- Devel - libvirt List Archives

[Libvir] --without-*
by Mark Johnson 30 Aug '07

30 Aug '07

Hi all, Merging up to 0.3.2 on Solaris finally.. I had a couple of things on my list of hacked patches, and not being a configure expert I figured I would ask here... How hard it is to change the config logic to account for the various --without's. e.g. --without-qemu, o ./Makefile.in - SUBDIRS shouldn't include qemud o ./src/Makefile.in o am__objects_1 shouldn't include: o libvirt_la-bridge.lo o libvirt_la-iptables.lo o libvirt_la-qemu_driver.lo o libvirt_la-qemu_conf.lo o client sources, build rules, etc for the above. --without-test (Incuding srcs, rules, etc) o libvirt_la-test.lo, etc... --without-xen (Incuding srcs, rules, etc) o libvirt_la-xen_internal.lo o libvirt_la-xs_internal.lo o libvirt_la-xend_internal.lo --without-remote (Incuding srcs, rules, etc) o libvirt_la-remote_internal.lo o am_libvirt_la_OBJECTS shouldn't include am__objects_2 One other thing I was curious about... I noticed the remote code uses remote_protocol* out of the ./qemud dir.. I was just curious if I should expect that to move somewhere else soon, or if you were going to leave it there? Thanks, Mark.

2 2

[Libvir] [PATCH] Fix endless loop of VirBufferVSprintf()
by Masayuki Sunou 30 Aug '07

30 Aug '07

Hi VirBufferVSprintf() loops endless when it receives over 2000 bytes, because the return value of vsnprintf() is more than "size - 1". It is because the maximum of "size" is 1999 bytes. --> the maximum of "buf->size - buf->use" = 2000 (1000(argument of virBufferGrow()) + 1000(set in virBufferGrow())) So, virBufferEscapeString() has the same problem. This patch sets to virBufferGrow() the size of the data which VirBufferVSprintf()(virBufferEscapeString()) received when it is over 1000 bytes. Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou. ---------------------------------------------------------------------- Index: src/buf.c =================================================================== RCS file: /data/cvs/libvirt/src/buf.c,v retrieving revision 1.3 diff -u -p -r1.3 buf.c --- src/buf.c 9 Jul 2007 11:24:52 -0000 1.3 +++ src/buf.c 24 Aug 2007 05:56:37 -0000 @@ -159,7 +159,7 @@ virBufferContentAndFree (virBufferPtr bu int virBufferVSprintf(virBufferPtr buf, const char *format, ...) { - int size, count; + int size, count, grow_size; va_list locarg, argptr; if ((format == NULL) || (buf == NULL)) { @@ -172,7 +172,8 @@ virBufferVSprintf(virBufferPtr buf, cons locarg)) < 0) || (count >= size - 1)) { buf->content[buf->use] = 0; va_end(locarg); - if (virBufferGrow(buf, 1000) < 0) { + grow_size = (count > 1000) ? count : 1000; + if (virBufferGrow(buf, grow_size) < 0) { return (-1); } size = buf->size - buf->use - 1; @@ -198,7 +199,7 @@ virBufferVSprintf(virBufferPtr buf, cons int virBufferEscapeString(virBufferPtr buf, const char *format, const char *str) { - int size, count, len; + int size, count, len, grow_size; char *escaped, *out; const char *cur; @@ -248,7 +249,8 @@ virBufferEscapeString(virBufferPtr buf, while (((count = snprintf(&buf->content[buf->use], size, format, (char *)escaped)) < 0) || (count >= size - 1)) { buf->content[buf->use] = 0; - if (virBufferGrow(buf, 1000) < 0) { + grow_size = (count > 1000) ? count : 1000; + if (virBufferGrow(buf, grow_size) < 0) { free(escaped); return (-1); } ----------------------------------------------------------------------

2 2

[Libvir] ANNOUNCE: virt-top 0.3.2.5 released
by Richard W.M. Jones 29 Aug '07

29 Aug '07

I'm pleased to announce a new release of virt-top, which is a utility like 'top' that works for virtual machines. Many keys and command line options are the same as for ordinary top. Thanks to all the people who have been playing with it and sending me feedback. This version adds block device and network statistics (if supported by the hypervisor and libvirt). This version also fixes a bug where it would fail silently if unable to contact the hypervisor. Main page: http://et.redhat.com/~rjones/virt-top/ Source and binaries for several architectures available from: http://et.redhat.com/~rjones/virt-top/download.html If you have Fedora 7, this will probably work: yum --enablerepo=development install virt-top Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

1 0

[Libvir] [PATCH] TX/RX network stats are swapped
by Richard W.M. Jones 29 Aug '07

29 Aug '07

I noticed that the stats reported by virDomainInterfaceStats appeared to be swapped around. If the domain was downloading a lot of information, then tx_bytes would increase quickly (I'd expect rx_bytes to increase more if the domain was downloading). The problem is that xen_internal.c measures the TX/RX from the point of view of dom0. So that if dom0 is TRANSMITTING a lot of data up to the domain, then that domain is really RECEIVING it. Yes, I feel like an idiot ... Fix attached. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

2 1

[Libvir] [PATCH] Block device and network stats (version 2)
by Richard W.M. Jones 29 Aug '07

29 Aug '07

This patch adds block device and network stats. The main changes over the previous version are: * Remote support. * Change Xen network interface names to have the form "vif<domid>.<n>"[1]. Discussions about the previous version may be found starting here: https://www.redhat.com/archives/libvir-list/2007-August/thread.html#00064 I have left use of stdint.h / int64_t, since it wasn't clear to me what conclusion people had arrived at. I left the explicit structure size parameter to allow for future extensibility. Also attached is the output of: virsh -c xen+tls://amd/ domifstat fc6_1 vif2.0 Rich. [1] There is currently no way to get the interface name except to just know it -- that is the subject of a separate patch. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 libvirt: virConnectOpen (name=xen+tls://amd/) libvirt: virInitialize () libvirt: virInitialize () libvirt: virInitialize () libvirt: virInitialize () libvirt: virInitialize () libvirt: virInitialize () libvirt: virInitialize () libvirt: virInitialize () libvirt: virInitialize () libvirt: do_open: proceeding with name=xen+tls://amd/ libvirt: do_open: trying driver 0 (Test) ... libvirt: do_open: driver 0 Test returned DECLINED libvirt: do_open: trying driver 1 (QEMU) ... libvirt: do_open: driver 1 QEMU returned DECLINED libvirt: do_open: trying driver 2 (Xen) ... libvirt: do_open: driver 2 Xen returned DECLINED libvirt: do_open: trying driver 3 (remote) ... libvirt: do_open: driver 3 remote returned SUCCESS libvirt: do_open: network driver 0 Test returned DECLINED libvirt: do_open: network driver 1 QEMU returned DECLINED libvirt: do_open: network driver 2 remote returned SUCCESS libvirt: virDomainLookupByName (conn=0x130ee020, name=fc6_1) libvirt: virDomainInterfaceStats (domain=0x13129bd0, path=vif2.0, stats=0x7fff688e6e20, size=64) libvirt: virDomainFree (domain=0x13129bd0) libvirt: virConnectClose (conn=0x130ee020) vif2.0 rx_bytes 18591 vif2.0 rx_packets 113 vif2.0 rx_errs 0 vif2.0 rx_drop 0 vif2.0 tx_bytes 125319 vif2.0 tx_packets 962 vif2.0 tx_errs 0 vif2.0 tx_drop 0

5 17

[Libvir] [PATCH] Fix return value of virsh vncdisplay
by Masayuki Sunou 29 Aug '07

29 Aug '07

Hi Virsh vncdisplay always returns 1, because return value is not set in the case of normal end. ------------------------------------ # virsh vncdisplay test :0 # echo $? 1 ------------------------------------ This patch fixes it. Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou. ---------------------------------------------------------------------- Index: libvirt/src/virsh.c =================================================================== RCS file: /data/cvs/libvirt/src/virsh.c,v retrieving revision 1.100 diff -u -p -r1.100 virsh.c --- libvirt/src/virsh.c 21 Aug 2007 11:53:52 -0000 1.100 +++ libvirt/src/virsh.c 28 Aug 2007 06:33:03 -0000 @@ -2919,6 +2919,7 @@ cmdVNCDisplay(vshControl * ctl, vshCmd * } xmlXPathFreeObject(obj); obj = NULL; + ret = TRUE; cleanup: if (obj) ----------------------------------------------------------------------

2 1

[Libvir] [PATCH] Delete a harmful variable in xend_parse_sexp_desc()
by Masayuki Sunou 27 Aug '07

27 Aug '07

Hi Xend_parse_sexp_desc() of xend_internal.c declares a variable named "ret" to buffer SXP. However, it actually uses "buf.content" not "ret" and allocates memory to "buf" when size of "ret" became insufficient. --> virBufferAdd(&buf, ...);, virBufferVSprintf(&buf, ...) So freeing "ret" fails, because "ret" refers to an address unlike "buf.contents" As a result, a segmentation fault occurs. This patch fixes so that xend_parse_sexp_desc() uses "buf.contents" not "ret". Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou. -------------------------------------------------------------------------------- Index: src/xend_internal.c =================================================================== RCS file: /data/cvs/libvirt/src/xend_internal.c,v retrieving revision 1.135 diff -u -p -r1.135 xend_internal.c --- src/xend_internal.c 21 Aug 2007 09:31:12 -0000 1.135 +++ src/xend_internal.c 27 Aug 2007 07:47:03 -0000 @@ -1346,7 +1346,6 @@ xend_parse_sexp_desc_os(virConnectPtr xe static char * xend_parse_sexp_desc(virConnectPtr conn, struct sexpr *root, int xendConfigVersion) { - char *ret; struct sexpr *cur, *node; const char *tmp; char *tty; @@ -1362,10 +1361,9 @@ xend_parse_sexp_desc(virConnectPtr conn, /* ERROR */ return (NULL); } - ret = malloc(4000); - if (ret == NULL) + buf.content = malloc(4000); + if (buf.content == NULL) return (NULL); - buf.content = ret; buf.size = 4000; buf.use = 0; @@ -1762,11 +1760,11 @@ xend_parse_sexp_desc(virConnectPtr conn, virBufferAdd(&buf, "</domain>\n", 10); buf.content[buf.use] = 0; - return (ret); + return (buf.content); error: - if (ret != NULL) - free(ret); + if (buf.content != NULL) + free(buf.content); return (NULL); } --------------------------------------------------------------------------------

2 1

[Libvir] OpenVZ driver enhancements
by Shuveb Hussain 24 Aug '07

24 Aug '07

Hi, I'm glad to make available patches for the OpenVZ driver that provide the following features: * virDomainDefineXML - Defines an OpenVZ domain and does not start it. Takes XML description of the domain as input. * virDomainCreateLinux - Starts a domain based on the provided XML description. There is no way to start a domain in OpenVZ without defining it. So, it is defined anyway, as of now. :-( There may be a way to get around this. We are looking into it. As of now, treat this as define + start. * virDomainUndefine - removes domain from OpenVZ management. Since OpenVZ manages the domain's root file system, it is also lost. This behaviour is different from Xen. The XML Format for OpenVZ: -------------------------- <domain type='openvz'> <name>108</name> <uuid>8dea22b31d52d8f32516782e98ab8fa0</uuid> <container> <filesystem> <template>fedora-core-3-i386-minimal</template> <quota level = 'first'>123</quota> <quota level = 'second' uid = '500'>534</quota> </filesystem> <network> <ipaddress>192.168.1.108</ipaddress> <hostname>fedora-minimal</hostname> <gateway>192.168.1.1</gateway> <nameserver>192.168.1.1</nameserver> <netmask>255.255.255.0</netmask> </network> <profile>vps.basic</profile> </container> </domain> The name is the VPS "ID". The VPS ID is not temporary in OpenVZ as in Xen. The "<template>" tag in the "<filesystem>" section tells libvirt which OS template cache to use to create the VPS file system. Quota is not implemented as yet. First and second level quotas are intended to be supported. The "<profile>" tag must be a valid profile name from which VPS parameter are inherited. Other things, I guess are self explanatory. Other issues: ------------- * Moved some static declarations from the header to the .c files. * Fixed a small bug that cause libvirtd to crash on remote client exit. Patches are against CVS head. Regards, -- Shuveb Hussain Unix is very user friendly. It is just a little choosy about who its friends are http://www.binarykarma.com

2 1

[Libvir] Libvirt with Gentoo and Udev
by Stefan Bottelier 22 Aug '07

22 Aug '07

Hello, I have the follow problem with libvirt and Gentoo. I will be use virt-manager for my xen server on my Gentoo machine, zo i have libvirt afcors to use. virt-manager are in the portage of Gentoo. When i start libvirt i get the follow message. libvir: QEMU error : failed to set bridge STP to on Failed to autostart network 'default': failed to set bridge STP to on So i watch in my /var/log/messages for the error. I see the follow Aug 20 17:51:04 localhost udev-net.sh: /etc/init.d/net.vnet0: does not exist or is not executable So i create the vnet0 bridge interface in gentoo. Then i will restart libvirt. I see in /var/log/messages the follow error rc-scripts: We only hotplug for ethernet interfaces if its possibale to run udev and gentoo with libvirt? Or is zombie the can help me with this problem of if this working on Gentoo ? Thanks , Stefan P.s the option of /etc/init.d/net.vnet0 start --verbose will not be working, http://forums.gentoo.org/viewtopic-t-577325.html

1 0

[Libvir] Release of libvirt-0.3.2
by Daniel Veillard 21 Aug '07

21 Aug '07

No releases for nearly a month, it was time to push a new one, so 0.3.2 is available from : ftp://libvirt.org/libvirt/ There is more changes that I initially though. I will probably make a new release within a week to update things which were left out from 0.3.2, * New features: - KVM migration and save/restore (Jim Paris) - added API for migration (Richard Jones) - added APIs for block device and interface statistic (Richard Jones). * Documentation: - examples for XML network APIs - fix typo and schedinfo synopsis in man page (Atsushi SAKAI) - hypervisor support page update (Richard Jones). * Bug fixes: - remove a couple of leaks in QEmu/KVM backend(Daniel berrange) - fix GnuTLS 1.0 compatibility (Richard Jones) - --config/-f option mistake for libvirtd (Richard Jones) - remove leak in QEmu backend (Jim Paris) - fix some QEmu communication bugs (Jim Paris) - UUID lookup though proxy fix - setvcpus checking bugs (with Atsushi SAKAI) - int checking in virsh parameters (with Masayuki Sunou) - deny devices attach/detach for < Xen 3.0.4 (Masayuki Sunou) - XenStore query memory leak (Masayuki Sunou) - virsh schedinfo cleanup (Saori Fukuta). * Improvement: - virsh new ttyconsole command - networking API implementation for test driver (Daniel berrange) - qemu/kvm feature reporting of ACPI/APIC (David Lutterkort) - checking of QEmu architectures (Daniel berrange) - improve devices XML errors reporting (Masayuki Sunou) - speedup of domain queries on Xen (Daniel berrange) - augment XML dumps with interface devices names (Richard Jones) - internal API to query drivers for features (Richard Jones). * Cleanups: - Improve virNodeGetInfo implementation (Daniel berrange) - general UUID code cleanup (Daniel berrange) - fix API generator file selection. thanks to everybody who helped contribute to this new release with ideas, reports or patches ! Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

2 1

[Libvir] ANNOUNCE: ocaml-libvirt 0.3.2.0 (libvirt bindings for OCaml) released
by Richard W.M. Jones 21 Aug '07

21 Aug '07

The changes are: - virDomainMigrate supported - virDomainBlockStats, virDomainInterfaceStats supported - mlvirsh migrate - mlvirsh domblkstat and domifstat - now compiles with any libvirt >= 0.3.0 - uses autoconf to generate the ./configure script http://et.redhat.com/~rjones/ocaml-libvirt/ Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

1 0

[Libvir] [PATCH] Show Xen vif<domid>.<n> device name in domain XML
by Richard W.M. Jones 21 Aug '07

21 Aug '07

This patch changes the Xen domain XML so that it shows the target device (ie. vif<domid>.<n>), which you can then use to collect interface stats. For example: # virsh dumpxml fc6_1 ... <interface type='bridge'> <source bridge='xenbr0'/> <target dev='vif2.0'/> <mac address='00:16:3e:1a:2f:8c'/> <script path='vif-bridge'/> </interface> Note that this patch only adds to the XML, it doesn't change any XML that was being returned before. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 3

[Libvir] [PATCH] Cleanup xs_internal.c
by Masayuki Sunou 21 Aug '07

21 Aug '07

Hi Blanks and Tabs coexist in xs_internal.c. This patch replaces a tab by blanks in xs_internal.c. Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou.

4 6

[Libvir] [PATCH] Update hvsupport page
by Richard W.M. Jones 21 Aug '07

21 Aug '07

The three independent patches attached update the hvsupport page: (1) Document that KVM now supports save & restore. [Q: This is just kvm, or any qemu?] (2) Document the introduction of virDomainMigrate call. (3) Document the introduction of virDomainBlockStats and virDomainInterfaceStats calls. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 2

[Libvir] [PATCH] Check availability of driver calls
by Richard W.M. Jones 21 Aug '07

21 Aug '07

This is a repost of a patch I sent before which adds the ability to check for driver features. This is completely internal to libvirt and does not add any public APIs. The original discussion is here: https://www.redhat.com/archives/libvir-list/2007-July/thread.html#00437 and continues here: https://www.redhat.com/archives/libvir-list/2007-August/thread.html#00000 Daniel Veillard said: >> +typedef int >> + (*virDrvSupportsFeature) (virConnectPtr conn, virDrvFeature feature); > >I would rather use , int features) where you OR the features, allows >to know in one call if you get what you want or not. My objection to that is here: https://www.redhat.com/archives/libvir-list/2007-August/msg00000.html Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

4 6

[Libvir] [PATCH] virDomainMigrate version 5 (repost)
by Richard W.M. Jones 20 Aug '07

20 Aug '07

This is a repost of version 5 patch. Previous discussion of version 5: https://www.redhat.com/archives/libvir-list/2007-July/thread.html#00357 Previous discussion of version 4: https://www.redhat.com/archives/libvir-list/2007-July/thread.html#00304 (This is separate from the feature patch, but in reality if the feature patch goes in, then I'd modify this patch slightly to make it use feature detection). Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

2 1

[Libvir] [PATCH] check the file name for --log
by Atsushi SAKAI 20 Aug '07

20 Aug '07

Hi, This patch fixes the trouble if execute virsh with virsh --log --debug. In this case FILE "--debug" is generated and --debug is off. This patch treat as error return if it uses the hyphen "-" at the head of log file name. Signed-off-by: Atsushi SAKAI <sakaia(a)jp.fujitsu.com> Thanks Atsushi SAKAI

4 8

[Libvir] question about --quiet option
by Atsushi SAKAI 20 Aug '07

20 Aug '07

Hi, I have a question about virsh --quiet specification. It seems virsh --quiet option supress virsh interactive mode boot-message and vshPrintExtra only.(not includes vshPrint) This is supposed specification. Thanks Atsushi SAKAI

2 1

[Libvir] Re: [et-mgmt-tools] ANNOUNCE: virt-viewer release 0.0.1
by Daniel P. Berrange 20 Aug '07

20 Aug '07

On Fri, Aug 17, 2007 at 04:10:52PM +0200, Gerd Hoffmann wrote: > Hi, > > > To connect to the guest with ID 7 running under QEMU > > > > virt-viewer --connect qemu:///system 7 > > Can we *please* think about a standardized environment variable to set > the default for --connect and which all tools are looking at? Yes, this would be useful. > virsh has VIRSH_DEFAULT_CONNECT_URI for that. > > I'd like to have *one* variable used by *all* libvirt-based tools, > virt-viewer, virt-manager, virsh, ... We could simply reuse the virsh > variable, or we could pick some other name without virsh in there, > preferable also shorter and thus easier to remember. VIRT_URI maybe? We should probably apply the change to libvirt, so that if 'NULL' is used for the connect URI for virConnectOpen, and VIRT_CONNECT_URI is set, then NULL gets substituted with the value of the env variable. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

4 3

[Libvir] Set memory (decrease) from Virt Manager
by Bruce Evans 20 Aug '07

20 Aug '07

Hi, I am trying to lower the amount of memory for a Domain using the Virt manager. I am denied from doing this if I want to decrease the memory below 50MB, even though I can do it from virsh. There appears to be a low end limit of 50MB in the Virt Manager. If the current memory allocation is greater than 50MB, I can decrease the amount from Virt Manager, but can't go under 50MB. But if I use virsh, I can decrease the memory to, say, 13000KB and it works. The Virt Manger will show this new amount but will only enable the 'increase' icon in the 'Change Allocation option box. The 'decrease' icon is disabled. Is this a bug in Virt Manager? If not, why can't the memory be decreased below 50MB? Thanks, -Bruce Sun Microsystems Systems Virtualiztion

2 1

[Libvir] Stability of virsh / libvirt interfaces
by John Levon 17 Aug '07

17 Aug '07

This has probably been discussed before so apologies. What is the current status of the interfaces, especially libvirt and virsh? Are they guaranteed to be stable? Does this apply to every new interface as soon as a release is done? (I ask after noticing that virsh schedinfo can be used to *set* parameters!) thanks john

3 3

[Libvir] [PATCH] Fix some memory leaks of virsh schedinfo
by Saori Fukuta 17 Aug '07

17 Aug '07

Hi, There are some memory leaks at the virsh schedinfo command. So I fix them. (e.g.) When the virDomainSetSchedulerParameters failed, the allocated memory as params is not freed. # valgrind --leak-check=full virsh schedinfo PV_RH5_8 --weight 0 ==18879== Memcheck, a memory error detector. ==18879== Copyright (C) 2002-2006, and GNU GPL'd, by Julian Seward et al. *snip* ==18879== 92 bytes in 1 blocks are definitely lost in loss record 6 of 23 ==18879== at 0x40053C0: malloc (vg_replace_malloc.c:149) ==18879== by 0x80531D5: _vshMalloc (virsh.c:4201) ==18879== by 0x804BC44: cmdSchedinfo (virsh.c:1033) ==18879== by 0x80526CF: vshCommandRun (virsh.c:3839) ==18879== by 0x805452A: main (virsh.c:4741) ==18879== ==18879== LEAK SUMMARY: ==18879== definitely lost: 92 bytes in 1 blocks. ==18879== possibly lost: 0 bytes in 0 blocks. ==18879== still reachable: 14,234 bytes in 293 blocks. ==18879== suppressed: 0 bytes in 0 blocks. Signed-off-by: Saori Fukuta <fukuta.saori(a)jp.fujitsu.com> Thanks, Saori Fukuta

2 1

[Libvir] [PATCH] Fix some memory leaks in xs_internal.c
by Masayuki Sunou 17 Aug '07

17 Aug '07

Hi Some memory releases is missed in xs_internal.c. This patch fixes it. Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou. ---------------------------------------------------------------------- Index: src/xs_internal.c =================================================================== RCS file: /data/cvs/libvirt/src/xs_internal.c,v retrieving revision 1.47 diff -u -p -r1.47 xs_internal.c --- src/xs_internal.c 10 Aug 2007 18:25:15 -0000 1.47 +++ src/xs_internal.c 17 Aug 2007 05:19:00 -0000 @@ -578,6 +578,8 @@ xenStoreListDomains(virConnectPtr conn, #endif ids[ret++] = (int) id; } + if (idlist) + free(idlist); return(ret); } @@ -637,11 +639,10 @@ xenStoreLookupByName(virConnectPtr conn, path = xs_get_domain_path(priv->xshandle, (unsigned int) id); if (!found) - return(NULL); + goto done; ret = virGetDomain(conn, name, NULL); if (ret == NULL) { - if (path != NULL) free(path); goto done; } ret->id = id; @@ -651,6 +652,8 @@ done: free(xenddomain); if (idlist != NULL) free(idlist); + if (path != NULL) + free(path); return(ret); } ----------------------------------------------------------------------

2 1

[Libvir] [PATCH] Check the version of Xen in xenDaemonAttachDevice()
by Masayuki Sunou 16 Aug '07

16 Aug '07

Hi When virsh attach-disk/attach-interface is executed in < Xen 3.0.4, the OS type of domain is NULL in xenDaemonAttachDevice(). As a result, the command becomes segmentation fault. This patch fixes it by checking the version of Xen in xenDaemonAttachDevice(). Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou. ---------------------------------------------------------------------- Index: src/xend_internal.c =================================================================== RCS file: /data/cvs/libvirt/src/xend_internal.c,v retrieving revision 1.132 diff -u -p -r1.132 xend_internal.c --- src/xend_internal.c 9 Aug 2007 20:19:12 -0000 1.132 +++ src/xend_internal.c 16 Aug 2007 07:38:42 -0000 @@ -3091,6 +3091,9 @@ xenDaemonAttachDevice(virDomainPtr domai priv = (xenUnifiedPrivatePtr) domain->conn->privateData; + if (domain->id < 0 && priv->xendConfigVersion < 3) + return(-1); + str = virDomainGetOSType(domain); if (strcmp(str, "linux")) hvm = 1; ----------------------------------------------------------------------

3 3

[Libvir] [PATCH] Check a value in vshCommandOptInt()
by Masayuki Sunou 16 Aug '07

16 Aug '07

Hi When the option which should set a number is set a character which is not a number, the value is changed into 0 in libvirt. It causes the movement not expected. ex) # virsh vcpupin dom_name A 0,1 => VCPU0 is pinned without the command becoming the error. This patch fixes it by checking the value in vshCommandOptInt(). Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou. ---------------------------------------------------------------------- Index: src/virsh.c =================================================================== RCS file: /data/cvs/libvirt/src/virsh.c,v retrieving revision 1.95 diff -u -p -r1.95 virsh.c --- src/virsh.c 14 Aug 2007 07:07:57 -0000 1.95 +++ src/virsh.c 16 Aug 2007 07:13:46 -0000 @@ -3699,12 +3699,16 @@ static int vshCommandOptInt(vshCmd * cmd, const char *name, int *found) { vshCmdOpt *arg = vshCommandOpt(cmd, name); - int res = 0; + int res = 0, num_found = TRUE; + char *end_p=NULL; if (arg) - res = atoi(arg->data); + res = strtol(arg->data, &end_p, 10); + if (arg->data == end_p) { + num_found = FALSE; + } if (found) - *found = arg ? TRUE : FALSE; + *found = (arg && num_found) ? TRUE : FALSE; return res; } ----------------------------------------------------------------------

2 1