepare_ret_list); + print "\n"; + } + if (@ret_list) { print " "; print join("\n ", @ret_list); diff --git a/src/remote/remote_protocol.x b/src/remote/remote_protocol.x index b136e41..0aa3dfb 100644 --- a/src/remote/remote_protocol.x +++ b/src/remote/remote_protocol.x @@ -2386,7 +2386,7 @@ enum remote_procedure { REMOTE_PROC_STORAGE_VOL_DOWNLOAD = 209, /* autogen autogen | readstream@1 */ REMOTE_PROC_DOMAIN_INJECT_NMI = 210, /* autogen autogen */ - REMOTE_PROC_DOMAIN_SCREENSHOT = 211, /* skipgen autogen | readstream@1 */ + REMOTE_PROC_DOMAIN_SCREENSHOT = 211, /* autogen autogen | readstream@1 */ REMOTE_PROC_DOMAIN_GET_STATE = 212, /* skipgen skipgen */ REMOTE_PROC_DOMAIN_MIGRATE_BEGIN3 = 213, /* skipgen skipgen */ REMOTE_PROC_DOMAIN_MIGRATE_PREPARE3 = 214, /* skipgen skipgen */ -- 1.7.0.4

2 2

[libvirt] [PATCH] remote generator: Allow to annotate arrays with typecasts
by Matthias Bolte 20 Jun '11

20 Jun '11

Removes special case code from the generator and handle additional methods. The generated version of remoteDispatchDomainPinVcpu(Flags) has no length check, but this check was useless anyway as it was applied to data that was already deserialized from its XDR form. --- daemon/remote.c | 82 -------------------------------------- daemon/remote_generator.pl | 90 ++++++++++++++++++++++++++++++------------ src/remote/remote_protocol.x | 17 +++++--- 3 files changed, 74 insertions(+), 115 deletions(-) diff --git a/daemon/remote.c b/daemon/remote.c index f2e193d..37fbed0 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -1269,88 +1269,6 @@ cleanup: } static int -remoteDispatchDomainPinVcpu(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client ATTRIBUTE_UNUSED, - virConnectPtr conn, - remote_message_header *hdr ATTRIBUTE_UNUSED, - remote_error *rerr, - remote_domain_pin_vcpu_args *args, - void *ret ATTRIBUTE_UNUSED) -{ - virDomainPtr dom = NULL; - int rv = -1; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - if (!(dom = get_nonnull_domain(conn, args->dom))) - goto cleanup; - - if (args->cpumap.cpumap_len > REMOTE_CPUMAP_MAX) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("cpumap_len > REMOTE_CPUMAP_MAX")); - goto cleanup; - } - - if (virDomainPinVcpu(dom, args->vcpu, - (unsigned char *) args->cpumap.cpumap_val, - args->cpumap.cpumap_len) < 0) - goto cleanup; - - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - if (dom) - virDomainFree(dom); - return rv; -} - -static int -remoteDispatchDomainPinVcpuFlags(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client ATTRIBUTE_UNUSED, - virConnectPtr conn, - remote_message_header *hdr ATTRIBUTE_UNUSED, - remote_error *rerr, - remote_domain_pin_vcpu_flags_args *args, - void *ret ATTRIBUTE_UNUSED) -{ - virDomainPtr dom = NULL; - int rv = -1; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - if (!(dom = get_nonnull_domain(conn, args->dom))) - goto cleanup; - - if (args->cpumap.cpumap_len > REMOTE_CPUMAP_MAX) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("cpumap_len > REMOTE_CPUMAP_MAX")); - goto cleanup; - } - - if (virDomainPinVcpuFlags(dom, - args->vcpu, - (unsigned char *) args->cpumap.cpumap_val, - args->cpumap.cpumap_len, - args->flags) < 0) - goto cleanup; - - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - if (dom) - virDomainFree(dom); - return rv; -} - -static int remoteDispatchDomainGetMemoryParameters(struct qemud_server *server ATTRIBUTE_UNUSED, struct qemud_client *client diff --git a/daemon/remote_generator.pl b/daemon/remote_generator.pl index d9d9ded..146c56b 100755 --- a/daemon/remote_generator.pl +++ b/daemon/remote_generator.pl @@ -403,20 +403,32 @@ elsif ($opt_b) { " virDomainSnapshotFree(snapshot);\n" . " if (dom)\n" . " virDomainFree(dom);"); - } elsif ($args_member =~ m/^(remote_string|remote_nonnull_string|remote_uuid|opaque) (\S+)<\S+>;/) { + } elsif ($args_member =~ m/^(?:remote_string|remote_uuid) (\S+)<\S+>;/) { if (! @args_list) { push(@args_list, "conn"); } - if ($call->{ProcName} eq "SecretSetValue") { - push(@args_list, "(const unsigned char *)args->$2.$2_val"); - } elsif ($call->{ProcName} eq "CPUBaseline") { - push(@args_list, "(const char **)args->$2.$2_val"); - } else { - push(@args_list, "args->$2.$2_val"); + push(@args_list, "args->$1.$1_val"); + push(@args_list, "args->$1.$1_len"); + } elsif ($args_member =~ m/^(?:opaque|remote_nonnull_string) (\S+)<\S+>;(.*)$/) { + if (! @args_list) { + push(@args_list, "conn"); + } + + my $cast = ""; + my $arg_name = $1; + my $annotation = $2; + + if ($annotation ne "") { + if ($annotation =~ m/\s*\/\*\s*(.*)\s*\*\//) { + $cast = $1; + } else { + die "malformed cast annotation for argument: $args_member"; + } } - push(@args_list, "args->$2.$2_len"); + push(@args_list, "${cast}args->$arg_name.${arg_name}_val"); + push(@args_list, "args->$arg_name.${arg_name}_len"); } elsif ($args_member =~ m/^(?:unsigned )?int (\S+)<\S+>;/) { if (! @args_list) { push(@args_list, "conn"); @@ -998,37 +1010,63 @@ elsif ($opt_k) { } elsif ($args_member =~ m/^remote_string (\S+);/) { push(@args_list, "const char *$1"); push(@setters_list, "args.$1 = $1 ? (char **)&$1 : NULL;"); - } elsif ($args_member =~ m/^remote_nonnull_string (\S+)<(\S+)>;/) { - push(@args_list, "const char **$1"); - push(@args_list, "unsigned int ${1}len"); - push(@setters_list, "args.$1.${1}_val = (char **)$1;"); - push(@setters_list, "args.$1.${1}_len = ${1}len;"); - push(@args_check_list, { name => "\"$1\"", arg => "${1}len", limit => $2 }); + } elsif ($args_member =~ m/^remote_nonnull_string (\S+)<(\S+)>;(.*)$/) { + my $type_name = "const char **"; + my $arg_name = $1; + my $limit = $2; + my $annotation = $3; + + if ($annotation ne "") { + if ($annotation =~ m/\s*\/\*\s*$(.*)$\s*\*\//) { + $type_name = $1; + } else { + die "malformed cast annotation for argument: $args_member"; + } + } + + push(@args_list, "$type_name$arg_name"); + push(@args_list, "unsigned int ${arg_name}len"); + push(@setters_list, "args.$arg_name.${arg_name}_val = (char **)$arg_name;"); + push(@setters_list, "args.$arg_name.${arg_name}_len = ${arg_name}len;"); + push(@args_check_list, { name => "\"$arg_name\"", arg => "${arg_name}len", limit => $2 }); } elsif ($args_member =~ m/^remote_nonnull_string (\S+);/) { push(@args_list, "const char *$1"); push(@setters_list, "args.$1 = (char *)$1;"); - } elsif ($args_member =~ m/^(remote_string|opaque) (\S+)<(\S+)>;/) { - my $type_name = $1; - my $arg_name = $2; - my $limit = $3; + } elsif ($args_member =~ m/^opaque (\S+)<(\S+)>;(.*)$/) { + my $type_name = "const char *"; + my $arg_name = $1; + my $limit = $2; + my $annotation = $3; + + if ($annotation ne "") { + if ($annotation =~ m/\s*\/\*\s*$(.*)$\s*\*\//) { + $type_name = $1; + } else { + die "malformed cast annotation for argument: $args_member"; + } + } + + push(@args_list, "$type_name$arg_name"); if ($call->{ProcName} eq "SecretSetValue") { - push(@args_list, "const unsigned char *$arg_name"); + # SPECIAL: virSecretSetValue uses size_t instead of int push(@args_list, "size_t ${arg_name}len"); - } elsif ($call->{ProcName} eq "DomainPinVcpu") { - push(@args_list, "unsigned char *$arg_name"); - push(@args_list, "int ${arg_name}len"); - } elsif ($call->{ProcName} eq "DomainPinVcpuFlags") { - push(@args_list, "unsigned char *$arg_name"); - push(@args_list, "int ${arg_name}len"); } else { - push(@args_list, "const char *$arg_name"); push(@args_list, "int ${arg_name}len"); } push(@setters_list, "args.$arg_name.${arg_name}_val = (char *)$arg_name;"); push(@setters_list, "args.$arg_name.${arg_name}_len = ${arg_name}len;"); push(@args_check_list, { name => "\"$arg_name\"", arg => "${arg_name}len", limit => $limit }); + } elsif ($args_member =~ m/^remote_string (\S+)<(\S+)>;/) { + my $arg_name = $1; + my $limit = $2; + + push(@args_list, "const char *$arg_name"); + push(@args_list, "int ${arg_name}len"); + push(@setters_list, "args.$arg_name.${arg_name}_val = (char *)$arg_name;"); + push(@setters_list, "args.$arg_name.${arg_name}_len = ${arg_name}len;"); + push(@args_check_list, { name => "\"$arg_name\"", arg => "${arg_name}len", limit => $limit }); } elsif ($args_member =~ m/^((?:unsigned )?int) (\S+)<(\S+)>;/) { my $type_name = $1; my $arg_name = $2; diff --git a/src/remote/remote_protocol.x b/src/remote/remote_protocol.x index 0aa3dfb..a77fd81 100644 --- a/src/remote/remote_protocol.x +++ b/src/remote/remote_protocol.x @@ -357,7 +357,10 @@ struct remote_node_get_memory_stats { * If the 'remote_CALL_ret' maps to a struct in the public API then it is * also filled via call-by-reference and must be annotated with a * insert@<offset> comment to indicate the offset in the parameter list of - * the function to be called. */ + * the function to be called. + * + * Dynamic opaque and remote_nonnull_string arrays can be annotated with an + * optional typecast */ struct remote_open_args { /* NB. "name" might be NULL although in practice you can't @@ -890,13 +893,13 @@ struct remote_domain_get_vcpus_flags_ret { struct remote_domain_pin_vcpu_args { remote_nonnull_domain dom; unsigned int vcpu; - opaque cpumap<REMOTE_CPUMAP_MAX>; + opaque cpumap<REMOTE_CPUMAP_MAX>; /* (unsigned char *) */ }; struct remote_domain_pin_vcpu_flags_args { remote_nonnull_domain dom; unsigned int vcpu; - opaque cpumap<REMOTE_CPUMAP_MAX>; + opaque cpumap<REMOTE_CPUMAP_MAX>; /* (unsigned char *) */ unsigned int flags; }; @@ -1701,7 +1704,7 @@ struct remote_secret_get_xml_desc_ret { struct remote_secret_set_value_args { remote_nonnull_secret secret; - opaque value<REMOTE_SECRET_VALUE_MAX>; + opaque value<REMOTE_SECRET_VALUE_MAX>; /* (const unsigned char *) */ unsigned int flags; }; @@ -1819,7 +1822,7 @@ struct remote_cpu_compare_ret { struct remote_cpu_baseline_args { - remote_nonnull_string xmlCPUs<REMOTE_CPU_BASELINE_MAX>; + remote_nonnull_string xmlCPUs<REMOTE_CPU_BASELINE_MAX>; /* (const char **) */ unsigned int flags; }; @@ -2182,7 +2185,7 @@ enum remote_procedure { REMOTE_PROC_DOMAIN_LOOKUP_BY_NAME = 23, /* autogen autogen */ REMOTE_PROC_DOMAIN_LOOKUP_BY_UUID = 24, /* autogen autogen */ REMOTE_PROC_NUM_OF_DEFINED_DOMAINS = 25, /* autogen autogen */ - REMOTE_PROC_DOMAIN_PIN_VCPU = 26, /* skipgen autogen */ + REMOTE_PROC_DOMAIN_PIN_VCPU = 26, /* autogen autogen */ REMOTE_PROC_DOMAIN_REBOOT = 27, /* autogen autogen */ REMOTE_PROC_DOMAIN_RESUME = 28, /* autogen autogen */ REMOTE_PROC_DOMAIN_SET_AUTOSTART = 29, /* autogen autogen */ @@ -2401,7 +2404,7 @@ enum remote_procedure { REMOTE_PROC_INTERFACE_CHANGE_ROLLBACK = 222, /* autogen autogen */ REMOTE_PROC_DOMAIN_GET_SCHEDULER_PARAMETERS_FLAGS = 223, /* skipgen autogen */ REMOTE_PROC_DOMAIN_EVENT_CONTROL_ERROR = 224, /* skipgen skipgen */ - REMOTE_PROC_DOMAIN_PIN_VCPU_FLAGS = 225, /* skipgen autogen */ + REMOTE_PROC_DOMAIN_PIN_VCPU_FLAGS = 225, /* autogen autogen */ REMOTE_PROC_DOMAIN_SEND_KEY = 226, /* autogen autogen */ REMOTE_PROC_NODE_GET_CPU_STATS = 227, /* skipgen skipgen */ REMOTE_PROC_NODE_GET_MEMORY_STATS = 228, /* skipgen skipgen */ -- 1.7.0.4

2 2

[libvirt] [PATCH] Fix preprocessor indentation in nodeinfo.c
by Matthias Bolte 20 Jun '11

20 Jun '11

--- I'm pushing this under the build breaker rule. src/nodeinfo.c | 12 ++++++------ 1 files changed, 6 insertions(+), 6 deletions(-) diff --git a/src/nodeinfo.c b/src/nodeinfo.c index c6dbb84..9e0f906 100644 --- a/src/nodeinfo.c +++ b/src/nodeinfo.c @@ -685,23 +685,23 @@ int nodeGetMemoryStats(virConnectPtr conn ATTRIBUTE_UNUSED, return -1; } } else { -#if HAVE_NUMACTL +# if HAVE_NUMACTL if (numa_available() < 0) { -#endif +# endif nodeReportError(VIR_ERR_NO_SUPPORT, "%s", _("NUMA not supported on this host")); return -1; -#if HAVE_NUMACTL +# if HAVE_NUMACTL } -#endif +# endif -#if HAVE_NUMACTL +# if HAVE_NUMACTL if (cellNum > numa_max_node()) { nodeReportError(VIR_ERR_INVALID_ARG, "%s", _("Invalid cell number")); return -1; } -#endif +# endif if (virAsprintf(&meminfo_path, "%s/node%d/meminfo", NODE_SYS_PATH, cellNum) < 0) { -- 1.7.0.4

2 2

Re: [libvirt] [libvirt-php] - zend_mm_heap corrupted
by Michal Novotny 20 Jun '11

20 Jun '11

Hi Henrik, what exactly did you to do to get the error page? I never run into those in any version I'm having in libvirt-php git repository since I always check but it may be Apache & PHP version connected and I'm running it as a module. Could you please provide me more information about this or at least provide me the results of `make check` command in the libvirt-php directory as cloned from git? Please note that to try to `make check` you need to install it first so you need to do `make`, then `(sudo) make install` and then `make check`. The `make check` command is running the tests and you shouldn't be seeing any segmentation fault there otherwise there's some bug in the code which obviously seems to be reproducible only on your environment since it's not reproducible on my environment at all. All the tests pass in my case: $ make check ... Making check in tests make[1]: Entering directory `tests' ./runtests.sh Test test-connect.phpt has been completed successfully Test test-version-check.phpt has been completed successfully Test test-version-get.phpt has been completed successfully Test test-domain-define-undefine.phpt has been completed successfully Test test-domain-define-create-destroy.phpt has been completed successfully Test test-domain-create.phpt has been completed successfully Test test-domain-create-and-get-xpath.phpt has been completed successfully Test test-domain-create-and-coredump.phpt has been completed successfully Test test-domain-snapshot.phpt has been completed successfully All tests passed successfully ... $ Could you please provide me with the Apache logs as well and also the Apache and PHP version numbers? You mention Drupal, did you integrate libvirt-php functionality into Drupal? The backtraces using xdebug should be sufficient. Please provide me with them. Thanks, Michal On 06/18/2011 06:51 PM, Henrik Jönsson wrote: > > I basicly just got an erronous page load and checked the apache logs. > After some further testing it seems the problem only occurs when ran > within Drupal and not when ran as a standalone script. I will do some > further testing in a different cms to see if it is Drupal specific. I > can also do backtraces with xdebug and coredump apache if that will help. > > Best Regards > > /Henrik > Den 15 jun 2011 09:03 skrev "Michal Novotny" <minovotn(a)redhat.com > <mailto:minovotn@redhat.com>>: > > > > Hi Henrik, > > thanks a lot for your observations. How did you find out about the PHP > > memory usage and about zend_mm_heap corrupted at the first place? > > > > This could help me debug this issue. > > > > Thanks, > > Michal > > > > On 06/12/2011 06:05 PM, Henrik Jönsson wrote: > > > [libvirt-php] > > > > > > Hello! > > > > > > I am using Libvirt-php for a private project, i do however have a > > > problem that seems to occur when i call any function that returns a > > > domain resource. I have gotten 3 different errors so far. First i had > > > a problem with php memory usage going extreme(2gb), then i have had > > > some apache segmentation fault ending up in apache killing the > > > process. I have also had zend_mm_heap corrupted which i think is the > > > most common one. > > > > > > I have tried with Ubuntu 10.04, 10.10 and 11.04 repository versions of > > > Apache. Other than that i am using libvirt version 0.9.1 and > > > libvirt-php version 0.4.1. This is within the contexts of Drupal as i > > > am using the code in a Drupal module. > > > > > > I have checked out the source code for libvirt-php but since i am no > > > C/C++ developer i have a hard time figuring out what the issue might > > > be. Some help would be greatly appreciated. :) > > > > > > Best Regards! > > > > > > /Henrik > > > > > > > > > -- > > > libvir-list mailing list > > > libvir-list(a)redhat.com <mailto:libvir-list@redhat.com> > > > https://www.redhat.com/mailman/listinfo/libvir-list > > > > > > -- > > Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>, RHCE > > Virtualization Team (xen userspace), Red Hat > > > -- Michal Novotny <minovotn(a)redhat.com>, RHCE Virtualization Team (xen userspace), Red Hat

1 0

[libvirt] [PATCH] Move XenAPI driver to correct spec file section
by Matthias Bolte 20 Jun '11

20 Jun '11

The XenAPI driver works like the ESX and PHyp driver by using its own HTTPS based remote protocol. --- libvirt.spec.in | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 97ebd65..75b145a 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -44,13 +44,13 @@ %define with_lxc 0%{!?_without_lxc:%{server_drivers}} %define with_vbox 0%{!?_without_vbox:%{server_drivers}} %define with_uml 0%{!?_without_uml:%{server_drivers}} -%define with_xenapi 0%{!?_without_xenapi:%{server_drivers}} %define with_libxl 0%{!?_without_libxl:%{server_drivers}} %define with_vmware 0%{!?_without_vmware:%{server_drivers}} # Then the hypervisor drivers that talk a native remote protocol %define with_phyp 0%{!?_without_phyp:1} %define with_esx 0%{!?_without_esx:1} +%define with_xenapi 0%{!?_without_xenapi:1} # Then the secondary host drivers %define with_network 0%{!?_without_network:%{server_drivers}} -- 1.7.0.4

2 2

[libvirt] [PATCH v4 RESEND 0/3] Add filesystem pool formatting
by Osier Yang 20 Jun '11

20 Jun '11

Hi, The following patches add the ability to format filesystem pools when the appropriate flags are passed to pool build. This patch set introduces two new flags: VIR_STORAGE_POOL_BUILD_NO_OVERWRITE causes the build to probe for an existing pool of the requested type. The build operation formats the filesystem if it does not find an existing filesystem of that type. VIR_STORAGE_POOL_BUILD_OVERWRITE causes the build to format unconditionally. [PATCH 1/3] storage: Add mkfs and libblkid to build system [PATCH 2/3] storage: Add fs pool formatting [PATCH 3/3] storage: Add virsh support for fs pool formating Regards Osier

1 4

[libvirt] [PATCH] build: fix building error when building without libvirtd
by Wen Congyang 20 Jun '11

20 Jun '11

When building libvirt without libvirtd, we will receive the following error message: make[3]: Entering directory `/home/wency/rpmbuild/BUILD/libvirt-0.9.2/tools' CC virsh-virsh.o CC virsh-console.o GEN virt-xml-validate GEN virt-pki-validate CCLD virsh ../src/.libs/libvirt.so: undefined reference to `numa_available' ../src/.libs/libvirt.so: undefined reference to `numa_max_node' collect2: ld returned 1 exit status The reason is that: we check numactl only when building qemu driver, and qemu driver will not be built when bulding without libvirtd. So with_numactl's value is check and we will not link libnuma.so. In the other function, we call numa_available() and numa_max_node() only when HAVE_NUMACTL is 1. We should do the same check in the function nodeGetMemoryStats(). --- src/nodeinfo.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/src/nodeinfo.c b/src/nodeinfo.c index 647cb1e..c6dbb84 100644 --- a/src/nodeinfo.c +++ b/src/nodeinfo.c @@ -685,17 +685,23 @@ int nodeGetMemoryStats(virConnectPtr conn ATTRIBUTE_UNUSED, return -1; } } else { +#if HAVE_NUMACTL if (numa_available() < 0) { +#endif nodeReportError(VIR_ERR_NO_SUPPORT, "%s", _("NUMA not supported on this host")); return -1; +#if HAVE_NUMACTL } +#endif +#if HAVE_NUMACTL if (cellNum > numa_max_node()) { nodeReportError(VIR_ERR_INVALID_ARG, "%s", _("Invalid cell number")); return -1; } +#endif if (virAsprintf(&meminfo_path, "%s/node%d/meminfo", NODE_SYS_PATH, cellNum) < 0) { -- 1.7.1

2 2

[libvirt] [PATCH v3 0/3] add support for changing blkio parameters for inactive domains
by Hu Tao 20 Jun '11

20 Jun '11

This series enables user to change blkio parameters for inactive domains from virsh command line. CHANGES: v2-v3: - based on new macros VIR_DOMAIN_AFFECT_XXX Hu Tao (3): Add new parameters for blkiotune update qemuDomainGetBlkioParameters to use flags Update qemuDomainSetBlkioParameters to use flags src/qemu/qemu_driver.c | 250 ++++++++++++++++++++++++++++++++++-------------- tools/virsh.c | 26 +++++- tools/virsh.pod | 7 ++ 3 files changed, 206 insertions(+), 77 deletions(-) -- 1.7.3.1

3 8

[libvirt] Schedule for 0.9.3 release
by Daniel Veillard 20 Jun '11

20 Jun '11

As stated before I really want the next release schedule to be aligned with month end, so I would like to get 0.9.3 out around Friday 1st July, which means that there is one week left before we enter the feature freeze. Make sure your patches get reviewed in time this week if they need to make 0.9.3 :-) 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/

1 0

[libvirt] [PATCH] apparmor: implement AppArmorSetFDLabel()
by Jamie Strandboge 20 Jun '11

20 Jun '11

During a savevm operation, libvirt will now use fd migration if qemu supports it. When the AppArmor driver is enabled, AppArmorSetFDLabel() is used but since this function simply returns '0', the dynamic AppArmor profile is not updated and AppArmor blocks access to the save file. This patch implements AppArmorSetFDLabel() to get the pathname of the file by resolving the fd symlink in /proc, and then gives that pathname to reload_profile(), which fixes 'virsh save' when AppArmor is enabled. Passes 'check' and 'syntax-check' (though po_check failed for unrelated reasons). Reference: https://launchpad.net/bugs/795800 -- Jamie Strandboge | http://www.canonical.com

3 2

[libvirt] [PATCH v3]: set and restore MAC address of a NIC when using PASSTHROUGH mode
by Gerhard Stenzel 19 Jun '11

19 Jun '11

This is a another rework of the patch Dirk sent out last week taking into account most propsosed changes The following patch addresses the problem that when a PASSTHROUGH mode DIRECT NIC connection is made the MAC address of the NIC is not automatically set and reset to the configured VM MAC and back again. The attached patch fixes this problem by setting and resetting the MAC while remembering the previous setting while the VM is running. This also works if libvirtd is restarted while the VM is running. the patch passes make syntax-check Signed-off-by: Dirk Herrendoerfer <d.herrendoerfer at herrendoerfer.name> Signed-off-by: Gerhard Stenzel <gerhard.stenzel(a)de.ibm.com> --- Index: libvirt/src/libvirt_macvtap.syms =================================================================== --- libvirt.orig/src/libvirt_macvtap.syms +++ libvirt/src/libvirt_macvtap.syms @@ -5,6 +5,8 @@ # macvtap.h delMacvtap; +getMacaddr; openMacvtapTap; +setMacaddr; vpAssociatePortProfileId; vpDisassociatePortProfileId; Index: libvirt/src/qemu/qemu_command.c =================================================================== --- libvirt.orig/src/qemu/qemu_command.c +++ libvirt/src/qemu/qemu_command.c @@ -128,7 +128,7 @@ qemuPhysIfaceConnect(virDomainDefPtr def rc = openMacvtapTap(net->ifname, net->mac, net->data.direct.linkdev, net->data.direct.mode, vnet_hdr, def->uuid, &net->data.direct.virtPortProfile, &res_ifname, - vmop); + vmop, driver->stateDir); if (rc >= 0) { qemuAuditNetDevice(def, net, res_ifname, true); VIR_FREE(net->ifname); @@ -149,7 +149,7 @@ qemuPhysIfaceConnect(virDomainDefPtr def if (err) { VIR_FORCE_CLOSE(rc); delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, - &net->data.direct.virtPortProfile); + net->data.direct.mode, &net->data.direct.virtPortProfile, driver->stateDir); VIR_FREE(net->ifname); } } Index: libvirt/src/qemu/qemu_process.c =================================================================== --- libvirt.orig/src/qemu/qemu_process.c +++ libvirt/src/qemu/qemu_process.c @@ -2707,7 +2707,8 @@ void qemuProcessStop(struct qemud_driver virDomainNetDefPtr net = def->nets[i]; if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, - &net->data.direct.virtPortProfile); + net->data.direct.mode, + &net->data.direct.virtPortProfile, driver->stateDir); VIR_FREE(net->ifname); } } Index: libvirt/src/util/macvtap.c =================================================================== --- libvirt.orig/src/util/macvtap.c +++ libvirt/src/util/macvtap.c @@ -87,6 +87,7 @@ # define LLDPAD_PID_FILE "/var/run/lldpad.pid" +#define MACADDRSIZE 6 enum virVirtualPortOp { ASSOCIATE = 0x1, @@ -191,6 +192,149 @@ err_exit: # if WITH_MACVTAP +/** + * getMacaddr: + * Get the MAC address of a network device + * + * @macaddress: Pointer where the MAC address will be stored + * @srcdev: The interface name of the NIC to get the MAC from + * + * Returns zero in case of success, + * negative value otherwise with error reported. + * + */ +int +getMacaddr(const unsigned char *macaddress, const char *srcdev ) +{ + int sockfd; + int io; + struct ifreq ifr; + + strcpy(ifr.ifr_name, srcdev); + + sockfd = socket(AF_INET, SOCK_STREAM, 0); + if(sockfd < 0){ + return -1; + } + + io = ioctl(sockfd, SIOCGIFHWADDR, (char *)&ifr); + if(io < 0){ + return -1; + } + + memcpy(macaddress, ifr.ifr_ifru.ifru_hwaddr.sa_data, MACADDRSIZE); + + return 0; +} + +/** + * setMacaddr: + * Set the MAC address of a network device + * + * @macaddress: MAC address to assign to the NIC + * @srcdev: The interface name of the NIC + * + * Returns zero in case of success, + * negative value otherwise with error reported. + * + */ +int +setMacaddr(const unsigned char *macaddress, const char *srcdev ) +{ + int rc = 0; + struct nlmsghdr *resp; + struct nlmsgerr *err; + struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC }; + int ifindex; + unsigned char *recvbuf = NULL; + unsigned int recvbuflen; + struct nl_msg *nl_msg; + + if (ifaceGetIndex(true, srcdev, &ifindex) != 0) + return -1; + + nl_msg = nlmsg_alloc_simple(RTM_SETLINK, NLM_F_REQUEST); + + if (!nl_msg) { + virReportOOMError(); + return -1; + } + + if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) + goto buffer_too_small; + + if (nla_put_u32(nl_msg, IFLA_LINK, ifindex) < 0) + goto buffer_too_small; + + if (nla_put(nl_msg, IFLA_ADDRESS, MACADDRSIZE, macaddress) < 0) + goto buffer_too_small; + + if (srcdev && + nla_put(nl_msg, IFLA_IFNAME, strlen(srcdev)+1, srcdev) < 0) + goto buffer_too_small; + + if (nlComm(nl_msg, &recvbuf, &recvbuflen, 0) < 0) { + rc = -1; + goto err_exit; + } + + if (recvbuflen < NLMSG_LENGTH(0) || recvbuf == NULL) + goto malformed_resp; + + resp = (struct nlmsghdr *)recvbuf; + + switch (resp->nlmsg_type) { + case NLMSG_ERROR: + err = (struct nlmsgerr *)NLMSG_DATA(resp); + if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) + goto malformed_resp; + + switch (err->error) { + + case 0: + break; + + case -EEXIST: + rc = -1; + break; + + default: + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("error setting device mac address")); + rc = -1; + } + break; + + case NLMSG_DONE: + break; + + default: + goto malformed_resp; + } + +err_exit: + nlmsg_free(nl_msg); + + VIR_FREE(recvbuf); + + return rc; + +malformed_resp: + nlmsg_free(nl_msg); + + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("malformed netlink response message")); + VIR_FREE(recvbuf); + return -1; + +buffer_too_small: + nlmsg_free(nl_msg); + + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("allocated netlink buffer is too small")); + return -1; +} + static int link_add(const char *type, const unsigned char *macaddress, int macaddrsize, @@ -575,7 +719,8 @@ openMacvtapTap(const char *tgifname, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, char **res_ifname, - enum virVMOperationType vmOp) + enum virVMOperationType vmOp, + char *stateDir) { const char *type = "macvtap"; int c, rc; @@ -589,6 +734,54 @@ openMacvtapTap(const char *tgifname, VIR_DEBUG("%s: VM OPERATION: %s", __FUNCTION__, virVMOperationTypeToString(vmOp)); + /** Note: When using PASSTHROUGH mode with MACVTAP devices the link + * device's MAC address must be set to the VMs MAC address. In + * order to not confuse the first switch or bridge in line this MAC + * address must be reset when the VM is shut down. + * This is especially important when using SRIOV capable cards that + * emulate their switch in firmware. + */ + if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + unsigned char oldmac[6]; + + rc = getMacaddr(&oldmac, linkdev); + if (rc) { + virReportSystemError(rc, + _("Getting MAC address from '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]); + } else { + char *path = NULL; + + char macstr[VIR_MAC_STRING_BUFLEN]; + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + return errno; + } + virFormatMacAddr(oldmac, macstr); + if (virFileWriteStr(path, macstr, O_CREAT|O_TRUNC|O_WRONLY) < 0) { + virReportSystemError(errno, _("Unable to preserve mac for %s"), + linkdev); + return errno; + } + } + + rc = setMacaddr(macaddress, linkdev); + if (rc) { + virReportSystemError(errno, + _("Setting MAC address on '%s' to " + "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), + linkdev, + macaddress[0], macaddress[1], macaddress[2], + macaddress[3], macaddress[4], macaddress[5]); + } + } + + if (tgifname) { if(ifaceGetIndex(false, tgifname, &ifindex) == 0) { if (STRPREFIX(tgifname, @@ -684,8 +886,43 @@ void delMacvtap(const char *ifname, const unsigned char *macaddr, const char *linkdev, - virVirtualPortProfileParamsPtr virtPortProfile) -{ + int mode, + virVirtualPortProfileParamsPtr virtPortProfile, + char *stateDir) +{ + if (mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + int ret; + char *oldmacname = NULL; + char *macstr = NULL; + char *path = NULL; + unsigned char oldmac[6]; + + if (virAsprintf(&path, "%s/%s", + stateDir, + linkdev) < 0) { + virReportOOMError(); + goto disassociate_exit; + } + + if (virFileReadAll(path, VIR_MAC_STRING_BUFLEN, &macstr) < 0) { + virReportSystemError(errno, _("Unable to preserve mac for %s"), + linkdev); + goto disassociate_exit; + } + + if (virParseMacAddr(macstr, &oldmac[0]) != 0) { + virReportSystemError(VIR_ERR_INTERNAL_ERROR, + _("Cannot parse MAC address from '%s'"), + oldmacname); + } + + /*reset mac and remove file-ignore results*/ + ret = setMacaddr(oldmac, linkdev); + ret = unlink(path); + VIR_FREE(macstr); + } + +disassociate_exit: if (ifname) { vpDisassociatePortProfileId(ifname, macaddr, linkdev, Index: libvirt/src/util/macvtap.h =================================================================== --- libvirt.orig/src/util/macvtap.h +++ libvirt/src/util/macvtap.h @@ -75,6 +75,10 @@ enum virVMOperationType { # include "internal.h" +int getMacaddr(const unsigned char *macaddress, const char *srcdev ); + +int setMacaddr(const unsigned char *macaddress, const char *srcdev ); + int openMacvtapTap(const char *ifname, const unsigned char *macaddress, const char *linkdev, @@ -83,12 +87,15 @@ int openMacvtapTap(const char *ifname, const unsigned char *vmuuid, virVirtualPortProfileParamsPtr virtPortProfile, char **res_ifname, - enum virVMOperationType vmop); + enum virVMOperationType vmop, + char *stateDir); void delMacvtap(const char *ifname, const unsigned char *macaddress, const char *linkdev, - virVirtualPortProfileParamsPtr virtPortProfile); + int mode, + virVirtualPortProfileParamsPtr virtPortProfile, + char *stateDir); int vpAssociatePortProfileId(const char *macvtap_ifname, const unsigned char *macvtap_macaddr, Index: libvirt/src/qemu/qemu_hotplug.c =================================================================== --- libvirt.orig/src/qemu/qemu_hotplug.c +++ libvirt/src/qemu/qemu_hotplug.c @@ -1610,7 +1610,7 @@ int qemuDomainDetachNetDevice(struct qem #if WITH_MACVTAP if (detach->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(detach->ifname, detach->mac, detach->data.direct.linkdev, - &detach->data.direct.virtPortProfile); + detach->data.direct.mode, &detach->data.direct.virtPortProfile, driver->stateDir); VIR_FREE(detach->ifname); } #endif =================================================================== Best regards, Gerhard Stenzel ------------------------------------------------------------------------------------- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Martin Jetter Geschaeftsfuehrung: Dirk Wittkopp Sitz der Gesellschaft: Boeblingen Registergericht: Amtsgericht Stuttgart, HRB 243294

2 1

[libvirt] [Libvirt-test-API][PATCH] Add a pair of keywords in parser.py for supportting testcases loop
by Nan Zhang 19 Jun '11

19 Jun '11

For example: ----------------- domain:start guestname vm1 domain:save start_loop guestname vm1 domain:restore end_loop 10 guestname vm1 domain:destroy guestname vm1 ----------------- --- parser.py | 46 +++++++++++++++++++++++++++++++++++++++++++++- 1 files changed, 45 insertions(+), 1 deletions(-) diff --git a/parser.py b/parser.py index 5b3ce58..e9cce58 100644 --- a/parser.py +++ b/parser.py @@ -31,13 +31,18 @@ class CaseFileParser(object): """ Parser the case configuration file to generate a data list. """ def __init__(self, casefile=None, debug=False): - """ Initialize the list and optionally parse case file. """ self.list = [[]] self.variables = {} self.missing_variables = [] self.debug = debug self.casefile = casefile self.env = env_parser.Envparser("env.cfg") + self.loop_finish = False + self.loop_start = 0 + self.loop_end = 0 + self.loop_times = 0 + self.loop_list = [] + if casefile: self.parse_file(casefile) @@ -306,6 +311,7 @@ class CaseFileParser(object): def parse(self, fh, list): """ For the testcase name parsing. """ + while True: if self.debug: self.debug_print("the list is", list) @@ -330,6 +336,44 @@ class CaseFileParser(object): self.debug_print("we begin to handle the case", tripped_casename) + if self.loop_finish: + for i in range(len(list)): + self.loop_list.append([]) + + i = 0 + for caselist in list: + for j in range(self.loop_start, self.loop_end): + self.loop_list[i].append(caselist.pop()) + + self.loop_list[i].reverse() + self.debug_print("loop_list is", self.loop_list) + caselist.extend(self.loop_list[i] * self.loop_times) + i += 1 + + self.loop_finish = False + + if len(tripped_caselist) == 2 and \ + tripped_caselist[1] == "start_loop": + for caselist in list: + newdict = {} + newdict[tripped_casename] = {} + caselist.append(newdict) + self.loop_start = len(caselist) - 1 + continue + + if len(tripped_caselist) == 3 and \ + tripped_caselist[1] == "end_loop": + looptimes = tripped_caselist[2] + self.debug_print("looptimes is", looptimes) + self.loop_times = int(looptimes) + self.loop_finish = True + for caselist in list: + newdict = {} + newdict[tripped_casename] = {} + caselist.append(newdict) + self.loop_end = len(caselist) + continue + if len(tripped_caselist) == 3 and \ tripped_caselist[1] == "times": times = tripped_caselist[2] -- 1.7.4.4

3 6

[libvirt] [PATCH]: C# bindings. Added some virStream bindings (not all right now)
by Anthony Fox 18 Jun '11

18 Jun '11

2 2

[libvirt] [PATCH 0/6] Introduce virDomainGetControlInfo API
by Jiri Denemark 16 Jun '11

16 Jun '11

The API can be used to query current state of an interface to VMM used to control a domain. In QEMU world this translates into monitor connection and one can use the API, e.g., to check whether (and for how long) libvirtd is currently waiting for a reply from qemu process. Jiri Denemark (6): Introduce virTimeMs for getting current time in ms Use virTimeMs when appropriate Introduce virDomainGetControlInfo API Wire protocol and remote driver for virDomainGetControlInfo qemu: Implement virDomainGetControlInfo virsh: Add support for virDomainGetControlInfo daemon/remote_generator.pl | 20 +++++++--- include/libvirt/libvirt.h.in | 40 +++++++++++++++++++++ python/generator.py | 1 + python/libvirt-override-api.xml | 6 +++ src/driver.h | 5 +++ src/libvirt.c | 48 ++++++++++++++++++++++++++ src/libvirt_private.syms | 1 + src/libvirt_public.syms | 5 +++ src/qemu/qemu_domain.c | 28 ++++++--------- src/qemu/qemu_domain.h | 2 + src/qemu/qemu_driver.c | 72 ++++++++++++++++++++++++++++++++++----- src/qemu/qemu_migration.c | 31 +++++----------- src/qemu/qemu_process.c | 3 ++ src/remote/remote_driver.c | 1 + src/remote/remote_protocol.x | 14 +++++++- src/remote_protocol-structs | 9 +++++ src/util/event_poll.c | 45 +++++++++--------------- src/util/util.c | 24 +++++++++++++ src/util/util.h | 2 + tools/virsh.c | 66 +++++++++++++++++++++++++++++++++++ tools/virsh.pod | 7 ++++ 21 files changed, 348 insertions(+), 82 deletions(-) -- 1.7.5.3

4 21

[libvirt] [PATCH RFC 0/5] support cpu bandwidth in libvirt
by Wen Congyang 16 Jun '11

16 Jun '11

I need this feature immediately after CFS bandwidth patchset is merged into upstream kernel. So I prepare this patchset, and post it here for reviewing before CFS bandwidth is merged into upstream kernel. quota is an optional attribute specified in microseconds not a percentage of period. TODO: 1. quota should be in the range [1000, 18446744073709551(2^64/1000)] or less than 0. But I only limit it less or equal than 2^64/1000. Does anyone have a better way to limit quota? Wen Congyang (5): cgroup: Implement cpu.cfs_period_us and cpu.cfs_quota_us tuning API Update XML Schema for new entries qemu: Implement period and quota tunable XML configuration and parsing. qemu: Implement cfs_period and cfs_quota's modification doc: Add documentation for new cputune elements period and quota docs/formatdomain.html.in | 19 +++ docs/schemas/domain.rng | 25 ++++- src/conf/domain_conf.c | 20 +++- src/conf/domain_conf.h | 2 + src/libvirt_private.syms | 4 + src/qemu/qemu_cgroup.c | 43 +++++- src/qemu/qemu_driver.c | 162 +++++++++++++++++++---- src/util/cgroup.c | 81 +++++++++++- src/util/cgroup.h | 6 + tests/qemuxml2argvdata/qemuxml2argv-cputune.xml | 2 + 10 files changed, 328 insertions(+), 36 deletions(-)

2 9

[libvirt] [RFC PATCH] Set and reset MAC for PASSTHROUGH mode
by Dirk Herrendoerfer 16 Jun '11

16 Jun '11

there is a problem present in libvirt that when a PASSTHROUGH mode DIRECT NIC connection is made the MAC address of the NIC is not automatically set and reset to the configured VM MAC and back again. The attached patch fixes this problem by setting and resetting the MAC while remembering the previous setting while the VM is running. This also works if libvirtd is restarted while the VM is running. The storing of the previous MAC address is done in a link ( currently in /tmp, but should probably go into /var/run somewhere ). Best regards Dirk

3 5