- Devel - libvirt List Archives

[Libvir] Strange warnings with gcc 4.1.2
by Richard W.M. Jones 07 Nov '07

07 Nov '07

This is with libvirt from CVS. They don't look like errors to me though... ------------ In function ‘open’, inlined from ‘readFile’ at virsh.c:850: /usr/include/bits/fcntl2.h:45: error: call to ‘__open_too_many_args’ declared with attribute error: open can be called either with 2 or 3 arguments, not more /usr/include/bits/fcntl2.h:51: error: call to ‘__open_missing_mode’ declared with attribute error: open with O_CREAT in second argument needs 3 arguments The code in question is: if ((fd = open(filename, O_RDONLY)) == -1) { Not a newly created file, so mode isn't needed AFAICS. ------------ In function ‘read’, inlined from ‘readFile’ at virsh.c:864: /usr/include/bits/unistd.h:43: warning: call to ‘__read_chk_warn’ declared with attribute warning: read called with bigger length than size of the destination buffer The code in question is: char buffer[1024]; char *new; int ret; if ((ret = read(fd, buffer, sizeof(buffer))) == 0) break; Also seems correct to me. ------------ Other warnings: virsh.c: In function ‘cmdSchedinfo’: virsh.c:1138: warning: ‘weight’ may be used uninitialized in this function virsh.c:1140: warning: ‘cap’ may be used uninitialized in this function For some reason gcc appears to have lots its ability to do the data flow analysis to find out that these warnings are bogus. In any case the fix is trivial: Index: src/virsh.c =================================================================== RCS file: /data/cvs/libvirt/src/virsh.c,v retrieving revision 1.105 diff -u -r1.105 virsh.c --- src/virsh.c 6 Nov 2007 09:41:18 -0000 1.105 +++ src/virsh.c 7 Nov 2007 12:37:07 -0000 @@ -1135,9 +1135,9 @@ int nr_inputparams = 0; int inputparams = 0; int weightfound = 0; - int weight; + int weight = 0; int capfound = 0; - int cap; + int cap = 0; char str_weight[] = "weight"; char str_cap[] = "cap"; int ret_val = FALSE; 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 3

[Libvir] PATCH: Avoid format string abuse (also avoids gcc warnings).
by Jim Meyering 07 Nov '07

07 Nov '07

This patch was prompted by warnings like this: util.c:56: warning: format not a string literal and no format arguments and they're legitimate. Imagine a format string contains "%%..." goes through the vnsprintf call, which reduces it to "%...". If the result string is then passed to __virRaiseError as the format string, then *boom*. Instead, use "%s" as the format, with the non-literal as the matching argument. Patch below. I searched the sources for %% and *did* find one potential problem: $ git-grep -B1 %% > k po/ms.po-msgid "too many drivers registered in %s" po/ms.po:msgstr "terlalu banyak spesifikasi penukaran %% pada suffiks" -- src/xend_internal.c- case '\n': src/xend_internal.c: snprintf(ptr, 4, "%%%02x", string[i]); since "% p" does happen to be a valid format string! So if someone using Malaysian messages provoked that particular diagnostic in a code path that takes it through __virRaiseError, bad things might happen. Big "if", of course :-) I didn't try. 2007-11-06 Jim Meyering <meyering(a)redhat.com> Avoid risk of format string abuse (also avoids gcc warnings). * src/util.c (ReportError): Use a literal "%s" format string. * src/remote_internal.c (server_error): Likewise. * src/qemu_conf.c (qemudReportError): Likewise.

3 3

[Libvir] PATCH: Avoid format string abuse (also avoids gcc warnings).
by Jim Meyering 06 Nov '07

06 Nov '07

[resending with the patch in-line, not as an attachment -- mailman removed my attached patch the first time! Bad mailman. Bad. ] This patch was prompted by warnings like this: util.c:56: warning: format not a string literal and no format arguments and they're legitimate. Imagine a format string contains "%%..." goes through the vnsprintf call, which reduces it to "%...". If the result string is then passed to __virRaiseError as the format string, then *boom*. Instead, use "%s" as the format, with the non-literal as the matching argument. Patch below. I searched the sources for %% and *did* find one potential problem: $ git-grep -B1 %% > k po/ms.po-msgid "too many drivers registered in %s" po/ms.po:msgstr "terlalu banyak spesifikasi penukaran %% pada suffiks" -- src/xend_internal.c- case '\n': src/xend_internal.c: snprintf(ptr, 4, "%%%02x", string[i]); since "% p" does happen to be a valid format string! So if someone using Malaysian messages provoked that particular diagnostic in a code path that takes it through __virRaiseError, bad things might happen. Big "if", of course :-) I didn't try. 2007-11-06 Jim Meyering <meyering(a)redhat.com> Avoid risk of format string abuse (also avoids gcc warnings). * src/util.c (ReportError): Use a literal "%s" format string. * src/remote_internal.c (server_error): Likewise. * src/qemu_conf.c (qemudReportError): Likewise. ----------------------------- >From 8e0cafec021ba3cbec55a4d691f19c14ed9e587f Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 6 Nov 2007 20:14:21 +0100 Subject: [PATCH] Avoid risk of format string abuse (also avoids gcc warnings). * src/util.c (ReportError): Use a literal "%s" format string. * src/remote_internal.c (server_error): Likewise. * src/qemu_conf.c (qemudReportError): Likewise. Signed-off-by: Jim Meyering <meyering(a)redhat.com> --- ChangeLog | 7 +++++++ src/qemu_conf.c | 2 +- src/remote_internal.c | 2 +- src/util.c | 2 +- 4 files changed, 10 insertions(+), 3 deletions(-) diff --git a/ChangeLog b/ChangeLog index 511ed6a..535173d 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,10 @@ +2007-11-06 Jim Meyering <meyering(a)redhat.com> + + Avoid risk of format string abuse (also avoids gcc warnings). + * src/util.c (ReportError): Use a literal "%s" format string. + * src/remote_internal.c (server_error): Likewise. + * src/qemu_conf.c (qemudReportError): Likewise. + Tue Nov 6 17:24:16 CET 2007 Daniel Veillard <veillard(a)redhat.com> * src/xs_internals.c: patch from Chris Lalancette, forgot to diff --git a/src/qemu_conf.c b/src/qemu_conf.c index 78f4699..3556a9a 100644 --- a/src/qemu_conf.c +++ b/src/qemu_conf.c @@ -64,7 +64,7 @@ void qemudReportError(virConnectPtr conn, errorMessage[0] = '\0'; } __virRaiseError(conn, dom, net, VIR_FROM_QEMU, code, VIR_ERR_ERROR, - NULL, NULL, NULL, -1, -1, errorMessage); + NULL, NULL, NULL, -1, -1, "%s", errorMessage); } int qemudLoadDriverConfig(struct qemud_driver *driver, diff --git a/src/remote_internal.c b/src/remote_internal.c index 3af326f..1420a88 100644 --- a/src/remote_internal.c +++ b/src/remote_internal.c @@ -3073,7 +3073,7 @@ server_error (virConnectPtr conn, remote_error *err) err->domain, err->code, err->level, str1, str2, str3, err->int1, err->int2, - message); + "%s", message); } /* get_nonnull_domain and get_nonnull_network turn an on-wire diff --git a/src/util.c b/src/util.c index eb57859..c964a63 100644 --- a/src/util.c +++ b/src/util.c @@ -53,7 +53,7 @@ ReportError(virConnectPtr conn, errorMessage[0] = '\0'; } __virRaiseError(conn, dom, net, VIR_FROM_NONE, code, VIR_ERR_ERROR, - NULL, NULL, NULL, -1, -1, errorMessage); + NULL, NULL, NULL, -1, -1, "%s", errorMessage); } static int virSetCloseExec(int fd) { -- 1.5.3.5.561.g140d

1 0

[Libvir] [PATCH]: Fix build warning in xs_internal.c
by Chris Lalancette 06 Nov '07

06 Nov '07

All, Attached is a trivial patch to fix a build warning in xs_internal.c; basically the xenUnifiedDriver structure doesn't have a domainDumpXML method, so we had one too many NULLs. Signed-off-by: Chris Lalancette <clalance(a)redhat.com>

2 1

[Libvir] [RFC] Linux-VServer support
by Daniel Hokka Zakrisson 06 Nov '07

06 Nov '07

Hi, This is an initial stab at adding Linux-VServer support to libvirt. There are still a couple of things missing, like scheduler support in the XML parsing, and proper network support. I've got a few questions though. virsh's schedinfo hardcodes the available options, should I be adding new ones there? Would better introspection from getSchedulerType make this a non-issue? How do the network domains interact and associate with the regular domains? -- Daniel Hokka Zakrisson

4 24

[Libvir] [PATCH] Fix failure of virsh attach-disk (BZ346281)
by Masayuki Sunou 06 Nov '07

06 Nov '07

Hi Virsh attach-disk fails when the disk is attached to domain. And error message displays that device_configure() failed. # virsh attach-disk PV_RH51RC_12 /dev/hda8 xvdb --driver phy libvir: Xen Daemon error : POST operation failed: (xend.err'device_configure() takes exactly 2 arguments (3 given)') --> BZ346281 : https://bugzilla.redhat.com/show_bug.cgi?id=346281 This problem occurs because search of device ID of disk doesn't operate correctly. This patch fixes it. And this fixes a little comment too. Signed-off-by: Masayuki Sunou <fj1826dm(a)aa.jp.fujitsu.com> Thanks, Masayuki Sunou.

2 1

[Libvir] [PATCH] Fix parameter validation of virsh schedinfo
by Masayuki Sunou 06 Nov '07

06 Nov '07

Hi User cannot judge whether virsh schedinfo successes or fails, because it does not become error when character string is set to an option. -------------------------------------------------- # virsh schedinfo 0 --weight aaa Scheduler : credit weight : 256 cap : 0 -------------------------------------------------- This patch fixes to become error when character string is set. 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 1.104 diff -u -p -r1.104 virsh.c --- src/virsh.c 30 Sep 2007 13:22:16 -0000 1.104 +++ src/virsh.c 5 Nov 2007 06:26:52 -0000 @@ -1149,11 +1149,25 @@ cmdSchedinfo(vshControl * ctl, vshCmd * return FALSE; /* Currently supports Xen Credit only */ - weight = vshCommandOptInt(cmd, "weight", &weightfound); - if (weightfound) nr_inputparams++; - - cap = vshCommandOptInt(cmd, "cap", &capfound); - if (capfound) nr_inputparams++; + if(vshCommandOptBool(cmd, "weight")) { + weight = vshCommandOptInt(cmd, "weight", &weightfound); + if (!weightfound) { + vshError(ctl, FALSE, _("Invalid value of weight")); + goto cleanup; + } else { + nr_inputparams++; + } + } + + if(vshCommandOptBool(cmd, "cap")) { + cap = vshCommandOptInt(cmd, "cap", &capfound); + if (!capfound) { + vshError(ctl, FALSE, _("Invalid value of cap")); + goto cleanup; + } else { + nr_inputparams++; + } + } params = vshMalloc(ctl, sizeof (virSchedParameter) * nr_inputparams); if (params == NULL) { --------------------------------------------------------------------------------

2 1

[Libvir] [Announce] libvirt-cim : An open-source CIM provider for libvirt
by Daniel Veillard 05 Nov '07

05 Nov '07

libvirt-cim : An open-source CIM provider for libvirt This is the announcement of a new open-source project called libvirt-cim based on libvirt and aiming at offering the complete functionality of libvirt via a CIM provider implementing the DMTF SVPC virtualization model and released under an LGPL licence. A CIM [1] provider is an implementation of a set of standardized interfaces (Common Information Model) whose goal are to provide well defined entry points allowing easier and interoperable management tools to be built. In the case of libvirt-cim, the goal is of course to export the SVPC virtualization model, which then can be used to manage storage, hosts and domains remotely. Currently CIM based management targets mostly enterprise deployments but has the potential of solving the more generic problem of remote control interfaces which affects nearly all virtualization use cases and many other aspects of the computing environment. The current state of the project is based on an initial implementation provided by IBM, and released for the libvirt project under an LGPL licence. It currently targets only Xen virtualization but the goal is to expand support to the other mechanisms supported by libvirt, like KVM. It is made of a library libcmpiutil of library functions which are not specific to libvirt and useful for CMPI, the programming interfaces used to develop CIM providers. The main part is in the libvirt-cim module. The CIM provider is currently tested with the Pegasus [2] and SFCB [3] environment and the goal is to ensure compatibility with the various CIM based management systems. The project is hosted on libvirt.org, the source code is managed using Mercurial [4] and Dan Smith [5] is the project maintainer doing the commits. The project home page is available at: http://libvirt.org/CIM/ source code download is available at ftp://libvirt.org/libvirt-cim/ and the Mercurial tree is located at: http://libvirt.org/hg/ A dedicated mailing-list has been set up, libvirt-cim(a)redhat.com with the usual list page and archive mailman pages at: http://www.redhat.com/mailman/listinfo/libvirt-cim Daniel [1] http://en.wikipedia.org/wiki/Common_Information_Model_%28computing%29 [2] http://www.openpegasus.org [3] http://sblim.wiki.sourceforge.net/Sfcb [4] Mercurial: http://www.selenic.com/mercurial/wiki/ [5] Dan Smith <danms(a)us.ibm.com> -- 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/

1 0

[Libvir] collectd libvirtstats plugin (version 1, for discussion only)
by Richard W.M. Jones 05 Nov '07

05 Nov '07

Attached is a plugin allowing collectd to use libvirt to collect CPU, disk and network statistics for domains. There are a few problems with it at the moment, described below. Firstly just to say that this is a patch against the released collectd 4.2.0. I need to pull out the GIT repository and rebase the patch against that. Any other comments, most welcome! Usage ----- With this patch you should be able to collect stats by adding the following to your collectd.conf file: LoadPlugin libvirtstats <Plugin libvirtstats> Connection "xen:///" </Plugin> 'collectd' only needs to be running on the host. It uses libvirt to collect stats for each guest, via mechanisms in the hypervisor or host operating system. Requirements ------------ libvirt >= 0.3.3 and libxml2. The configure file should auto-detect the requirements and configure libvirtstats plugin only if they are available. Problems -------- I can't seem to get the ignorelist to do anything at all. It seems like it includes all the domains and devices, no matter what I specify in the configuration file and what the state of IgnoreSelected is set to. In the types.db, I have set the max for virt_cpu_total to be 256,000,000,000. The max CPU usage for a single CPU is 1 billion nanoseconds per second, but virt_cpu_total is for all CPUs in the system, and there could be (in theory) an unlimited number of CPUs, so an unlimited max. Therefore I have assumed here that max CPUs = 256. I'm using CSV output, which works, but it seems to report the aggregate counter values instead of the differences. Is that to be expected? 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] finish NUMA code reorg, plug cpuset at creat time support
by Daniel Veillard 05 Nov '07

05 Nov '07

The following patch finishes the cleanup for NUMA parsing code: - the cpuset parsing is moved to xml.c - some comments and cleanups of the include then add the output of a (cpus '...') line based on the /domain/vcpu/@cpuset attributes, this is parsed and reserialized as ranges to avoid any possibility of misinterpretation of say ^ or any special syntax we may want to add in the future. A few things to note: - dependant on the tiny patch I sent earlier today - if we notice that the set covers all CPU maybe we should avoid outputing (cpus '...'), trivial to add - mostly untested yet 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/

3 10