- Devel - libvirt List Archives

[libvirt] [PATCH v3 0/4] Introduce new cputune event
by Pavel Hrdina 15 Sep '14

15 Sep '14

This patch series introduce new cputune event to inform management applications about every change of cputune values for running domains. There is missing documentation for all events so the documentation for this event will be part of the patches to document all events. Changes from v2: - removed virSnprintf helper - fix typo and PM -> PowerManagement in the first patch - fixed memory leak - updated virConnectDomainEventCputuneCallback docs - virConnectDomainEventCputuneCallback now consumes params - updated call sequence in virConnectDomainEventCputuneCallback Pavel Hrdina (4): domain_conf: separate structures from virDomainDef event: introduce new event for cputune add an example how to use cputune event cputune_event: queue the event for cputune updates daemon/remote.c | 45 +++++++++++++++++ examples/object-events/event-test.c | 52 +++++++++++++++++++- include/libvirt/libvirt.h.in | 22 +++++++++ src/conf/domain_conf.h | 98 ++++++++++++++++++++++--------------- src/conf/domain_event.c | 93 +++++++++++++++++++++++++++++++++++ src/conf/domain_event.h | 9 ++++ src/libvirt_private.syms | 2 + src/qemu/qemu_cgroup.c | 18 ++++++- src/qemu/qemu_driver.c | 74 ++++++++++++++++++++++++++++ src/remote/remote_driver.c | 42 ++++++++++++++++ src/remote/remote_protocol.x | 14 +++++- src/remote_protocol-structs | 9 ++++ tools/virsh-domain.c | 33 +++++++++++++ 13 files changed, 469 insertions(+), 42 deletions(-) -- 1.8.5.5

2 7

[libvirt] [PATCH] configure: Require setns in glibc
by Michal Privoznik 15 Sep '14

15 Sep '14

Instead of writing our own wrapper over setns function, require it in glibc (when compiling with LXC). Our implementation uses private macros from kernel header files which should not be done. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- This is an alternative approach as requested by Martin. configure.ac | 11 ++++++++++- src/util/virprocess.c | 33 --------------------------------- 2 files changed, 10 insertions(+), 34 deletions(-) diff --git a/configure.ac b/configure.ac index b4fb99a..0366b78 100644 --- a/configure.ac +++ b/configure.ac @@ -274,9 +274,18 @@ dnl Availability of various common functions (non-fatal if missing), dnl and various less common threadsafe functions AC_CHECK_FUNCS_ONCE([cfmakeraw fallocate geteuid getgid getgrnam_r \ getmntent_r getpwuid_r getuid kill mmap newlocale posix_fallocate \ - posix_memalign prlimit regexec sched_getaffinity setgroups setns \ + posix_memalign prlimit regexec sched_getaffinity setgroups \ setrlimit symlink sysctlbyname getifaddrs]) +AC_CHECK_FUNC([setns]) +if test "$ac_cv_func_setns" = "no" ; then + if "$with_lxc" = "check" ; then + with_lxc=no + elif "$with_lxc" = "yes" ; then + AC_MSG_ERROR([setns function is required for LXC]) + fi +fi + dnl Availability of pthread functions. Because of $LIB_PTHREAD, we dnl cannot use AC_CHECK_FUNCS_ONCE. LIB_PTHREAD and LIBMULTITHREAD dnl were set during gl_INIT by gnulib. diff --git a/src/util/virprocess.c b/src/util/virprocess.c index 3dae1bd..7c5f770 100644 --- a/src/util/virprocess.c +++ b/src/util/virprocess.c @@ -61,39 +61,6 @@ VIR_LOG_INIT("util.process"); -/* - * Workaround older glibc. While kernel may support the setns - * syscall, the glibc wrapper might not exist. If that's the - * case, use our own. - */ -#ifndef __NR_setns -# if defined(__x86_64__) -# define __NR_setns 308 -# elif defined(__i386__) -# define __NR_setns 346 -# else -# error "__NR_setns is not defined" -# endif -#endif - -#ifndef HAVE_SETNS -# ifndef WIN32 -# include <sys/syscall.h> - -static inline int setns(int fd, int nstype) -{ - return syscall(__NR_setns, fd, nstype); -} -# else -static inline int setns(int fd ATTRIBUTE_UNUSED, int nstype ATTRIBUTE_UNUSED) -{ - virReportSystemError(ENOSYS, "%s", - _("Namespaces are not supported on windows.")); - return -1; -} -# endif /* WIN32 */ -#endif /* HAVE_SETNS */ - /** * virProcessTranslateStatus: * @status: child exit status to translate -- 1.8.5.5

2 1

[libvirt] [PATCH v2 0/3] add support for mapping hugepages as shared
by Martin Kletzander 15 Sep '14

15 Sep '14

Changes to v1: - 1/3 is now not connected to the rest of the patches any more - whitespaces - memShared='on/off' -> memAccess='private/shared' Martin Kletzander (3): schemas: finish virTristate{Bool,Switch} transition docs, conf, schema: add support for shared memory mapping qemu: add support for shared memory mapping docs/formatdomain.html.in | 7 +- docs/schemas/basictypes.rng | 19 ++- docs/schemas/capability.rng | 10 +- docs/schemas/domaincaps.rng | 5 +- docs/schemas/domaincommon.rng | 163 +++++---------------- docs/schemas/interface.rng | 19 +-- docs/schemas/network.rng | 29 +--- docs/schemas/nwfilter.rng | 5 +- docs/schemas/secret.rng | 10 +- src/conf/cpu_conf.c | 30 +++- src/conf/cpu_conf.h | 17 ++- src/qemu/qemu_command.c | 27 ++++ .../qemuxml2argv-cpu-numa-memshared.xml | 28 ++++ .../qemuxml2argv-hugepages-shared.args | 16 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++++++ tests/qemuxml2argvtest.c | 5 + tests/qemuxml2xmltest.c | 2 + 17 files changed, 240 insertions(+), 197 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml -- 2.1.0

2 4

[libvirt] [PATCH 0/3] add support for hugepages mapping as shared
by Martin Kletzander 15 Sep '14

15 Sep '14

- BLURB Martin Kletzander (3): schemas: finish virTristate{Bool,Switch} transition docs, conf, schema: add support for shared memory mapping qemu: add support for shared memory mapping docs/formatdomain.html.in | 7 +- docs/schemas/basictypes.rng | 19 ++- docs/schemas/capability.rng | 10 +- docs/schemas/domaincaps.rng | 5 +- docs/schemas/domaincommon.rng | 160 +++++---------------- docs/schemas/interface.rng | 19 +-- docs/schemas/network.rng | 29 +--- docs/schemas/nwfilter.rng | 5 +- docs/schemas/secret.rng | 10 +- src/conf/cpu_conf.c | 25 +++- src/conf/cpu_conf.h | 7 +- src/qemu/qemu_command.c | 19 +++ .../qemuxml2argv-cpu-numa-memshared.xml | 28 ++++ .../qemuxml2argv-hugepages-shared.args | 16 +++ .../qemuxml2argv-hugepages-shared.xml | 45 ++++++ tests/qemuxml2argvtest.c | 5 + tests/qemuxml2xmltest.c | 2 + 17 files changed, 214 insertions(+), 197 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml -- 2.1.0

5 15

[libvirt] [PATCHv2] network: check for invalid forward delay time
by Erik Skultety 15 Sep '14

15 Sep '14

When spanning tree protocol is allowed in bridge settings, forward delay value is set as well (default is 0 if omitted). Until now, there was no check for delay value validity. Delay makes sense only as a positive numerical value. Note: However, even if you provide positive numerical value, brctl utility only uses values from range <2,30>, so the number provided can be modified (kernel most likely) to fall within this range. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1125764 --- docs/schemas/network.rng | 2 +- src/conf/network_conf.c | 38 +++++++++++++++++++++++--------------- src/util/virxml.c | 2 +- 3 files changed, 25 insertions(+), 17 deletions(-) diff --git a/docs/schemas/network.rng b/docs/schemas/network.rng index 0e7da89..ab41814 100644 --- a/docs/schemas/network.rng +++ b/docs/schemas/network.rng @@ -62,7 +62,7 @@ <optional> <attribute name="delay"> - <data type="integer"/> + <data type="unsignedLong"/> </attribute> </optional> diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index 9571ee1..e9ea746 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -2003,7 +2003,7 @@ static virNetworkDefPtr virNetworkDefParseXML(xmlXPathContextPtr ctxt) { virNetworkDefPtr def; - char *tmp; + char *tmp = NULL; char *stp = NULL; xmlNodePtr *ipNodes = NULL; xmlNodePtr *routeNodes = NULL; @@ -2037,7 +2037,6 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) } } else { if (virUUIDParse(tmp, def->uuid) < 0) { - VIR_FREE(tmp); virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("malformed uuid element")); goto error; @@ -2078,8 +2077,16 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) stp = virXPathString("string(./bridge[1]/@stp)", ctxt); def->stp = (stp && STREQ(stp, "off")) ? false : true; - if (virXPathULong("string(./bridge[1]/@delay)", ctxt, &def->delay) < 0) - def->delay = 0; + tmp = virXPathString("string(./bridge[1]/@delay)", ctxt); + if (tmp) { + if (virStrToLong_ulp(tmp, NULL, 10, &def->delay) < 0) { + virReportError(VIR_ERR_XML_ERROR, + _("Invalid delay value in network '%s'"), + def->name); + goto error; + } + } + VIR_FREE(tmp); tmp = virXPathString("string(./mac[1]/@address)", ctxt); if (tmp) { @@ -2087,14 +2094,12 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) virReportError(VIR_ERR_XML_ERROR, _("Invalid bridge mac address '%s' in network '%s'"), tmp, def->name); - VIR_FREE(tmp); goto error; } if (virMacAddrIsMulticast(&def->mac)) { virReportError(VIR_ERR_XML_ERROR, _("Invalid multicast bridge mac address '%s' in network '%s'"), tmp, def->name); - VIR_FREE(tmp); goto error; } VIR_FREE(tmp); @@ -2126,9 +2131,9 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) goto error; /* parse each portgroup */ for (i = 0; i < nPortGroups; i++) { - int ret = virNetworkPortGroupParseXML(&def->portGroups[i], - portGroupNodes[i], ctxt); - if (ret < 0) + if (virNetworkPortGroupParseXML(&def->portGroups[i], + portGroupNodes[i], + ctxt) < 0) goto error; def->nPortGroups++; } @@ -2147,9 +2152,10 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) goto error; /* parse each addr */ for (i = 0; i < nIps; i++) { - int ret = virNetworkIPDefParseXML(def->name, ipNodes[i], - ctxt, &def->ips[i]); - if (ret < 0) + if (virNetworkIPDefParseXML(def->name, + ipNodes[i], + ctxt, + &def->ips[i]) < 0) goto error; def->nips++; } @@ -2168,9 +2174,10 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) goto error; /* parse each definition */ for (i = 0; i < nRoutes; i++) { - int ret = virNetworkRouteDefParseXML(def->name, routeNodes[i], - ctxt, &def->routes[i]); - if (ret < 0) + if (virNetworkRouteDefParseXML(def->name, + routeNodes[i], + ctxt, + &def->routes[i]) < 0) goto error; def->nroutes++; } @@ -2289,6 +2296,7 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) return def; error: + VIR_FREE(tmp); VIR_FREE(routeNodes); VIR_FREE(stp); virNetworkDefFree(def); diff --git a/src/util/virxml.c b/src/util/virxml.c index cc4a85c..a91da05 100644 --- a/src/util/virxml.c +++ b/src/util/virxml.c @@ -420,7 +420,7 @@ virXPathULongLong(const char *xpath, } /** - * virXPathULongLong: + * virXPathLongLong: * @xpath: the XPath string to evaluate * @ctxt: an XPath context * @value: the returned long long value -- 1.9.3

2 1

[libvirt] [PATCH] openvz: fixed an config file parsing error
by Hongbin Lu 15 Sep '14

15 Sep '14

The OpenVZ driver reported an error on parsing some OpenVZ config parameters (e.g. diskspace). This issue is due to the driver made two incorrect assumptions about the value of the parameters: 1. Assume paramaeter is just a number (without unit suffix). 2. Assume the number is an integer. Actually, an OpenVZ config parameter may consist of a scalar and a unit, and the scalar is not necessary an integer (e.g. 2.2G). This patch is for fixing this issue. --- src/openvz/openvz_conf.c | 109 +++++++++++++++++++++++++++++++++++++++++----- src/util/virutil.c | 84 ++++++++++++++++++++++++----------- src/util/virutil.h | 3 + 3 files changed, 157 insertions(+), 39 deletions(-) diff --git a/src/openvz/openvz_conf.c b/src/openvz/openvz_conf.c index 856c9f5..0b163b4 100644 --- a/src/openvz/openvz_conf.c +++ b/src/openvz/openvz_conf.c @@ -51,6 +51,7 @@ #include "virfile.h" #include "vircommand.h" #include "virstring.h" +#include "c-ctype.h" #define VIR_FROM_THIS VIR_FROM_OPENVZ @@ -127,6 +128,32 @@ int openvzExtractVersion(struct openvz_driver *driver) } +/* Split a string, which is a config parameter, into a scalar and a unit. + * Recognized unit: + * - Gigabyte: 'G' or 'g' + * - Megabyte: 'M' or 'm' + * - Kilobyte: 'K' or 'k' + * - Page: 'P' or 'g' */ +static int +openvzParseParamUnit(char* str, char *unit) +{ + char last; + int len = strlen(str); + + if (len == 0) + return -1; + + last = c_tolower(str[len - 1]); + if (last == 'g' || last == 'm' || last == 'k' || last == 'p') { + str[len - 1] = '\0'; + if (unit) + *unit = last; + } + + return 0; +} + + /* Parse config values of the form barrier:limit into barrier and limit */ static int openvzParseBarrierLimit(const char* value, @@ -146,7 +173,59 @@ openvzParseBarrierLimit(const char* value, goto error; } else { if (barrier != NULL) { - if (virStrToLong_ull(token, NULL, 10, barrier)) + if (openvzParseParamUnit(token, NULL) < 0) + goto error; + + if (virStrToLong_ull(token, NULL, 10, barrier) < 0) + goto error; + } + } + token = strtok_r(NULL, ":", &saveptr); + if (token == NULL) { + goto error; + } else { + if (limit != NULL) { + if (openvzParseParamUnit(token, NULL) < 0) + goto error; + + if (virStrToLong_ull(token, NULL, 10, limit) < 0) + goto error; + } + } + ret = 0; + error: + VIR_FREE(str); + return ret; +} + + +/* Just like openvzParseBarrierLimit, above, but produce "double" + * barrier and limit. This version also outputs the units of barrier + * and limit, which are bunit and lunit, if they are present.*/ +static int +openvzParseBarrierLimitDouble(const char* value, + double *barrier, + double *limit, + char *bunit, + char *lunit) +{ + char *token; + char *saveptr = NULL; + char *str; + int ret = -1; + + if (VIR_STRDUP(str, value) < 0) + goto error; + + token = strtok_r(str, ":", &saveptr); + if (token == NULL) { + goto error; + } else { + if (barrier != NULL) { + if (openvzParseParamUnit(token, bunit) < 0) + goto error; + + if (virStrToDouble(token, NULL, barrier) < 0) goto error; } } @@ -155,7 +234,10 @@ openvzParseBarrierLimit(const char* value, goto error; } else { if (limit != NULL) { - if (virStrToLong_ull(token, NULL, 10, limit)) + if (openvzParseParamUnit(token, lunit) < 0) + goto error; + + if (virStrToDouble(token, NULL, limit) < 0) goto error; } } @@ -352,7 +434,8 @@ openvzReadFSConf(virDomainDefPtr def, char *veid_str = NULL; char *temp = NULL; const char *param; - unsigned long long barrier, limit; + double barrier, limit; + char bunit, lunit; ret = openvzReadVPSConfigParam(veid, "OSTEMPLATE", &temp); if (ret < 0) { @@ -396,21 +479,23 @@ openvzReadFSConf(virDomainDefPtr def, param = "DISKSPACE"; ret = openvzReadVPSConfigParam(veid, param, &temp); if (ret > 0) { - if (openvzParseBarrierLimit(temp, &barrier, &limit)) { + bunit = 'k'; /* default unit is kilobyte */ + lunit = 'k'; /* default unit is kilobyte */ + + if (openvzParseBarrierLimitDouble(temp, &barrier, &limit, &bunit, &lunit)) { virReportError(VIR_ERR_INTERNAL_ERROR, _("Could not read '%s' from config for container %d"), param, veid); goto error; } else { - /* Ensure that we can multiply by 1024 without overflowing. */ - if (barrier > ULLONG_MAX / 1024 || - limit > ULLONG_MAX / 1024) { - virReportSystemError(VIR_ERR_OVERFLOW, "%s", - _("Unable to parse quota")); + if (virScaleDouble(&barrier, &bunit) < 0) goto error; - } - fs->space_soft_limit = barrier * 1024; /* unit is bytes */ - fs->space_hard_limit = limit * 1024; /* unit is bytes */ + + if (virScaleDouble(&limit, &lunit) < 0) + goto error; + + fs->space_soft_limit = (unsigned long long)barrier; + fs->space_hard_limit = (unsigned long long)limit; } } diff --git a/src/util/virutil.c b/src/util/virutil.c index 2edbec5..3b6e617 100644 --- a/src/util/virutil.c +++ b/src/util/virutil.c @@ -268,28 +268,12 @@ virHexToBin(unsigned char c) } } -/* Scale an integer VALUE in-place by an optional case-insensitive - * SUFFIX, defaulting to SCALE if suffix is NULL or empty (scale is - * typically 1 or 1024). Recognized suffixes include 'b' or 'bytes', - * as well as power-of-two scaling via binary abbreviations ('KiB', - * 'MiB', ...) or their one-letter counterpart ('k', 'M', ...), and - * power-of-ten scaling via SI abbreviations ('KB', 'MB', ...). - * Ensure that the result does not exceed LIMIT. Return 0 on success, - * -1 with error message raised on failure. */ int -virScaleInteger(unsigned long long *value, const char *suffix, - unsigned long long scale, unsigned long long limit) +virSuffixToScale(const char *suffix, unsigned long long *scale) { - if (!suffix || !*suffix) { - if (!scale) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("invalid scale %llu"), scale); - return -1; - } - suffix = ""; - } else if (STRCASEEQ(suffix, "b") || STRCASEEQ(suffix, "byte") || + if (STRCASEEQ(suffix, "b") || STRCASEEQ(suffix, "byte") || STRCASEEQ(suffix, "bytes")) { - scale = 1; + *scale = 1; } else { int base; @@ -299,28 +283,28 @@ virScaleInteger(unsigned long long *value, const char *suffix, base = 1000; } else { virReportError(VIR_ERR_INVALID_ARG, - _("unknown suffix '%s'"), suffix); + _("unknown suffix '%s'"), suffix); return -1; } - scale = 1; + *scale = 1; switch (c_tolower(*suffix)) { case 'e': - scale *= base; + *scale *= base; /* fallthrough */ case 'p': - scale *= base; + *scale *= base; /* fallthrough */ case 't': - scale *= base; + *scale *= base; /* fallthrough */ case 'g': - scale *= base; + *scale *= base; /* fallthrough */ case 'm': - scale *= base; + *scale *= base; /* fallthrough */ case 'k': - scale *= base; + *scale *= base; break; default: virReportError(VIR_ERR_INVALID_ARG, @@ -329,6 +313,34 @@ virScaleInteger(unsigned long long *value, const char *suffix, } } + return 0; +} + +/* Scale an integer VALUE in-place by an optional case-insensitive + * SUFFIX, defaulting to SCALE if suffix is NULL or empty (scale is + * typically 1 or 1024). Recognized suffixes include 'b' or 'bytes', + * as well as power-of-two scaling via binary abbreviations ('KiB', + * 'MiB', ...) or their one-letter counterpart ('k', 'M', ...), and + * power-of-ten scaling via SI abbreviations ('KB', 'MB', ...). + * Ensure that the result does not exceed LIMIT. Return 0 on success, + * -1 with error message raised on failure. */ +int +virScaleInteger(unsigned long long *value, const char *suffix, + unsigned long long scale, unsigned long long limit) +{ + if (!suffix || !*suffix) { + if (!scale) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("invalid scale %llu"), scale); + return -1; + } + suffix = ""; + } else if (virSuffixToScale(suffix, &scale) < 0) { + virReportError(VIR_ERR_INVALID_ARG, + _("unable to parse suffix '%s'"), suffix); + return -1; + } + if (*value && *value > (limit / scale)) { virReportError(VIR_ERR_OVERFLOW, _("value too large: %llu%s"), *value, suffix); @@ -338,6 +350,24 @@ virScaleInteger(unsigned long long *value, const char *suffix, return 0; } +/* Just like virScaleInteger, above, but produce an "double" value. */ +int +virScaleDouble(double *value, const char *suffix) +{ + unsigned long long scale; + + if (!suffix || !*suffix) + return -1; + + if (virSuffixToScale(suffix, &scale) < 0) { + virReportError(VIR_ERR_INVALID_ARG, + _("unable to parse suffix '%s'"), suffix); + return -1; + } + *value *= scale; + return 0; +} + /** * virParseNumber: diff --git a/src/util/virutil.h b/src/util/virutil.h index 89b7923..d74b581 100644 --- a/src/util/virutil.h +++ b/src/util/virutil.h @@ -57,6 +57,9 @@ int virScaleInteger(unsigned long long *value, const char *suffix, unsigned long long scale, unsigned long long limit) ATTRIBUTE_NONNULL(1) ATTRIBUTE_RETURN_CHECK; +int virScaleDouble(double *value, const char *suffix) + ATTRIBUTE_NONNULL(1) ATTRIBUTE_RETURN_CHECK; + int virHexToBin(unsigned char c); int virParseNumber(const char **str); -- 1.7.1

3 2

[libvirt] [PATCH] virprocess: Introduce our own setns() wrapper
by Michal Privoznik 15 Sep '14

15 Sep '14

>From time to time weird bugreports occur on the list, e.g [1]. Even though the kernel supports setns syscall, there's an older glibc in the system that misses a wrapper over the syscall. Hence, after the configure phase we think there's no setns support in the system, which is obviously wrong. On the other hand, we can't rely on linux distributions to provide newer glibc soon. Therefore we need to introduce the wrapper on or own. 1: https://www.redhat.com/archives/libvir-list/2014-September/msg00492.html Signed-off-by: Stephan Sachse <ste.sachse(a)gmail.com> Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- The original patch was ACKed already, I've just rebased it to the current master, and moved the wrapper declaration into virprocess.c as we don't need to pollute the header file. I've pushed this then. src/util/virprocess.c | 57 ++++++++++++++++++++++----------------------------- 1 file changed, 24 insertions(+), 33 deletions(-) diff --git a/src/util/virprocess.c b/src/util/virprocess.c index 97cce4f..b6fcbac 100644 --- a/src/util/virprocess.c +++ b/src/util/virprocess.c @@ -27,6 +27,8 @@ #include <errno.h> #include <stdlib.h> #include <sys/wait.h> +#include <unistd.h> +#include <sys/syscall.h> #if HAVE_SETRLIMIT # include <sys/time.h> # include <sys/resource.h> @@ -60,6 +62,28 @@ VIR_LOG_INIT("util.process"); +/* + * Workaround older glibc. While kernel may support the setns + * syscall, the glibc wrapper might not exist. If that's the + * case, use our own. + */ +#ifndef __NR_setns +# if defined(__x86_64__) +# define __NR_setns 308 +# elif defined(__i386__) +# define __NR_setns 346 +# else +# error "__NR_setns is not defined" +# endif +#endif + +#ifndef HAVE_SETNS +static inline int setns(int fd, int nstype) +{ + return syscall(__NR_setns, fd, nstype); +} +#endif /* HAVE_SETNS */ + /** * virProcessTranslateStatus: * @status: child exit status to translate @@ -559,7 +583,6 @@ int virProcessGetAffinity(pid_t pid ATTRIBUTE_UNUSED, #endif /* HAVE_SCHED_GETAFFINITY */ -#if HAVE_SETNS int virProcessGetNamespaces(pid_t pid, size_t *nfdlist, int **fdlist) @@ -630,26 +653,6 @@ int virProcessSetNamespaces(size_t nfdlist, } return 0; } -#else /* ! HAVE_SETNS */ -int virProcessGetNamespaces(pid_t pid, - size_t *nfdlist ATTRIBUTE_UNUSED, - int **fdlist ATTRIBUTE_UNUSED) -{ - virReportSystemError(ENOSYS, - _("Cannot get namespaces for %llu"), - (unsigned long long)pid); - return -1; -} - - -int virProcessSetNamespaces(size_t nfdlist ATTRIBUTE_UNUSED, - int *fdlist ATTRIBUTE_UNUSED) -{ - virReportSystemError(ENOSYS, "%s", - _("Cannot set namespaces")); - return -1; -} -#endif /* ! HAVE_SETNS */ #if HAVE_PRLIMIT static int @@ -905,7 +908,6 @@ int virProcessGetStartTime(pid_t pid, #endif -#ifdef HAVE_SETNS static int virProcessNamespaceHelper(int errfd, pid_t pid, virProcessNamespaceCallback cb, @@ -992,17 +994,6 @@ virProcessRunInMountNamespace(pid_t pid, VIR_FORCE_CLOSE(errfd[1]); return ret; } -#else /* !HAVE_SETNS */ -int -virProcessRunInMountNamespace(pid_t pid ATTRIBUTE_UNUSED, - virProcessNamespaceCallback cb ATTRIBUTE_UNUSED, - void *opaque ATTRIBUTE_UNUSED) -{ - virReportSystemError(ENOSYS, "%s", - _("Mount namespaces are not available on this platform")); - return -1; -} -#endif /** -- 1.8.5.5

3 3

[libvirt] [PATCH] node_device_udev: Try harder to get human readable vendor:product
by Lubomir Rintel 15 Sep '14

15 Sep '14

The manufacurer and product from USB device itself are usually not particularly useful -- they tend to be missing, or ugly (all-uppercase, padded with spaces, etc.). Prefer what's in the usb id database and fall back to descriptors only if the device is too now to be in database. --- Before: http://v3.sk/~lkundrak/virt/old.png After: http://v3.sk/~lkundrak/virt/libvirt.png src/node_device/node_device_udev.c | 27 +++++++++++++++++++++------ 1 file changed, 21 insertions(+), 6 deletions(-) diff --git a/src/node_device/node_device_udev.c b/src/node_device/node_device_udev.c index 0fe474d..53792f0 100644 --- a/src/node_device/node_device_udev.c +++ b/src/node_device/node_device_udev.c @@ -580,6 +580,7 @@ static int udevProcessUSBDevice(struct udev_device *device, { union _virNodeDevCapData *data = &def->caps->data; int ret = -1; + int err; if (udevGetUintProperty(device, "BUSNUM", @@ -602,10 +603,17 @@ static int udevProcessUSBDevice(struct udev_device *device, goto out; } - if (udevGetStringSysfsAttr(device, - "manufacturer", - &data->usb_dev.vendor_name) == PROPERTY_ERROR) { + err = udevGetStringProperty(device, + "ID_VENDOR_FROM_DATABASE", + &data->usb_dev.vendor_name); + if (err == PROPERTY_ERROR) goto out; + if (err == PROPERTY_MISSING) { + if (udevGetStringSysfsAttr(device, + "manufacturer", + &data->usb_dev.vendor_name) == PROPERTY_ERROR) { + goto out; + } } if (udevGetUintProperty(device, @@ -615,10 +623,17 @@ static int udevProcessUSBDevice(struct udev_device *device, goto out; } - if (udevGetStringSysfsAttr(device, - "product", - &data->usb_dev.product_name) == PROPERTY_ERROR) { + err = udevGetStringProperty(device, + "ID_MODEL_FROM_DATABASE", + &data->usb_dev.product_name); + if (err == PROPERTY_ERROR) goto out; + if (err == PROPERTY_MISSING) { + if (udevGetStringSysfsAttr(device, + "product", + &data->usb_dev.product_name) == PROPERTY_ERROR) { + goto out; + } } if (udevGenerateDeviceName(device, def, NULL) != 0) { -- 1.9.3

3 2

[libvirt] [PATCH v2 0/5] Introduce new cputune event
by Pavel Hrdina 15 Sep '14

15 Sep '14

This patch series introduce new cputune event to inform management applications about every change of cputune values for running domains. The first patch introduces a new string helper virSnprintf and the second patch cleans up the virDomanDef structure. There is missing documentation for all events so the documentation for this event will be part of the patches to document all events. Changes from v1: - new virSnprintf helper - complete cleanup of virDomainDef structure - cputune event now uses virTypedParameter Pavel Hrdina (5): introduce virSnprintf helper domain_conf: separate structures from virDomainDef event: introduce new event for cputune add an example how to use cputune event cputune_event: queue the event for cputune updates daemon/remote.c | 45 ++++++++++++++++++ examples/object-events/event-test.c | 52 +++++++++++++++++++- include/libvirt/libvirt.h.in | 19 ++++++++ src/conf/domain_conf.h | 94 ++++++++++++++++++++++--------------- src/conf/domain_event.c | 84 +++++++++++++++++++++++++++++++++ src/conf/domain_event.h | 9 ++++ src/libvirt_private.syms | 3 ++ src/qemu/qemu_cgroup.c | 18 ++++++- src/qemu/qemu_driver.c | 73 ++++++++++++++++++++++++++++ src/remote/remote_driver.c | 41 ++++++++++++++++ src/remote/remote_protocol.x | 14 +++++- src/remote_protocol-structs | 9 ++++ src/util/virstring.c | 25 ++++++++++ src/util/virstring.h | 20 ++++++++ tools/virsh-domain.c | 33 +++++++++++++ 15 files changed, 499 insertions(+), 40 deletions(-) -- 1.8.5.5

3 19

[libvirt] [PATCH V2 0/2] cpu: Handle only high order 16 bits of PVR for IBM Power processors
by Pradipta Kr. Banerjee 15 Sep '14

15 Sep '14

IBM Power processors encode PVR as CPU family in high order 16 bits and a CPU version in lower 16 bits. Since there is no significant change in behavior between versions, there is no point adding every single CPU version in cpu_map.xml. Qemu for ppc64 already have the necessary changes. V2: - Incorporate Martin comments from V1 - Rewrite the cpu_map.xml for IBM Power processors Pradipta Kr. Banerjee (2): Handle only high order 16 bits for PVR Update cputest test cases for the updated IBM Power processor models src/cpu/cpu_map.xml | 22 ++++++++-------------- src/cpu/cpu_powerpc.c | 7 ++++++- tests/cputest.c | 4 ++-- tests/cputestdata/ppc64-baseline-1-result.xml | 2 +- .../ppc64-baseline-incompatible-vendors.xml | 4 ++-- .../ppc64-baseline-no-vendor-result.xml | 2 +- tests/cputestdata/ppc64-baseline-no-vendor.xml | 2 +- tests/cputestdata/ppc64-exact.xml | 2 +- tests/cputestdata/ppc64-guest-nofallback.xml | 2 +- tests/cputestdata/ppc64-guest.xml | 2 +- .../ppc64-host+guest,ppc_models-result.xml | 2 +- ...st-nofallback,ppc_models,POWER7_v2.1-result.xml | 5 ----- tests/cputestdata/ppc64-host.xml | 2 +- tests/cputestdata/ppc64-strict.xml | 2 +- .../qemuxml2argv-pseries-cpu-exact.args | 4 ++-- .../qemuxml2argv-pseries-cpu-exact.xml | 2 +- 16 files changed, 30 insertions(+), 36 deletions(-) delete mode 100644 tests/cputestdata/ppc64-host+guest-nofallback,ppc_models,POWER7_v2.1-result.xml -- 1.9.3

3 4