- Devel - libvirt List Archives

[libvirt] [PATCH] cmdPoolDiscoverSources: initialize earlier to avoid FP from clang
by Jim Meyering 01 Mar '10

01 Mar '10

This change hoists the initialization of srcSpec to the top, to keep clang from reporting a false-positive used-uninitialized problem. >From f487655124c887310be3bc82d7beae3cd47e7d21 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 1 Mar 2010 15:41:15 +0100 Subject: [PATCH] cmdPoolDiscoverSources: initialize earlier to avoid FP from clang * tools/virsh.c (cmdPoolDiscoverSources): Always initialize srcSpec. Otherwise, clang would report that srcSpec could be used uninitialized in the call to virConnectFindStoragePoolSources. --- tools/virsh.c | 7 +++---- 1 files changed, 3 insertions(+), 4 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 3b1011c..191e7ba 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -4431,17 +4431,16 @@ static const vshCmdOptDef opts_find_storage_pool_sources[] = { static int cmdPoolDiscoverSources(vshControl * ctl, const vshCmd * cmd ATTRIBUTE_UNUSED) { - char *type, *srcSpec, *srcSpecFile, *srcList; + char *type, *srcSpecFile, *srcList; + char *srcSpec = NULL; int found; type = vshCommandOptString(cmd, "type", &found); if (!found) return FALSE; srcSpecFile = vshCommandOptString(cmd, "srcSpec", &found); - if (!found) { + if (!found) srcSpecFile = NULL; - srcSpec = NULL; - } if (!vshConnectionUsability(ctl, ctl->conn, TRUE)) return FALSE; -- 1.7.0.1.414.g89213d

2 1

[libvirt] [PATCH] qemu: avoid null dereference on failed migration
by Eric Blake 01 Mar '10

01 Mar '10

From: Eric Blake <eblake(a)redhat.com> * src/qemu/qemu_monitor_text.c (qemuMonitorTextGetMigrationStatus): Check for failed strchr, to silence a coverity warning. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- src/qemu/qemu_monitor_text.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 62ffcc6..e7b4b1f 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -989,6 +989,11 @@ int qemuMonitorTextGetMigrationStatus(qemuMonitorPtr mon, if ((tmp = strstr(reply, MIGRATION_PREFIX)) != NULL) { tmp += strlen(MIGRATION_PREFIX); end = strchr(tmp, '\r'); + if (end == NULL) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, + _("unexpected migration status in %s"), reply); + goto cleanup; + } *end = '\0'; if ((*status = qemuMonitorMigrationStatusTypeFromString(tmp)) < 0) { -- 1.6.6.1

3 2

[libvirt] [PATCH] Do not search xenstore for disk device IDs
by Jim Fehlig 01 Mar '10

01 Mar '10

Disk devices can be referenced by name in Xen, e.g. when modifying their configuration or remvoving them. As such, don't search xenstore for a device ID corresponding to the disk device. Instead, search the disks contained in the domain definition and use the disk's target name if found. This approach allows removing a disk when domain is inactive. We obviously can't search xenstore when the domain is inactive. --- src/xen/xend_internal.c | 38 +++++++++++++++++++++++--------------- 1 files changed, 23 insertions(+), 15 deletions(-) diff --git a/src/xen/xend_internal.c b/src/xen/xend_internal.c index 9d95291..d0d32e2 100644 --- a/src/xen/xend_internal.c +++ b/src/xen/xend_internal.c @@ -93,6 +93,7 @@ xenDaemonFormatSxprOnePCI(virConnectPtr conn, static int virDomainXMLDevID(virDomainPtr domain, + virDomainDefPtr domDef, virDomainDeviceDefPtr dev, char *class, char *ref, @@ -4212,7 +4213,7 @@ xenDaemonAttachDeviceFlags(virDomainPtr domain, const char *xml, sexpr = virBufferContentAndReset(&buf); - if (virDomainXMLDevID(domain, dev, class, ref, sizeof(ref))) { + if (virDomainXMLDevID(domain, def, dev, class, ref, sizeof(ref))) { /* device doesn't exist, define it */ ret = xend_op(domain->conn, domain->name, "op", "device_create", "config", sexpr, NULL); @@ -4306,7 +4307,7 @@ xenDaemonDetachDeviceFlags(virDomainPtr domain, const char *xml, def, xml, VIR_DOMAIN_XML_INACTIVE))) goto cleanup; - if (virDomainXMLDevID(domain, dev, class, ref, sizeof(ref))) + if (virDomainXMLDevID(domain, def, dev, class, ref, sizeof(ref))) goto cleanup; if (dev->type == VIR_DOMAIN_DEVICE_HOSTDEV) { @@ -6048,6 +6049,7 @@ error: */ static int virDomainXMLDevID(virDomainPtr domain, + virDomainDefPtr domDef, virDomainDeviceDefPtr dev, char *class, char *ref, @@ -6056,27 +6058,33 @@ virDomainXMLDevID(virDomainPtr domain, xenUnifiedPrivatePtr priv = domain->conn->privateData; char *xref; char *tmp; + unsigned int i; + virDomainDiskDefPtr disk; if (dev->type == VIR_DOMAIN_DEVICE_DISK) { + if (dev->data.disk->dst == NULL) + return -1; if (dev->data.disk->driverName && STREQ(dev->data.disk->driverName, "tap")) strcpy(class, "tap"); else strcpy(class, "vbd"); - if (dev->data.disk->dst == NULL) - return -1; - xenUnifiedLock(priv); - xref = xenStoreDomainGetDiskID(domain->conn, domain->id, - dev->data.disk->dst); - xenUnifiedUnlock(priv); - if (xref == NULL) - return -1; - - tmp = virStrcpy(ref, xref, ref_len); - VIR_FREE(xref); - if (tmp == NULL) - return -1; + /* For disks, the device name can be used directly. + * If disk device exists in domain definintion, + * copy it to ref and return success. + */ + for (i = 0; i < domDef->ndisks; i++) { + disk = domDef->disks[i]; + if (STREQ(dev->data.disk->dst, disk->dst)) { + tmp = virStrcpy(ref, disk->dst, ref_len); + if (tmp == NULL) + return -1; + else + return 0; + } + } + return -1; } else if (dev->type == VIR_DOMAIN_DEVICE_NET) { char mac[30]; virDomainNetDefPtr def = dev->data.net; -- 1.6.0.2

3 4

[libvirt] running libvirt on Xen Cloud Platform 0.7.1
by Gadi Naor 01 Mar '10

01 Mar '10

Hi, I am trying to run libvirt daemon in Dom0 of Xen Cloud Platform 0.7.1. For that purpose I downloaded the libvirt source files 0.7.6 and trying to build the libvirt rpm with XEN driver using the XCP DDK. I am running into to many problems getting it done - so I was wondering whether this is supported or if anyone is using it in such way. Thank You, Gadi _________________________________________________________________ Hotmail: Free, trusted and rich email service. https://signup.live.com/signup.aspx?id=60969

1 0

[libvirt] and for LXC?
by Mihamina Rakotomandimby 01 Mar '10

01 Mar '10

Manao ahoana, Hello, Bonjour, Is there some plan/schedule to make libvirt LXC friendly? Or is it already? Misaotra, Thanks, Merci. -- Architecte Informatique chez Blueline/Gulfsat: Administration Systeme, Recherche & Developpement +261 34 29 155 34 / +261 33 11 207 36

3 2

[libvirt] (no subject)
by Daniel P. Berrange 01 Mar '10

01 Mar '10

Repost of http://www.redhat.com/archives/libvir-list/2010-February/msg00177.html The virsh output is now scaled to show KB/MB as needed, instead of always bytes. The public API gained extra status VIR_DOMAIN_JOB_COMPLETED = 3, /* Job has finished, but isn't cleaned up */ VIR_DOMAIN_JOB_FAILED = 4, /* Job hit error, but isn't cleaned up */ VIR_DOMAIN_JOB_CANCELLED = 6, /* Job was aborted, but isn't cleaned up */ to allow to more easily detecting end of a job, once we allow for truely async usage

2 21

[libvirt] [PATCH] [macvtap] interface type direct web docs addition
by Stefan Berger 01 Mar '10

01 Mar '10

This adds a description about the 'direct' type of interface recently added for macvtap device type support on the host.

3 2

[libvirt] [PATCH] Free the macvtap mode string
by Stefan Berger 01 Mar '10

01 Mar '10

I forgot to free the macvtap mode string. This fixes it. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

3 2

[libvirt] Hypervisor raw options
by Daniel Dehennin 01 Mar '10

01 Mar '10

Hello, I wonder if it's possible to add hypervisor pass-through options, I add some usage of kvm -startdate and need to use a script which call kvm with the desired option. Something like: <emulator>/usr/bin/kvm</emulator> <emulator-options> <option>-option1</option> <option>-option2</option> </emulator-options> The several <option/> will be passed to the <emulator/> command. Regards. -- Daniel Dehennin Récupérer ma clef GPG: gpg --keyserver pgp.mit.edu --recv-keys 0x6A2540D1

2 1

[libvirt] reload libvirtd.conf has a little problem.
by Osier Yang 01 Mar '10

01 Mar '10

hi When libvirtd is running, I changed "unix_sock_ro_perms=0777" into ""unix_sock_ro_perms=0755", and then reload the configuration with: kill -SIGHUP `pidof libvirtd` or "service libvirtd reload" the permission of /var/lib/libvirt/libvirt-sock-ro still is 0777, but not 0755. and after "service libvirtd restart", the permission is changed into 0755. think it's a little problem. libvirt version: libvirt-0.7.6-1.fc12.x86_64 Thanks and Regards osier

2 2

[libvirt] [PATCH] build: avoid warning about unused variables
by Jim Meyering 01 Mar '10

01 Mar '10

>From 8ddff3e6cdccc2b4289509c6941b43f2ddf8e643 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 14:19:33 +0100 Subject: [PATCH] build: avoid warning about unused variables * tools/virsh.c (cmdCPUBaseline): Remove declarations of unused variables, p and cur. --- tools/virsh.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 5fdbbe5..89eefcf 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -7048,12 +7048,11 @@ cmdCPUBaseline(vshControl *ctl, const vshCmd *cmd) int found; int ret = TRUE; char *buffer; - char *p; char *result = NULL; const char **list = NULL; unsigned int count = 0; xmlDocPtr doc = NULL; - xmlNodePtr node_list, cur; + xmlNodePtr node_list; xmlXPathContextPtr ctxt = NULL; xmlSaveCtxtPtr sctxt = NULL; xmlBufferPtr buf = NULL; -- 1.7.0.401.g84adb

2 1

[libvirt] [PATCH] libvirtd: do not ignore failure to set group ID in privileged mode
by Jim Meyering 26 Feb '10

26 Feb '10

Coverity spotted these unchecked setgid calls: >From 7eddaf457d54974859a8c22cced6f43e2d267746 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 26 Feb 2010 10:42:14 +0100 Subject: [PATCH] libvirtd: do not ignore failure to set group ID in privileged mode * daemon/libvirtd.c (qemudListenUnix): Diagnose and fail upon failure to set or restore group-ID. --- daemon/libvirtd.c | 12 ++++++++---- 1 files changed, 8 insertions(+), 4 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index cee6f13..9bdbecb 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -560,8 +560,10 @@ static int qemudListenUnix(struct qemud_server *server, oldgrp = getgid(); oldmask = umask(readonly ? ~unix_sock_ro_mask : ~unix_sock_rw_mask); - if (server->privileged) - setgid(unix_sock_gid); + if (server->privileged && setgid(unix_sock_gid)) { + VIR_ERROR(_("Failed to set group ID to %d"), unix_sock_gid); + goto cleanup; + } if (bind(sock->fd, (struct sockaddr *)&addr, sizeof(addr)) < 0) { VIR_ERROR(_("Failed to bind socket to '%s': %s"), @@ -569,8 +571,10 @@ static int qemudListenUnix(struct qemud_server *server, goto cleanup; } umask(oldmask); - if (server->privileged) - setgid(oldgrp); + if (server->privileged && setgid(oldgrp)) { + VIR_ERROR(_("Failed to restore group ID to %d"), oldgrp); + goto cleanup; + } if (listen(sock->fd, 30) < 0) { VIR_ERROR(_("Failed to listen for connections on '%s': %s"), -- 1.7.0.442.g55ad1

2 2

[libvirt] [PATCH] build: avoid warning about return-with-value in void function
by Jim Meyering 26 Feb '10

26 Feb '10

I saw this: virsh.c: In function 'virshReportError': virsh.c:363: warning: 'return' with a value, in function returning void >From fc83f08ac204cec5fba474a23bd19386d1a415d7 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 26 Feb 2010 11:50:34 +0100 Subject: [PATCH] build: avoid warning about return-with-value in void function * tools/virsh.c: Just "return;", not "return NULL;". --- tools/virsh.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index f7fd7d4..3b1011c 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -360,7 +360,7 @@ virshReportError(vshControl *ctl) * no error was ever raised, so just ignore */ last_error = virSaveLastError(); if (!last_error || last_error->code == VIR_ERR_OK) - return NULL; + return; } if (last_error->code == VIR_ERR_OK) { -- 1.7.0.442.g55ad1

2 1

[libvirt] build: update gnulib submodule to latest
by Jim Meyering 26 Feb '10

26 Feb '10

Thanks, Eric. >From bd3ff6b34892fe75cfcc67f1aeab43e8a273c1b6 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 26 Feb 2010 17:14:01 +0100 Subject: [PATCH] build: update gnulib submodule to latest * .gnulib: Update to latest. Commit 89bdf84bcd9c6032e37 inadvertently rewound the .gnulib submodule by 51 commits. This corrects it. Spotted by Eric Blake. --- .gnulib | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/.gnulib b/.gnulib index 11fbc57..80cd995 160000 --- a/.gnulib +++ b/.gnulib @@ -1 +1 @@ -Subproject commit 11fbc57405a118e6ec9a3ebc19bbf5ececdae4d6 +Subproject commit 80cd995cdcbf4b9ded895a43621a11f11806ad8d -- 1.7.0.442.g55ad1

2 1

[libvirt] [PATCH] Only build virDomainObjFormat if not building proxy.
by Chris Lalancette 26 Feb '10

26 Feb '10

While building under RHEL-5, I got a compile warning because virDomainObjFormat was defined but not used. That came about because in RHEL-5 we build with "#define PROXY", and virDomainObjFormat is only used with !PROXY. Move the define. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/conf/domain_conf.c | 5 ++--- 1 files changed, 2 insertions(+), 3 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index b6d2c61..1d7f458 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -5668,6 +5668,8 @@ char *virDomainDefFormat(virDomainDefPtr def, return NULL; } +#ifndef PROXY + static char *virDomainObjFormat(virCapsPtr caps, virDomainObjPtr obj, int flags) @@ -5703,9 +5705,6 @@ error: return NULL; } - -#ifndef PROXY - int virDomainSaveXML(const char *configDir, virDomainDefPtr def, const char *xml) -- 1.6.6.1

3 3

[libvirt] Odd implementation of libvirt qemudDomainSetMaxMemory
by Chris Lalancette 26 Feb '10

26 Feb '10

Hey Cole, In commit 09a33fd8a9cc85e5e1064168472bff7aecff15fb, you made it so that qemudDomainSetMaxMemory() was hotplug only. However, with the changes you pushed, this function no longer makes a lot of sense. It looks up the domain, checks to make sure it is active, and then if it is active, it fails. Assuming that qemu doesn't support changing the maximum amount of memory on the fly, wouldn't it be better to just remove this function entirely and let the higher libvirt layers report "function not supported"? -- Chris Lalancette

2 1

[libvirt] [PATCH] openvzGetVEID: don't leak (memory + file descriptor)
by Jim Meyering 25 Feb '10

25 Feb '10

Coverity spotted the leaks. Here's a proposed fix. I didn't bother with a separate diagnostic for the unlikely event that we read a negative number from the pipe. Seems far-fetched enough not to bother, but it's easy to add, if anyone cares. >From f3439c7eae46681eacf5b469a6f0e22cb8fec1b4 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 19:24:50 +0100 Subject: [PATCH] openvzGetVEID: don't leak (memory + file descriptor) * src/openvz/openvz_conf.c (openvzGetVEID): Always call fclose. Diagnose parse failure also when vzlist output is empty. If somehow we read a -1, diagnose that (albeit as a parse failure). --- src/openvz/openvz_conf.c | 19 +++++++------------ 1 files changed, 7 insertions(+), 12 deletions(-) diff --git a/src/openvz/openvz_conf.c b/src/openvz/openvz_conf.c index ce0c4d3..3713a45 100644 --- a/src/openvz/openvz_conf.c +++ b/src/openvz/openvz_conf.c @@ -964,6 +964,7 @@ int openvzGetVEID(const char *name) { char *cmd; int veid; FILE *fp; + bool ok; if (virAsprintf(&cmd, "%s %s -ovpsid -H", VZLIST, name) < 0) { openvzError(NULL, VIR_ERR_INTERNAL_ERROR, "%s", @@ -979,18 +980,12 @@ int openvzGetVEID(const char *name) { return -1; } - if (fscanf(fp, "%d\n", &veid ) != 1) { - if (feof(fp)) - return -1; - - openvzError(NULL, VIR_ERR_INTERNAL_ERROR, - "%s", _("Failed to parse vzlist output")); - goto cleanup; - } - - return veid; - - cleanup: + ok = fscanf(fp, "%d\n", &veid ) == 1; fclose(fp); + if (ok && veid >= 0) + return veid; + + openvzError(NULL, VIR_ERR_INTERNAL_ERROR, + "%s", _("Failed to parse vzlist output")); return -1; } -- 1.7.0.401.g84adb

2 3

[libvirt] avoid "make rpm" failure in doc/
by Jim Meyering 25 Feb '10

25 Feb '10

"make rpm" fails like this: Rebuilding the HTML pages from the XML API make[3]: *** No rule to make target `html/libvirt-libvirt.html', needed by `distdir'. Stop. make[3]: *** Waiting for unfinished jobs.... Validating the resulting XHTML pages These two patches fix that: (though "make rpm" still fails for me: extracting debug info from /j/rpmbuild/BUILDROOT/libvirt-0.7.6-1.fc12.x86_6h.. *** ERROR: No build ID note found in /j/rpmbuild/BUILDROOT/libvirt-0.7.6-1.h.. The first (1/2) change is not officially required to solve this problem, but it overlapped, and does fix a non-distdir build failure. >From ace4dca5dfbc022b057ce1d9f3d2770ad3a34716 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 22:51:47 +0100 Subject: [PATCH 1/2] build: teach apibuild.py to work in a non-srcdir build * docs/Makefile.am (libvirt-api.xml libvirt-refs.xml): Generalize apibuild.py to work in a non-srcdir build. Pass "srcdir" to it. * docs/apibuild.py (rebuild): Honor the $srcdir envvar. --- docs/Makefile.am | 2 +- docs/apibuild.py | 10 +++++++--- 2 files changed, 8 insertions(+), 4 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index e47b668..dbbd1ef 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -132,7 +132,7 @@ devhelp/index.html devhelp/libvirt.devhelp: libvirt-api.xml $(devhelpxsl) libvirt-api.xml libvirt-refs.xml: apibuild.py \ $(srcdir)/../include/libvirt/*.h \ $(srcdir)/../src/libvirt.c $(srcdir)/../src/util/virterror.c - -$(srcdir)/apibuild.py + -srcdir=$(srcdir) $(srcdir)/apibuild.py clean-local: rm -f *~ *.bak *.hierarchy *.signals *-unused.txt diff --git a/docs/apibuild.py b/docs/apibuild.py index e233c8b..0ab5db2 100755 --- a/docs/apibuild.py +++ b/docs/apibuild.py @@ -2102,10 +2102,14 @@ class docBuilder: def rebuild(): builder = None - if glob.glob("../src/libvirt.c") != [] : + srcdir = os.environ["srcdir"] + if glob.glob(srcdir + "/../src/libvirt.c") != [] : print "Rebuilding API description for libvirt" - builder = docBuilder("libvirt", ["../src", "../src/util", "../include/libvirt"], - []) + builder = docBuilder("libvirt", + [srcdir + "/../src", + srcdir + "/../src/util", + srcdir + "/../include/libvirt"], + []) elif glob.glob("src/libvirt.c") != [] : print "Rebuilding API description for libvir" builder = docBuilder("libvirt", ["src", "src/util", "include/libvirt"], -- 1.7.0.401.g84adb >From a79fb1c491197a91094cd2345c3d7bbba40da2fd Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 10:35:20 +0100 Subject: [PATCH 2/2] build: avoid "make rpm" failure in docs/ Add missing rule to build html/libvirt-libvirt.html. Use a GNU Make pattern rule to avoid running apibuild.py once for each out-of-date target, in a parallel build. * docs/Makefile.am --- docs/Makefile.am | 15 +++++++++------ 1 files changed, 9 insertions(+), 6 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index dbbd1ef..edf6fc8 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -123,17 +123,20 @@ html/index.html: libvirt-api.xml newapi.xsl page.xsl sitemap.html.in $(XMLLINT) --nonet --valid --noout html/*.html ; \ else echo "missing XHTML1 DTD" ; fi ; fi ); -devhelp/index.html devhelp/libvirt.devhelp: libvirt-api.xml $(devhelpxsl) +$(devhelphtml): libvirt-api.xml $(devhelpxsl) -@(echo Rebuilding devhelp files) -@(if [ -x $(XSLTPROC) ] ; then \ - $(XSLTPROC) --nonet -o devhelp/libvirt.devhelp $(top_srcdir)/docs/devhelp/devhelp.xsl $(top_srcdir)/docs/libvirt-api.xml ; fi ); + $(XSLTPROC) --nonet -o devhelp/libvirt.devhelp \ + $(top_srcdir)/docs/devhelp/devhelp.xsl libvirt-api.xml ; fi ); - -libvirt-api.xml libvirt-refs.xml: apibuild.py \ - $(srcdir)/../include/libvirt/*.h \ - $(srcdir)/../src/libvirt.c $(srcdir)/../src/util/virterror.c +html/%-%.html html/%-virterror.html %-api.xml %-refs.xml: $(srcdir)/apibuild.py -srcdir=$(srcdir) $(srcdir)/apibuild.py +html/%-%.html html/%-virterror.html %-api.xml %-refs.xml: \ + $(srcdir)/../include/%/*.h \ + $(srcdir)/../src/%.c \ + $(srcdir)/../src/util/virterror.c + clean-local: rm -f *~ *.bak *.hierarchy *.signals *-unused.txt -- 1.7.0.401.g84adb

2 3

[libvirt] [PATCH] build: avoid non-srcdir "make distcheck" failures (CLEANFILES)
by Jim Meyering 25 Feb '10

25 Feb '10

A quick review won't hurt, but this doesn't really need one. >From 9ac893a1342b31ef7df6fc57c03767c49ef98d5b Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 09:28:51 +0100 Subject: [PATCH 1/4] build: avoid non-srcdir "make distcheck" failures (CLEANFILES) * docs/Makefile.am (MAINTAINERCLEANFILES): Use this variable for generated-and-distributed files, not "CLEANFILES". Besides, "make clean" and "make distclean" should not delete distributed files. --- docs/Makefile.am | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index 07ea9bf..dbbd1ef 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -17,7 +17,7 @@ apipng = \ html/home.png \ html/right.png -devhelphtml = \ +devhelphtml = \ devhelp/libvirt.devhelp \ devhelp/index.html \ devhelp/general.html \ @@ -79,7 +79,7 @@ EXTRA_DIST= \ sitemap.html.in \ ChangeLog.awk -CLEANFILES = $(dot_html) $(apihtml) $(devhelphtml) +MAINTAINERCLEANFILES = $(dot_html) $(apihtml) $(devhelphtml) all: web -- 1.7.0.401.g84adb

2 1

[libvirt] [PATCH] qemu: Report binary path if error parsing -help
by Cole Robinson 25 Feb '10

25 Feb '10

Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/qemu/qemu_conf.c | 14 ++++++++------ src/qemu/qemu_conf.h | 3 ++- tests/qemuhelptest.c | 3 ++- 3 files changed, 12 insertions(+), 8 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index e3da08a..1f8a9c3 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -1237,7 +1237,8 @@ static unsigned long long qemudComputeCmdFlags(const char *help, #define SKIP_BLANKS(p) do { while ((*(p) == ' ') || (*(p) == '\t')) (p)++; } while (0) -int qemudParseHelpStr(const char *help, +int qemudParseHelpStr(const char *qemu, + const char *help, unsigned long long *flags, unsigned int *version, unsigned int *is_kvm, @@ -1308,8 +1309,8 @@ fail: p = strndup(help, p - help); qemuReportError(VIR_ERR_INTERNAL_ERROR, - _("cannot parse QEMU version number in '%s'"), - p ? p : help); + _("cannot parse %s version number in '%s'"), + qemu, p ? p : help); VIR_FREE(p); @@ -1340,12 +1341,13 @@ int qemudExtractVersionInfo(const char *qemu, enum { MAX_HELP_OUTPUT_SIZE = 1024*64 }; int len = virFileReadLimFD(newstdout, MAX_HELP_OUTPUT_SIZE, &help); if (len < 0) { - virReportSystemError(errno, "%s", - _("Unable to read QEMU help output")); + virReportSystemError(errno, + _("Unable to read %s help output"), qemu); goto cleanup2; } - if (qemudParseHelpStr(help, &flags, &version, &is_kvm, &kvm_version) == -1) + if (qemudParseHelpStr(qemu, help, &flags, + &version, &is_kvm, &kvm_version) == -1) goto cleanup2; if (retversion) diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h index ae187dc..c8757c5 100644 --- a/src/qemu/qemu_conf.h +++ b/src/qemu/qemu_conf.h @@ -175,7 +175,8 @@ int qemudExtractVersionInfo (const char *qemu, unsigned int *version, unsigned long long *qemuCmdFlags); -int qemudParseHelpStr (const char *str, +int qemudParseHelpStr (const char *qemu, + const char *str, unsigned long long *qemuCmdFlags, unsigned int *version, unsigned int *is_kvm, diff --git a/tests/qemuhelptest.c b/tests/qemuhelptest.c index 0f2b509..ad355d7 100644 --- a/tests/qemuhelptest.c +++ b/tests/qemuhelptest.c @@ -49,7 +49,8 @@ static int testHelpStrParsing(const void *data) if (virtTestLoadFile(path, &help, MAX_HELP_OUTPUT_SIZE) < 0) return -1; - if (qemudParseHelpStr(help, &flags, &version, &is_kvm, &kvm_version) == -1) + if (qemudParseHelpStr("QEMU", help, &flags, + &version, &is_kvm, &kvm_version) == -1) return -1; if (flags != info->flags) { -- 1.6.5.2

2 2

[libvirt] [PATCH] remote: Improve daemon startup error reporting
by Cole Robinson 25 Feb '10

25 Feb '10

If I toggle enable_tcp in libvirtd.conf and add --listen in /etc/init.d/libvirtd, I get the unhelpful error: Starting libvirtd daemon: error: Unable to initialize network sockets. Running without --daemon provides much more useful info: sudo libvirtd --listen 11:29:26.117: error : remoteCheckCertFile:270 : Cannot access CA certificate '/etc/pki/CA/cacert.pem': No such file or directory The daemon architecture makes it difficult to report this useful info if daemonized, so point users to /var/log/messages and dropping the --daemon flag if they want more info. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- daemon/libvirtd.c | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index fefa7aa..cee6f13 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -474,7 +474,10 @@ static int daemonForkIntoBackground(void) { goto again; if (ret == 1 && status != 0) { - fprintf(stderr, "error: %s\n", virDaemonErrTypeToString(status)); + fprintf(stderr, + "error: %s. Check /var/log/messages or run without " + "--daemon for more info.\n", + virDaemonErrTypeToString(status)); } _exit(ret == 1 && status == 0 ? 0 : 1); } -- 1.6.5.2

2 1

[libvirt] [PATCH] virsh: Show errors reported by nonAPI functions
by Cole Robinson 25 Feb '10

25 Feb '10

Only API calls trigger the error callback, which is required for proper virsh error reporting. Since we use non API functions from util/, make sure we properly report these errors. Fixes lack of error message from 'virsh create idontexit.xml' Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- tools/virsh.c | 12 ++++++++++-- 1 files changed, 10 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 5fdbbe5..74e3fad 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -352,8 +352,16 @@ virshErrorHandler(void *unused ATTRIBUTE_UNUSED, virErrorPtr error) static void virshReportError(vshControl *ctl) { - if (last_error == NULL) - return; + if (last_error == NULL) { + /* Calling directly into libvirt util functions won't trigger the + * error callback (which sets last_error), so check it ourselves. + * + * If the returned error has CODE_OK, this most likely means that + * no error was ever raised, so just ignore */ + last_error = virSaveLastError(); + if (!last_error || last_error->code == VIR_ERR_OK) + return NULL; + } if (last_error->code == VIR_ERR_OK) { vshError(ctl, "%s", _("unknown error")); -- 1.6.5.2

2 1

[libvirt] [PATCH] Use standard spacing for user/pass prompt
by Cole Robinson 24 Feb '10

24 Feb '10

Kind of minor, but it annoys me that the default auth callback doesn't put a space between the prompt and the input, like a typical terminal, ssh, etc. This patch changes the current prompt: Please enter your authentication name:myuser to Please enter your authentication name: myuser Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/libvirt.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index be17668..9d50c92 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -159,7 +159,7 @@ static int virConnectAuthCallbackDefault(virConnectCredentialPtr cred, case VIR_CRED_AUTHNAME: case VIR_CRED_ECHOPROMPT: case VIR_CRED_REALM: - if (printf("%s:", cred[i].prompt) < 0) + if (printf("%s: ", cred[i].prompt) < 0) return -1; if (fflush(stdout) != 0) return -1; @@ -178,7 +178,7 @@ static int virConnectAuthCallbackDefault(virConnectCredentialPtr cred, case VIR_CRED_PASSPHRASE: case VIR_CRED_NOECHOPROMPT: - if (printf("%s:", cred[i].prompt) < 0) + if (printf("%s: ", cred[i].prompt) < 0) return -1; if (fflush(stdout) != 0) return -1; -- 1.6.5.2

2 1

[libvirt] the price we pay for using gettext-0.14
by Jim Meyering 24 Feb '10

24 Feb '10

Make distcheck currently fails like this: make[2]: Entering directory `/t/libvirt-0.7.6/_build/po' /bin/sh @MKINSTALLDIRS@ /t/libvirt-0.7.6/_inst/share /bin/sh: @MKINSTALLDIRS@: No such file or directory make[2]: *** [install-data-yes] Error 127 That is due to our use of a version of gettext that is way too old: $ grep GETT configure.ac AM_GNU_GETTEXT_VERSION([0.14.1]) The work-around is to define (via AC_SUBST) the now-obsolete symbol, MKINSTALLDIRS, so that @MKINSTALLDIRS@ in the autopoint-provided Makefile.in.in (from the old version of gettext) ends up being substituted. As I recall, people insist on using this old version of gettext, because nothing newer is supplied by RHEL5.x. IMHO, that is RHEL5's problem, and should not be allowed to cause continuing problems in upstream projects. It is easy and safe to install a newer version of gettext, especially since any project using autopoint will automatically continue to get the same bit-rotting old files: a gettext installation includes an archive of all older versions of its files, to provide just this type of backwards compatibility. We had a similar problem in libguestfs, and it was not trivial to diagnose there, either, even though I'd seen it at least once or twice before then. Here's the work-around: [just realized this requires a comment. adding that now] >From 16b9418427101a495b7fdd6917f518f254d22720 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 10:53:44 +0100 Subject: [PATCH] build: ensure that MKINSTALLDIRS is AC_SUBST-defined since we're using gettext-0.14.1, which uses that now-obsolete automake symbol. Otherwise, make distcheck would fails like this: make[2]: Entering directory `/t/libvirt-0.7.6/_build/po' /bin/sh @MKINSTALLDIRS@ /t/libvirt-0.7.6/_inst/share /bin/sh: @MKINSTALLDIRS@: No such file or directory make[2]: *** [install-data-yes] Error 127 * configure.ac (MKINSTALLDIRS): Define. --- configure.ac | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/configure.ac b/configure.ac index 117cb20..2f79970 100644 --- a/configure.ac +++ b/configure.ac @@ -49,6 +49,8 @@ AC_PROG_CPP AC_OBJEXT +AC_SUBST([MKINSTALLDIRS], [$ac_aux_dir/mkinstalldirs]) + dnl gl_INIT uses m4_foreach_w, yet that is not defined in autoconf-2.59. dnl In order to accommodate developers with such old tools, here's a dnl replacement definition. -- 1.7.0.367.g566c3

3 4

[libvirt] [PATCH] maint: relax git minimum version
by Eric Blake 24 Feb '10

24 Feb '10

From: Eric Blake <eblake(a)redhat.com> Requiring git 1.6.4, just for the optional GNULIB_SRCDIR support, was too harsh. Resynchronize from gnulib. * .gnulib: Import from latest gnulib. * bootstrap: Re-synchronize from .gnulib/build-aux. * bootstrap.conf: Drop git to 1.5.5. * README-hacking: Document use of GNULIB_SRCDIR. --- Chris is currently testing that RHEL 5 with git 1.5.5 can actually grok this patch. We don't have a README-prereq (but README-hacking was borrowed from coreutils, which does). Should we add one? The README-prereq in coreutils is not quite relevant for a straight-across copy. .gnulib | 2 +- README-hacking | 9 +++++++-- bootstrap | 41 +++++++++++++++++++++++++++++++++++------ bootstrap.conf | 2 +- 4 files changed, 44 insertions(+), 10 deletions(-) diff --git a/.gnulib b/.gnulib index 11fbc57..108cfb8 160000 --- a/.gnulib +++ b/.gnulib @@ -1 +1 @@ -Subproject commit 11fbc57405a118e6ec9a3ebc19bbf5ececdae4d6 +Subproject commit 108cfb803174abf17eb101e9999b9d1a642ce1ba diff --git a/README-hacking b/README-hacking index 4105a3e..80b022c 100644 --- a/README-hacking +++ b/README-hacking @@ -12,8 +12,7 @@ requirements on anyone wishing to build from the just-checked-out sources. Note the requirements to build the released archive are much less and are just the requirements of the standard ./configure && make procedure. Specific development tools and versions will be checked for and listed by -the bootstrap script. See README-prereq for specific notes on obtaining -these prerequisite tools. +the bootstrap script. Valgrind <http://valgrind.org/> is also highly recommended, if Valgrind supports your architecture. See also README-valgrind. @@ -28,6 +27,12 @@ You can get a copy of the source repository like this: $ git clone git://libvirt.org/libvirt $ cd libvirt +As an optional step, if you already have a copy of the gnulib git +repository on your hard drive, then you can use it as a reference to +reduce download time and disk space requirements: + + $ export GNULIB_SRCDIR=/path/to/gnulib + The next step is to get all required pieces from gnulib, to run autoreconf, and to invoke ./configure: diff --git a/bootstrap b/bootstrap index be96444..365a3d9 100755 --- a/bootstrap +++ b/bootstrap @@ -1,4 +1,6 @@ #! /bin/sh +# Print a version string. +scriptversion=2010-02-24.17; # UTC # Bootstrap this package from checked-out sources. @@ -17,7 +19,15 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. -# Written by Paul Eggert. +# Originally written by Paul Eggert. The canonical version of this +# script is maintained as build-aux/bootstrap in gnulib, however, to +# be useful to your project, you should place a copy of it under +# version control in the top-level directory of your project. The +# intent is that all customization can be done with a bootstrap.conf +# file also maintained in your version control; gnulib comes with a +# template build-aux/bootstrap.conf to get you started. + +# Please report bugs or propose patches to bug-gnulib(a)gnu.org. nl=' ' @@ -161,7 +171,6 @@ vc_ignore=auto # die otherwise. find_tool () { - # Find sha1sum, named gsha1sum on MacPorts. find_tool_envvar=$1 shift find_tool_names=$@ @@ -424,12 +433,28 @@ case ${GNULIB_SRCDIR--} in GNULIB_SRCDIR=$gnulib_path ;; *) - # Use GNULIB_SRCDIR as a reference. Assumes git 1.6.4 or newer. + # Use GNULIB_SRCDIR as a reference. if test -d "$GNULIB_SRCDIR"/.git && \ git_modules_config submodule.gnulib.url >/dev/null; then echo "$0: getting gnulib files..." - git submodule update --init --reference "$GNULIB_SRCDIR" \ + if git submodule --help|grep reference > /dev/null; then + # Prefer the one-liner available in git 1.6.4 or newer. + git submodule update --init --reference "$GNULIB_SRCDIR" \ "$gnulib_path" || exit $? + else + # This fallback allows at least git 1.5.5. + if test -f "$gnulib_path"/gnulib-tool; then + # Since file already exists, assume submodule init already complete. + git submodule update || exit $? + else + # Older git can't clone into an empty directory. + rmdir "$gnulib_path" 2>/dev/null + git clone --reference "$GNULIB_SRCDIR" \ + "$(git_modules_config submodule.gnulib.url)" "$gnulib_path" \ + && git submodule init && git submodule update \ + || exit $? + fi + fi GNULIB_SRCDIR=$gnulib_path fi ;; @@ -819,6 +844,10 @@ bootstrap_epilogue echo "$0: done. Now you can run './configure'." -# Local Variables: -# indent-tabs-mode: nil +# Local variables: +# eval: (add-hook 'write-file-hooks 'time-stamp) +# time-stamp-start: "scriptversion=" +# time-stamp-format: "%:y-%02m-%02d.%02H" +# time-stamp-time-zone: "UTC" +# time-stamp-end: "; # UTC" # End: diff --git a/bootstrap.conf b/bootstrap.conf index d315e04..58ef2ab 100644 --- a/bootstrap.conf +++ b/bootstrap.conf @@ -124,7 +124,7 @@ autoconf 2.59 automake 1.9.6 autopoint - gettext - -git 1.6.4 +git 1.5.5 gzip - libtool - perl 5.5 -- 1.6.6.1

2 2