June 2016 - Devel - libvirt List Archives

[libvirt] [PATCH 0/3] perf: Unbreak virsh and XML parser
by Peter Krempa 17 Jun '16

17 Jun '16

Yet another breakage in the perf event code. Add missing docs, fix schemas and rewrite the XML parser since it was broken. Peter Krempa (3): qemu: perf: Don't set state of first event for every other event docs: Add at least some docs and fix schema entry for perf events conf: Fix perf event parser docs/formatdomain.html.in | 44 +++++++++++++++++++++++++++ docs/schemas/domaincommon.rng | 2 ++ src/conf/domain_conf.c | 47 ++++++++++++----------------- src/qemu/qemu_driver.c | 2 +- tests/genericxml2xmlindata/generic-perf.xml | 22 ++++++++++++++ tests/genericxml2xmltest.c | 2 ++ 6 files changed, 90 insertions(+), 29 deletions(-) create mode 100644 tests/genericxml2xmlindata/generic-perf.xml -- 2.8.3

2 8

[libvirt] [PATCH] spec: Make driver-qemu require driver-storage
by Martin Kletzander 17 Jun '16

17 Jun '16

Without that we might get similar messages in the log: error : virDriverLoadModule:73 : failed to load module /usr/lib64/libvirt/connection-driver/libvirt_driver_qemu.so /usr/lib64/libvirt/connection-driver/libvirt_driver_qemu.so: undefined symbol: virStorageFileCreate Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- libvirt.spec.in | 1 + 1 file changed, 1 insertion(+) diff --git a/libvirt.spec.in b/libvirt.spec.in index 6b3e888d0048..816a950ef624 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -607,6 +607,7 @@ Group: Development/Libraries Requires: libvirt-daemon = %{version}-%{release} # There really is a hard cross-driver dependency here Requires: libvirt-daemon-driver-network = %{version}-%{release} +Requires: libvirt-daemon-driver-storage = %{version}-%{release} Requires: /usr/bin/qemu-img # For image compression Requires: gzip -- 2.9.0

2 1

[libvirt] [PATCH v2] util: Make failure to get suplementary group list for a uid non-fatal
by Peter Krempa 17 Jun '16

17 Jun '16

Since introduction of the DAC security driver we've documented that seclabels with a leading + can be used with numerical uid. This would not work though with the rest of libvirt if the uid was not actually used in the system as we'd fail when trying to get a list of suplementary groups for the given uid. Since a uid without entry in /etc/passwd (or other user database) will not have any suppolementary groups we can treat the failure to obtain them as such. This patch modifies virGetGroupList to not report the error of missing user and tweaks callers to treat the missing list as having 0 supplementary groups. The only place reporting errors is virt-login-shell as it's used to determine whether the given user is allowed to access the shell. --- Although it was ACKed I'm reposting a rebased version including the fixes and fixed merge conflicts. I'd like to hear feedback from the reporter of this issue. CC: Roy Keene <rkeene(a)knightpoint.com> src/security/security_dac.c | 13 +++++++------ src/util/vircommand.c | 4 +++- src/util/virfile.c | 28 ++++++++++++++++------------ src/util/virutil.c | 27 +++++++++++++++++---------- tools/virt-login-shell.c | 6 +++++- 5 files changed, 48 insertions(+), 30 deletions(-) diff --git a/src/security/security_dac.c b/src/security/security_dac.c index 442ce70..9dec201 100644 --- a/src/security/security_dac.c +++ b/src/security/security_dac.c @@ -266,14 +266,15 @@ static int virSecurityDACPreFork(virSecurityManagerPtr mgr) { virSecurityDACDataPtr priv = virSecurityManagerGetPrivateData(mgr); - int ngroups; VIR_FREE(priv->groups); - priv->ngroups = 0; - if ((ngroups = virGetGroupList(priv->user, priv->group, - &priv->groups)) < 0) - return -1; - priv->ngroups = ngroups; + + /* ignore a possible problem in getting supplementary groups just assume + * we have none and continue with uid/gid only */ + if ((priv->ngroups = virGetGroupList(priv->user, priv->group, + &priv->groups)) < 0) + priv->ngroups = 0; + return 0; } diff --git a/src/util/vircommand.c b/src/util/vircommand.c index f5bd7af..58af06a 100644 --- a/src/util/vircommand.c +++ b/src/util/vircommand.c @@ -554,8 +554,10 @@ virExec(virCommandPtr cmd) childerr = null; } + /* ignore a possible problem in getting supplementary groups just assume + * we have none and continue with uid/gid only */ if ((ngroups = virGetGroupList(cmd->uid, cmd->gid, &groups)) < 0) - goto cleanup; + ngroups = 0; pid = virFork(); diff --git a/src/util/virfile.c b/src/util/virfile.c index 9d460b9..4298ec5 100644 --- a/src/util/virfile.c +++ b/src/util/virfile.c @@ -1978,9 +1978,10 @@ virFileAccessibleAs(const char *path, int mode, gid == getegid()) return access(path, mode); - ngroups = virGetGroupList(uid, gid, &groups); - if (ngroups < 0) - return -1; + /* ignore a possible problem in getting supplementary groups just assume + * we have none and continue with uid/gid only */ + if ((ngroups = virGetGroupList(uid, gid, &groups)) == -1) + ngroups = 0; pid = virFork(); @@ -2104,9 +2105,10 @@ virFileOpenForked(const char *path, int openflags, mode_t mode, * following dance avoids problems caused by root-squashing * NFS servers. */ - ngroups = virGetGroupList(uid, gid, &groups); - if (ngroups < 0) - return -errno; + /* ignore a possible problem in getting supplementary groups just assume + * we have none and continue with uid/gid only */ + if ((ngroups = virGetGroupList(uid, gid, &groups)) == -1) + ngroups = 0; if (socketpair(AF_UNIX, SOCK_STREAM, 0, pair) < 0) { ret = -errno; @@ -2407,9 +2409,10 @@ virFileRemove(const char *path, if (gid == (gid_t) -1) gid = getegid(); - ngroups = virGetGroupList(uid, gid, &groups); - if (ngroups < 0) - return -errno; + /* ignore a possible problem in getting supplementary groups just assume + * we have none and continue with uid/gid only */ + if ((ngroups = virGetGroupList(uid, gid, &groups)) == -1) + ngroups = 0; pid = virFork(); @@ -2583,9 +2586,10 @@ virDirCreate(const char *path, if (gid == (gid_t) -1) gid = getegid(); - ngroups = virGetGroupList(uid, gid, &groups); - if (ngroups < 0) - return -errno; + /* ignore a possible problem in getting supplementary groups just assume + * we have none and continue with uid/gid only */ + if ((ngroups = virGetGroupList(uid, gid, &groups)) == -1) + ngroups = 0; pid = virFork(); diff --git a/src/util/virutil.c b/src/util/virutil.c index ff58054..c11278f 100644 --- a/src/util/virutil.c +++ b/src/util/virutil.c @@ -755,9 +755,10 @@ virGetUserDirectory(void) #ifdef HAVE_GETPWUID_R /* Look up fields from the user database for the given user. On - * error, set errno, report the error, and return -1. */ + * error, set errno, report the error if not instructed otherwise via @quiet, + * and return -1. */ static int -virGetUserEnt(uid_t uid, char **name, gid_t *group, char **dir, char **shell) +virGetUserEnt(uid_t uid, char **name, gid_t *group, char **dir, char **shell, bool quiet) { char *strbuf; struct passwd pwbuf; @@ -792,12 +793,19 @@ virGetUserEnt(uid_t uid, char **name, gid_t *group, char **dir, char **shell) if (VIR_RESIZE_N(strbuf, strbuflen, strbuflen, strbuflen) < 0) goto cleanup; } + if (rc != 0) { + if (quiet) + goto cleanup; + virReportSystemError(rc, _("Failed to find user record for uid '%u'"), (unsigned int) uid); goto cleanup; } else if (pw == NULL) { + if (quiet) + goto cleanup; + virReportError(VIR_ERR_SYSTEM_ERROR, _("Failed to find user record for uid '%u'"), (unsigned int) uid); @@ -882,7 +890,7 @@ char * virGetUserDirectoryByUID(uid_t uid) { char *ret; - virGetUserEnt(uid, NULL, NULL, &ret, NULL); + virGetUserEnt(uid, NULL, NULL, &ret, NULL, false); return ret; } @@ -890,7 +898,7 @@ virGetUserDirectoryByUID(uid_t uid) char *virGetUserShell(uid_t uid) { char *ret; - virGetUserEnt(uid, NULL, NULL, NULL, &ret); + virGetUserEnt(uid, NULL, NULL, NULL, &ret, false); return ret; } @@ -940,7 +948,7 @@ char *virGetUserRuntimeDirectory(void) char *virGetUserName(uid_t uid) { char *ret; - virGetUserEnt(uid, &ret, NULL, NULL, NULL); + virGetUserEnt(uid, &ret, NULL, NULL, NULL, false); return ret; } @@ -1113,8 +1121,9 @@ virGetGroupID(const char *group, gid_t *gid) /* Compute the list of primary and supplementary groups associated * with @uid, and including @gid in the list (unless it is -1), * storing a malloc'd result into @list. Return the size of the list - * on success, or -1 on failure with error reported and errno set. May - * not be called between fork and exec. */ + * on success, or -1 on failure with no error reported as this usually isn't + * a fatal problem for callers. errno is set on error. May not be called + * between fork and exec. */ int virGetGroupList(uid_t uid, gid_t gid, gid_t **list) { @@ -1126,14 +1135,12 @@ virGetGroupList(uid_t uid, gid_t gid, gid_t **list) if (uid == (uid_t)-1) return 0; - if (virGetUserEnt(uid, &user, &primary, NULL, NULL) < 0) + if (virGetUserEnt(uid, &user, &primary, NULL, NULL, true) < 0) return -1; ret = mgetgroups(user, primary, list); if (ret < 0) { sa_assert(!*list); - virReportSystemError(errno, - _("cannot get group list for '%s'"), user); goto cleanup; } diff --git a/tools/virt-login-shell.c b/tools/virt-login-shell.c index 38fcb9e..2ad6634 100644 --- a/tools/virt-login-shell.c +++ b/tools/virt-login-shell.c @@ -303,8 +303,12 @@ main(int argc, char **argv) if (!(conf = virConfReadFile(login_shell_path, 0))) goto cleanup; - if ((ngroups = virGetGroupList(uid, gid, &groups)) < 0) + if ((ngroups = virGetGroupList(uid, gid, &groups)) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("failed to query supplementary group list for uid '%u'"), + (unsigned int) uid); goto cleanup; + } if (virLoginShellAllowedUser(conf, name, groups) < 0) goto cleanup; -- 2.8.3

2 1

[libvirt] [PATCH v4 0/3] qemu: expand domain memory statistics
by Derbyshev Dmitriy 16 Jun '16

16 Jun '16

From: Derbyshev Dmitry <dderbyshev(a)virtuozzo.com> QEMU reports timestamp and available along with other memory statistics. This information was not saved into domain statistics. Also, to collect all balloon statistics for all guests it was necessary to make several libvirt requests (one per VE). Last patch allows doing this via qemuConnectGetAllDomainStats in one request. Changes since v1: * Enum numeration fixed * Macro getting "usage" field fixed Changes since v2: * previous patches were on wrong branch * qemu's stat name was "stat-available-memory" Changes since v3: * 3rd patch added Derbyshev Dmitry (3): qemu: expand domain memory statistics with 'usable' qemu: expand domain memory statistics with 'last-update' timestamp qemu: return balloon statistics alongside all domain stats include/libvirt/libvirt-domain.h | 11 +++++- src/libvirt-domain.c | 5 +++ src/qemu/qemu_driver.c | 84 +++++++++++++++++++++++++++++++--------- src/qemu/qemu_monitor_json.c | 22 ++++++----- tools/virsh-domain-monitor.c | 4 ++ 5 files changed, 98 insertions(+), 28 deletions(-) -- 1.9.5.msysgit.0

2 6

[libvirt] [PATCH] xenconfig: fix conversion of <driver> to backendtype
by Jim Fehlig 16 Jun '16

16 Jun '16

When converting domXML to xen xl.cfg, backendtype should not be emitted if <driver> is not specified. Moreover, <driver name='file'/> should be converted to backendtype qdisk, similar to handling of <driver> in libxlMakeDisk() in libxl_conf.c. Prior to this change, connectDomainXMLToNative would produce incorrect xl.cfg when the input domXML contained <driver name='file'/> domXML: <disk type="file" device="disk"> <driver name="file"/> <source file="/image/file/path"/> <target dev="xvda" bus="xen"/> </disk> virsh domxml-to-native xen-xl domXML disk = [ "format=raw,vdev=xvda,access=rw,backendtype=target=/image/file/path" ] xl create xl.cfg config parsing error in disk specification: unknown value for backendtype: near `target=/image/file/path' in `format=raw,vdev=xvda,access=rw,backendtype=target=/image/file/path' Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- src/xenconfig/xen_xl.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/src/xenconfig/xen_xl.c b/src/xenconfig/xen_xl.c index 5879c66..d524a82 100644 --- a/src/xenconfig/xen_xl.c +++ b/src/xenconfig/xen_xl.c @@ -857,13 +857,15 @@ xenFormatXLDisk(virConfValuePtr list, virDomainDiskDefPtr disk) } /* backendtype */ - virBufferAddLit(&buf, "backendtype="); - if (STREQ_NULLABLE(driver, "qemu")) - virBufferAddLit(&buf, "qdisk,"); - else if (STREQ_NULLABLE(driver, "tap")) - virBufferAddLit(&buf, "tap,"); - else if (STREQ_NULLABLE(driver, "phy")) - virBufferAddLit(&buf, "phy,"); + if (driver) { + virBufferAddLit(&buf, "backendtype="); + if (STREQ(driver, "qemu") || STREQ(driver, "file")) + virBufferAddLit(&buf, "qdisk,"); + else if (STREQ(driver, "tap")) + virBufferAddLit(&buf, "tap,"); + else if (STREQ(driver, "phy")) + virBufferAddLit(&buf, "phy,"); + } /* devtype */ if (disk->device == VIR_DOMAIN_DISK_DEVICE_CDROM) -- 2.1.4

2 2

[libvirt] [PATCH] util: fix missing broadcast address in bridge and tap device IP addresses
by Laine Stump 16 Jun '16

16 Jun '16

Commit b3d069872ce53eb added peer address setting to the low level virNetDevSetIPAddress() function, but ended up causing a segfault in cases where the caller passed NULL for peer address. Commit a3510e33d33e52c fixed the segfault, but managed to cause us to skip setting the broadcast address when setting an interface's IP address if the peer address was NULL. The result is that the broadcast address is 0.0.0.0 for all libvirt-created bridges (and for now also in interfaces in lxc containers with IP addresses set by libvirt). This was reported on the mailing list: https://www.redhat.com/archives/libvir-list/2016-June/msg00027.html but I was too busy to investigate at the time. I found it by accident today while refactoring virNetDevSetIPAddress(). Since this regression is present in the 1.3.5 release, I'm sending the bugfix as a separate simpler patch from my larger refactoring patchset. --- src/util/virnetdev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/util/virnetdev.c b/src/util/virnetdev.c index 925b41f..5a4ccc6 100644 --- a/src/util/virnetdev.c +++ b/src/util/virnetdev.c @@ -1129,7 +1129,8 @@ int virNetDevSetIPAddress(const char *ifname, unsigned int recvbuflen; /* The caller needs to provide a correct address */ - if (VIR_SOCKET_ADDR_FAMILY(addr) == AF_INET && peer && !VIR_SOCKET_ADDR_VALID(peer)) { + if (VIR_SOCKET_ADDR_FAMILY(addr) == AF_INET && + !(peer && VIR_SOCKET_ADDR_VALID(peer))) { /* compute a broadcast address if this is IPv4 */ if (VIR_ALLOC(broadcast) < 0) return -1; -- 2.4.11

2 2

[libvirt] [PATCH] util: remove redundant comments
by Chen Hanxiao 16 Jun '16

16 Jun '16

From: Chen Hanxiao <chenhanxiao(a)gmail.com> Signed-off-by: Chen Hanxiao <chenhanxiao(a)gmail.com> --- src/util/vireventpoll.c | 1 - 1 file changed, 1 deletion(-) diff --git a/src/util/vireventpoll.c b/src/util/vireventpoll.c index 5e99b3c..81ecab4 100644 --- a/src/util/vireventpoll.c +++ b/src/util/vireventpoll.c @@ -405,7 +405,6 @@ static struct pollfd *virEventPollMakePollFDs(int *nfds) { fds[*nfds].events = eventLoop.handles[i].events; fds[*nfds].revents = 0; (*nfds)++; - //EVENT_DEBUG("Wait for %d %d", eventLoop.handles[i].fd, eventLoop.handles[i].events); } return fds; -- 1.8.3.1

2 2

[libvirt] [PATCH python v2 0/2] Bindings for storage pool lifecycle events
by Jovanka Gulicoska 16 Jun '16

16 Jun '16

Changes since v1: use double newline spacing, change LIBVIRT_CHECK_VERSION to 2.0.0 Python buindings and tests for storage pool lifecycle events API. Function storageDetailToString() is not implemented for now. Jovanka Gulicoska (2): Python binding for storage pool lifecycle events API event-test: Add storage pool lifecycle event tests examples/event-test.py | 18 +++++ generator.py | 2 + libvirt-override-virConnect.py | 35 ++++++++++ libvirt-override.c | 153 +++++++++++++++++++++++++++++++++++++++++ sanitytest.py | 2 + 5 files changed, 210 insertions(+) -- 2.5.5

2 3

[libvirt] [PATCH v3 0/4] Some patches for native TLS encrypted chardev TCP support
by John Ferlan 16 Jun '16

16 Jun '16

v2: http://www.redhat.com/archives/libvir-list/2016-June/msg01044.html Patches 3 and 4 from original already pushed. Changes since v2: Patch 1: (the two bullets were ACK'd, but 3rd introduced new change so repost) * Change to using /etc/pki/qemu as the default directory name * Add description for dh-params.pem * Added default_tls_x509_verify Patch 2: * Change to using /etc/pki/qemu-chardev * Change charTCP/chartcp to chardev * Added chardev_tls_x509_verify Patch 3: (former patch 5) * Added processing for verify-peer and set in .args output Patch 4: (was ACK'd, but...) * Needed update due to new argument to qemuBuildTLSx509BackendProps John Ferlan (4): conf: Add new default TLS X.509 certificate default directory conf: Introduce chartcp_tls_x509_cert_dir qemu: Add support for TLS X.509 path to TCP chardev backend qemu: Add the ability to hotplug the TLS X.509 environment src/conf/domain_conf.h | 1 + src/qemu/libvirtd_qemu.aug | 11 ++- src/qemu/qemu.conf | 83 ++++++++++++---- src/qemu/qemu_command.c | 109 ++++++++++++++++++++- src/qemu/qemu_command.h | 7 ++ src/qemu/qemu_conf.c | 59 +++++++++-- src/qemu/qemu_conf.h | 7 ++ src/qemu/qemu_hotplug.c | 30 +++++- src/qemu/qemu_monitor_json.c | 9 ++ src/qemu/test_libvirtd_qemu.aug.in | 5 + .../qemuxml2argv-serial-tcp-tlsx509-chardev.args | 33 +++++++ .../qemuxml2argv-serial-tcp-tlsx509-chardev.xml | 41 ++++++++ tests/qemuxml2argvtest.c | 6 ++ .../qemuxml2xmlout-serial-tcp-tlsx509-chardev.xml | 50 ++++++++++ tests/qemuxml2xmltest.c | 1 + 15 files changed, 424 insertions(+), 28 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-serial-tcp-tlsx509-chardev.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-serial-tcp-tlsx509-chardev.xml create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-serial-tcp-tlsx509-chardev.xml -- 2.5.5

1 4

[libvirt] [PATCH 0/3] Tiny cleanups
by Martin Kletzander 16 Jun '16

16 Jun '16

John suggested I clean up some things, so here they are. 2 out of 3 are trivial, but the first patch would look really really ugly if I wquashed them together (feel free to try). Martin Kletzander (3): qemu: Unify automatic coredump filenames qemu: Follow coding style convention qemu: Remove useless block in processWatchdogEvent src/qemu/qemu_driver.c | 117 +++++++++++++++++++++++++------------------------ 1 file changed, 60 insertions(+), 57 deletions(-) -- 2.9.0

2 4