December 2009 - Devel - libvirt List Archives

[libvirt] [Patch v0.5] iSCSI Multi-IQN (Libvirt Support)
by Shyam Iyer 08 Jan '10

08 Jan '10

This patch set realizes the multi-IQN concept discussed in an earlier thread http://www.mail-archive.com/libvir-list@redhat.com/msg16706.html And here .. http://www.mail-archive.com/libvir-list@redhat.com/msg17499.html The patch realizes an XML schema like the one below and allows libvirt to read through it to create storage pools. These XMLs when created using a virtualization manager realize unique VM to storage LUN mappings through a single console and thus opening up possibilities for the following scenarios - * possibility of multiple IQNs for a single Guest * option for hypervisor's initiator to use these IQNs on behalf of the guest Change Log from v0.4: 1) Set default tab space to 4(Hopefully this is corrected this time ;) ) 2) Review comments from Dave Allan a) Use output of "iscsiadm -m iface" to search for existing iqn names in the iface files. 3) Create new unique iface file names for user provided iqn names if the iqns are not present in the existing iface files. Change Log from v0.3: 1) Set default tab space to 4 2) Use Case Description for Commit Log 3) Review comments from Dave Allan a) No initiator iqn in the xml would mean use of the default initiator iqn name b) Initiator iqn provided would mean a unique session to be created using the provided iqn name. c) Single iSCSI session per pool 4) Added syntax in doc/schemas/storagepool.rng There are no new errors introduced by this patch with "make check" and "make syntax-check" tests. Signed-off-by: Sudhir Bellad <sudhir_bellad(a)dell.com> Signed-off-by: Shyam Iyer <shyam_iyer(a)dell.com>

2 1

[libvirt] [PATCH] [RFC] xen domctl version 6
by Jim Fehlig 08 Jan '10

08 Jan '10

xen-unstable c/s 20685 changed the domctl interface, adding a field to xen_domctl_getdomaininfo structure. This additional field causes stack corruption in libvirt. xen-unstable c/s 20711 rightly bumped the domctl interface version so it is at least possible to handle the new field. The attached patch accounts for shr_pages field added to xen_domctl_getdomaininfo structure. I'm not thrilled about the changes to all the macros - suggestions for improvement welcomed. Tested with domctl version 5 and 6. Regards, Jim

2 3

[libvirt] Fwd: Re: [Qemu-devel] Re: qemu-kvm-0.12 bug?
by Thomas Treutner 07 Jan '10

07 Jan '10

Hi, there is a new parameter in qemu-0.12 required for virtio_balloon support. Is there an appropriate XML-entry in libvirt's domain config yet or any plans to introduce one? A quick googling for "libvirt balloon" didn't reveal any hints. Adam, how did you test the memstat reporting feature? kr, tom ---------- Forwarded Message ---------- Subject: Re: [Qemu-devel] Re: qemu-kvm-0.12 bug? Date: Tuesday 29 December 2009 From: Luiz Capitulino <lcapitulino(a)redhat.com> To: Thomas Treutner <thomas(a)scripty.at> On Tue, 29 Dec 2009 13:08:26 +0100 Thomas Treutner <thomas(a)scripty.at> wrote: > On Tuesday 29 December 2009 12:48:39 Luiz Capitulino wrote: > > On Tue, 29 Dec 2009 13:10:36 +0200 Avi Kivity <avi(a)redhat.com> wrote: > > > Hm, I get "The balloon device has not been activated by the guest" in > > > the qemu monitor. > > > > Is the "-balloon virtio" parameter passed on the command-line? > > # grep balloon /usr/local/var/log/libvirt/qemu/* | wc -l > 0 > > These logs include some history - so the parameter isn't used by libvirt even > when ballooning works? This parameter is new in qemu-0.12.

5 8

[libvirt] [PATCH 0/1] Don't update vol details after build
by David Allan 07 Jan '10

07 Jan '10

This patch fixes the bug reported at: https://bugzilla.redhat.com/show_bug.cgi?id=510450 See the patch for details. Dave

3 3

[libvirt] [PATCH] Disable building of static Python module.
by Diego Elio Pettenò 07 Jan '10

07 Jan '10

Python modules are loaded at runtime so the static version of it is not really needed, this avoids duplicating the build for the PIC and non-PIC cases. --- python/Makefile.am | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/python/Makefile.am b/python/Makefile.am index 04342b7..58c6729 100644 --- a/python/Makefile.am +++ b/python/Makefile.am @@ -39,7 +39,7 @@ libvirtmod_la_SOURCES = libvirt-override.c typewrappers.c libvirt.c libvirt.h # need extra flags here libvirtmod_la_CFLAGS = @WARN_PYTHON_CFLAGS@ -libvirtmod_la_LDFLAGS = -module -avoid-version -L$(top_builddir)/src/.libs \ +libvirtmod_la_LDFLAGS = -module -avoid-version -shared -L$(top_builddir)/src/.libs \ @CYGWIN_EXTRA_LDFLAGS@ libvirtmod_la_LIBADD = $(mylibs) \ @CYGWIN_EXTRA_LIBADD@ @CYGWIN_EXTRA_PYTHON_LIBADD@ -- 1.6.6.rc4

2 1

[libvirt] [PATCH] Fix parsing of 'info chardev' line endings
by Matthew Booth 07 Jan '10

07 Jan '10

This change makes the 'info chardev' parser ignore any trailing whitespace on a line. This fixes a specific problem handling a '\r\n' line ending. * src/qemu/qemu_monitor_text.c: Ignore trailing whitespace in 'info chardev' output. --- src/qemu/qemu_monitor_text.c | 26 +++++++++++++++++--------- 1 files changed, 17 insertions(+), 9 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 0cb9ea6..4fd8c4a 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -1622,15 +1622,26 @@ int qemuMonitorTextGetPtyPaths(qemuMonitorPtr mon, goto cleanup; } - char *pos = reply; /* The current start of searching */ - char *end = pos + strlen(reply); /* The end of the reply string */ + char *pos; /* The current start of searching */ + char *next = reply; /* The start of the next line */ char *eol; /* The character which ends the current line */ + char *end = reply + strlen(reply); /* The end of the reply string */ + + while (next) { + pos = next; - while (pos < end) { /* Split the output into lines */ eol = memchr(pos, '\n', end - pos); - if (eol == NULL) + if (eol == NULL) { eol = end; + next = NULL; + } else { + next = eol + 1; + } + + /* Ignore all whitespace immediately before eol */ + while (eol > pos && c_isspace(*(eol-1))) + eol -= 1; /* Look for 'filename=pty:' */ #define NEEDLE "filename=pty:" @@ -1638,13 +1649,13 @@ int qemuMonitorTextGetPtyPaths(qemuMonitorPtr mon, /* If it's not there we can ignore this line */ if (!needle) - goto next; + continue; /* id is everthing from the beginning of the line to the ':' * find ':' and turn it into a terminator */ char *colon = memchr(pos, ':', needle - pos); if (colon == NULL) - goto next; + continue; *colon = '\0'; char *id = pos; @@ -1664,9 +1675,6 @@ int qemuMonitorTextGetPtyPaths(qemuMonitorPtr mon, goto cleanup; } #undef NEEDLE - - next: - pos = eol + 1; } ret = 0; -- 1.6.5.2

3 2

[libvirt] [PATCH] vbox_tmpl.c: avoid NULL deref upon vboxDomainCreateXML failure
by Jim Meyering 06 Jan '10

06 Jan '10

Similar to others, >From 5dcb4d95a351cb6fa32ebaafb96756275e1079e2 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 16 Dec 2009 13:56:57 +0100 Subject: [PATCH] vbox_tmpl.c: avoid NULL deref upon vboxDomainCreateXML failure * src/vbox/vbox_tmpl.c (vboxDomainCreateXML): Don't call vboxDomainUndefine on a NULL "dom". --- src/vbox/vbox_tmpl.c | 16 +++++----------- 1 files changed, 5 insertions(+), 11 deletions(-) diff --git a/src/vbox/vbox_tmpl.c b/src/vbox/vbox_tmpl.c index d6b681c..ba70053 100644 --- a/src/vbox/vbox_tmpl.c +++ b/src/vbox/vbox_tmpl.c @@ -990,8 +990,6 @@ cleanup: static virDomainPtr vboxDomainCreateXML(virConnectPtr conn, const char *xml, unsigned int flags ATTRIBUTE_UNUSED) { - virDomainPtr dom = NULL; - /* VirtualBox currently doesn't have support for running * virtual machines without actually defining them and thus * for time being just define new machine and start it. @@ -1000,17 +998,13 @@ static virDomainPtr vboxDomainCreateXML(virConnectPtr conn, const char *xml, * change this behaviour to the expected one. */ - dom = vboxDomainDefineXML(conn, xml); - if (dom) { - if (vboxDomainCreate(dom) < 0) - goto cleanup; - } else { - goto cleanup; - } + virDomainPtr dom = vboxDomainDefineXML(conn, xml); + if (dom == NULL) + return NULL; - return dom; + if (0 < vboxDomainCreate(dom)) + return dom; -cleanup: vboxDomainUndefine(dom); return NULL; } -- 1.6.6.rc2.275.g51e2d

2 7

[libvirt] [PATCH] qemu_driver.c: avoid NULL dereference upon disk-op failure
by Jim Meyering 06 Jan '10

06 Jan '10

The local, "cgroup" is initialized to NULL, and may still have that value when the code below is reached. That would provoke a NULL-dereference in virCgroupDenyDevicePath. >From c1caed370a7b2eae2e964a6059b014530143075c Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 16 Dec 2009 14:15:50 +0100 Subject: [PATCH] qemu_driver.c: avoid NULL dereference upon disk-op failure * src/qemu/qemu_driver.c (qemudDomainAttachDevice): Call virCgroupDenyDevicePath only if cgroup is non-NULL. --- src/qemu/qemu_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 9ef6c35..81afecf 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -5485,7 +5485,7 @@ static int qemudDomainAttachDevice(virDomainPtr dom, virDomainDiskDeviceTypeToString(dev->data.disk->device)); /* Fallthrough */ } - if (ret != 0) { + if (ret != 0 && cgroup) { virCgroupDenyDevicePath(cgroup, dev->data.disk->src); } -- 1.6.6.rc2.275.g51e2d

2 2

[libvirt] [PATCH] openvz_conf.c: don't dereference NULL upon failure
by Jim Meyering 05 Jan '10

05 Jan '10

"dom" is set to NULL within the while loop: virDomainObjUnlock(dom); dom = NULL; If on a subsequent iteration something fails, we goto "cleanup" or "no_memory", both of which have us run this code: fclose(fp); virDomainObjUnref(dom); return -1; And the virDomainObjUnref function would dereference "dom". >From 3971ff17c7e9f1ddbc443d48b86fe6ba60a2d4a0 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 15 Dec 2009 16:16:57 +0100 Subject: [PATCH] openvz_conf.c: don't dereference NULL upon failure * src/openvz/openvz_conf.c (openvzLoadDomains): Avoid NULL deref of "dom". --- src/openvz/openvz_conf.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/openvz/openvz_conf.c b/src/openvz/openvz_conf.c index 7e9abbf..43bbaf2 100644 --- a/src/openvz/openvz_conf.c +++ b/src/openvz/openvz_conf.c @@ -535,7 +535,8 @@ int openvzLoadDomains(struct openvz_driver *driver) { cleanup: fclose(fp); - virDomainObjUnref(dom); + if (dom) + virDomainObjUnref(dom); return -1; } -- 1.6.6.rc2.275.g51e2d

2 2

[libvirt] [PATCH 0/3] Block assignment of PCI devices below non-ACS capable switch
by Jiri Denemark 05 Jan '10

05 Jan '10

Hi. This is a patchset for blocking assignment of PCI devices below non-ACS capable switch. In case user still wants to assign such device even though it might not be safe, she can specify permissive='yes' attribute for <hostdev>. Special thanks to Chris L. who created a standalone program the PCI check. This code is a port of that check to libvirt. Jiri Denemark (3): Tests for ACS in PCIe switches New 'permissive' attribute for hostdev Use pciDeviceIsAssignable in qemu driver docs/schemas/domain.rng | 8 +++ src/conf/domain_conf.c | 14 ++++- src/conf/domain_conf.h | 1 + src/libvirt_private.syms | 3 + src/qemu/qemu_driver.c | 9 +++- src/util/pci.c | 147 ++++++++++++++++++++++++++++++++++++++++++++++ src/util/pci.h | 7 ++ 7 files changed, 186 insertions(+), 3 deletions(-)

5 12