- Devel - libvirt List Archives

[libvirt] [PATCH] virsh: check if specified debug level is in range
by Ján Tomko 26 Jul '12

26 Jul '12

--- tools/virsh.c | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 5658796..34ae171 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -3097,6 +3097,10 @@ vshParseArgv(vshControl *ctl, int argc, char **argv) vshError(ctl, "%s", _("option -d takes a numeric argument")); exit(EXIT_FAILURE); } + if (ctl->debug < VSH_ERR_DEBUG || ctl->debug > VSH_ERR_ERROR) { + vshError(ctl, _("ignoring debug level %d out of range [0-4]"), + ctl->debug); + } break; case 'h': vshUsage(); -- 1.7.8.6

2 3

[libvirt] No /dev/tty device within LXC guest?
by Yufang Zhang 26 Jul '12

26 Jul '12

Hi all, When I create a LXC guest(RHEL5.4) via libvirt, the guest could boot successfully. I could ping to guest, but can not ssh into the it. The secure log in the guest shows that this is due to lack of '/dev/tty' device: # tail /path/to/rootfs/of/container/var/log/secure Jul 27 04:25:25 xen2143v sshd[671]: Accepted password for zhangyufang from 10.0.0.1 port 34102 ssh2 Jul 27 04:25:25 xen2143v sshd[671]: pam_unix(sshd:session): session opened for user zhangyufang by (uid=0) Jul 27 04:25:25 xen2143v sshd[671]: pam_loginuid(sshd:session): set_loginuid failed *Jul 27 04:25:25 xen2143v sshd[671]: error: PAM: pam_open_session(): Cannot make/remove an entry for the specified session* *Jul 27 04:25:25 xen2143v sshd[674]: error: open /dev/tty failed - could not set controlling tty: No such file or directory* *Jul 27 04:25:25 xen2143v sshd[673]: Received disconnect from 10.0.0.1: 11: disconnected by user* The xml of the guest: # virsh --connect lxc:/// dumpxml instance-00000034 <domain type='lxc' id='30090'> <name>instance-00000034</name> <uuid>8387e0f1-6df3-41c8-82ba-cde7d2b60844</uuid> <memory>524288</memory> <currentMemory>524288</currentMemory> <vcpu>1</vcpu> <os> <type arch='x86_64'>exe</type> <init>/sbin/init</init> </os> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <emulator>/usr/libexec/libvirt_lxc</emulator> <filesystem type='mount' accessmode='passthrough'> <source dir='/var/lib/nova/instances/instance-00000034/rootfs'/> <target dir='/'/> </filesystem> <interface type='bridge'> <mac address='fa:16:3e:3c:1d:95'/> <source bridge='demonetbr0'/> <target dev='veth0'/> <filterref filter='nova-instance-instance-00000034-fa163e3c1d95'> <parameter name='DHCPSERVER' value='10.0.0.1'/> <parameter name='IP' value='10.0.0.3'/> </filterref> </interface> <console type='pty' tty='/dev/pts/0'> <source path='/dev/pts/0'/> <target type='serial' port='0'/> </console> </devices> </domain> Could anyone point me to where the problem is? Thanks. Best Regards. Yufang

1 0

[libvirt] [PATCH 0/2] Refactor and fix error resetting on fallback paths in virsh
by Peter Krempa 26 Jul '12

26 Jul '12

This series fixes places where errors were not reset correctly and adds a function for resetting libvirt errors in virsh. Peter Krempa (2): virsh: Refactor error clearing on graceful fallback paths virsh: Fix error resetting on fallback paths tools/virsh-domain-monitor.c | 16 ++++++---------- tools/virsh-domain.c | 35 ++++++++++++----------------------- tools/virsh-network.c | 3 +-- tools/virsh-pool.c | 3 +-- tools/virsh-snapshot.c | 19 +++++++------------ tools/virsh.c | 13 +++++++++++-- 6 files changed, 38 insertions(+), 51 deletions(-) -- 1.7.8.6

1 3

[libvirt] [PATCH v3][PATCH] add test case for block job lifecycle testing
by Kyla Zhang 26 Jul '12

26 Jul '12

v3: Add skip for block_pull not implemented v2: Add skip block for qemu only and 120s timeout for test v1: Add tests for block job lifecyle and the test flow is as follows: create 50M qed img with qed backing img-> block pull->abort block job->resume block pull->set block job speed-> wait to finish --- scripts/qemu/300-blockjob-lifecycle.t | 146 +++++++++++++++++++++++++++++++++ 1 files changed, 146 insertions(+), 0 deletions(-) create mode 100644 scripts/qemu/300-blockjob-lifecycle.t diff --git a/scripts/qemu/300-blockjob-lifecycle.t b/scripts/qemu/300-blockjob-lifecycle.t new file mode 100644 index 0000000..e43dfe2 --- /dev/null +++ b/scripts/qemu/300-blockjob-lifecycle.t @@ -0,0 +1,146 @@ +# -*- perl -*- +# +# Copyright (C) 2011-2012 Red Hat, Inc. +# Copyright (C) 2011 Xiaoqiang Hu <xhu redhat com> +# +# This program is free software; You can redistribute it and/or modify +# it under the GNU General Public License as published by the Free +# Software Foundation; either version 2, or (at your option) any +# later version +# +# The file "LICENSE" distributed along with this file provides full +# details of the terms and conditions +# + +=pod + +=head1 NAME + +qemu/300-blockjob-lifecycle.t - verify the lifecycle of block job: +block pull, set block job speed, get block job info and abort block job + +=head1 DESCRIPTION + +The test case validates that it is possible to block pull, set block job +speed, get block job info and abort block job for domain using qed img +with qed backing img + +=cut + +use strict; +use warnings; + +use Test::More tests => 16; + +use Sys::Virt::TCK; +use Test::Exception; +use File::Spec::Functions qw(catfile); +use File::stat; + +my $tck = Sys::Virt::TCK->new(); +my $conn = eval { $tck->setup(); }; +BAIL_OUT "failed to setup test harness: $@" if $@; +END { + $tck->cleanup if $tck; +} + +SKIP:{ + skip "Only relevant to QEMU driver", 16 unless $conn->get_type() eq "QEMU"; + my $xml = $tck->generic_pool("dir") + ->mode("0755")->as_xml; + + diag "Defining transient storage pool $xml"; + my $pool; + + ok_pool(sub { $pool = $conn->define_storage_pool($xml) }, "define transient storage pool"); + lives_ok(sub { $pool->build(0) }, "built storage pool"); + lives_ok(sub { $pool->create }, "started storage pool"); + + my $volbackxml = $tck->generic_volume("tck-back", "qed", 1024*1024*50)->allocation(0)->as_xml; + + my ($volback, $pathback); + diag "back $volbackxml"; + ok_volume(sub { $volback = $pool->create_volume($volbackxml) }, "create qed backing file volume"); + + my $st; + $pathback = xpath($volback, "string(/volume/target/path)"); + $st = stat($pathback); + + ok($st, "path $pathback exists"); + + ok($st->size < 1024*1024, "size is < 1M"); + + my $volmainxml = $tck->generic_volume("tck-main", "qed", 1024*1024*50) + ->backing_file($pathback) + ->backing_format("qed") + ->allocation(0)->as_xml; + + my ($volmain, $pathmain); + diag "main $volmainxml"; + ok_volume(sub { $volmain = $pool->create_volume($volmainxml) }, "create qed backing file volume"); + + $pathmain = xpath($volmain, "string(/volume/target/path)"); + $st = stat($pathmain); + + ok($st, "path $pathmain exists"); + + ok($st->size < 1024*1024, "size is < 1M"); + + # define the guest at a qed image + # and the backing store in this qed image. + $xml = $tck->generic_domain("tck") + ->disk(format => { name => "qemu", type => "qed" }, + type => "file", + src => $pathmain, + dst => "vdb") + ->as_xml; + + diag "Defining an inactive domain config $xml"; + my $dom; + ok_domain(sub { $dom = $conn->define_domain($xml) }, "defined persistent domain config"); + + diag "Starting inactive domain config"; + $dom->create; + ok($dom->get_id() > 0, "running domain has an ID > 0"); + + # start to block pull and bandwidth is 1MB/S + my ($bandwidth, $flags, $jobinfo, $timeout); + # 1MB/S + $bandwidth = 1; + $flags=0; + eval { $dom->block_pull($pathmain, $bandwidth, $flags); }; + SKIP: { + skip "block_pull not implemented", 5 if $@ && $@->code() == 55; + + # $jobinfo is a hash reference summarising the execution state of the block job + # and it has four keys:cur, end, bandwidth, type + $jobinfo = $dom->get_block_job_info($pathmain, $flags); + ok($jobinfo->{bandwidth} == $bandwidth, "start to block pull and block job bandwidth is $bandwidth"."MB/S"); + + $dom->abort_block_job($pathmain, $flags); + $jobinfo = $dom->get_block_job_info($pathmain, $flags); + ok($jobinfo->{type} == 0, "abort block job"); + + $dom->block_pull($pathmain, $bandwidth, $flags); + $jobinfo = $dom->get_block_job_info($pathmain, $flags); + ok($jobinfo->{bandwidth} == $bandwidth, "continue to block pull and block job bandwidth is $bandwidth"."MB/S"); + + # set block job bandwidth to 2MB/S + $bandwidth = 2; + $dom->set_block_job_speed($pathmain, $bandwidth, $flags); + $jobinfo = $dom->get_block_job_info($pathmain, $flags); + ok($jobinfo->{bandwidth} == $bandwidth, "block job bandwidth is set to $bandwidth"."MB/S"); + + # wait for the end of block pull and timeout is 120s + $timeout = 120; + while($jobinfo->{cur} < $jobinfo->{end} && $jobinfo->{type} == 1 && $timeout > 0) { + sleep(1); + $jobinfo = $dom->get_block_job_info($pathmain, $flags); + $timeout--; + } + + diag "block pull is not finished in 120S" if $jobinfo->{type} == 1 && $timeout == 0; + ok($jobinfo->{type} == 0, "block pull is finished"); + } +} +# end -- 1.7.1

1 0

[libvirt] [PATCH] virsh: Check for existence of storage before undefining the domain
by Peter Krempa 26 Jul '12

26 Jul '12

When undefining a domain and removing associated storage using "virsh undefine --storage" the domain was at first undefined and after that the storage removal proces was started. If the user specified an invalid disk to remove, the error could not be corrected. This patch moves enumeration and filtering of volumes that should be removed before the domain is undefined, but the removal process is still kept after the domain has been undefined. --- tools/virsh.c | 280 +++++++++++++++++++++++++++++---------------------------- 1 files changed, 144 insertions(+), 136 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 0354822..e6b8e5f 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -3007,6 +3007,12 @@ static const vshCmdOptDef opts_undefine[] = { {NULL, 0, 0, NULL} }; +typedef struct { + virStorageVolPtr vol; + char *source; + char *target; +} vshUndefineVolume; + static bool cmdUndefine(vshControl *ctl, const vshCmd *cmd) { @@ -3019,7 +3025,6 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) bool managed_save = vshCommandOptBool(cmd, "managed-save"); bool snapshots_metadata = vshCommandOptBool(cmd, "snapshots-metadata"); bool wipe_storage = vshCommandOptBool(cmd, "wipe-storage"); - bool remove_storage = false; bool remove_all_storage = vshCommandOptBool(cmd, "remove-all-storage"); /* Positive if these items exist. */ int has_managed_save = 0; @@ -3031,6 +3036,8 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) int rc = -1; int running; /* list of volumes to remove along with this domain */ + vshUndefineVolume *vlist = NULL; + int nvols = 0; const char *volumes_arg = NULL; char *volumes = NULL; char **volume_tokens = NULL; @@ -3045,8 +3052,10 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) xmlXPathContextPtr ctxt = NULL; xmlNodePtr *vol_nodes = NULL; int nvolumes = 0; - virStorageVolPtr vol = NULL; - bool vol_del_failed = false; + bool vol_not_found = false; + + ignore_value(vshCommandOptString(cmd, "storage", &volumes_arg)); + volumes = vshStrdup(ctl, volumes_arg); if (managed_save) { flags |= VIR_DOMAIN_UNDEFINE_MANAGED_SAVE; @@ -3063,36 +3072,21 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) if (!(dom = vshCommandOptDomain(ctl, cmd, &name))) return false; - /* check if a string that should contain list of volumes to remove is present */ - if (vshCommandOptString(cmd, "storage", &volumes_arg) > 0) { - volumes = vshStrdup(ctl, volumes_arg); - - if (remove_all_storage) { - vshError(ctl, _("Specified both --storage and --remove-all-storage")); - goto cleanup; - } - remove_storage = true; - } - /* Do some flag manipulation. The goal here is to disable bits * from flags to reduce the likelihood of a server rejecting * unknown flag bits, as well as to track conditions which are * safe by default for the given hypervisor and server version. */ - running = virDomainIsActive(dom); - if (running < 0) { - virshReportError(ctl); - goto cleanup; - } + if ((running = virDomainIsActive(dom)) < 0) + goto error; + if (!running) { /* Undefine with snapshots only fails for inactive domains, * and managed save only exists on inactive domains; if * running, then we don't want to remove anything. */ has_managed_save = virDomainHasManagedSaveImage(dom, 0); if (has_managed_save < 0) { - if (last_error->code != VIR_ERR_NO_SUPPORT) { - virshReportError(ctl); - goto cleanup; - } + if (last_error->code != VIR_ERR_NO_SUPPORT) + goto error; virFreeError(last_error); last_error = NULL; has_managed_save = 0; @@ -3100,10 +3094,8 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) has_snapshots = virDomainSnapshotNum(dom, 0); if (has_snapshots < 0) { - if (last_error->code != VIR_ERR_NO_SUPPORT) { - virshReportError(ctl); - goto cleanup; - } + if (last_error->code != VIR_ERR_NO_SUPPORT) + goto error; virFreeError(last_error); last_error = NULL; has_snapshots = 0; @@ -3137,16 +3129,116 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) } /* Stash domain description for later use */ - if (remove_storage || remove_all_storage) { + if (volumes || remove_all_storage) { if (running) { vshError(ctl, _("Storage volume deletion is supported only on stopped domains")); goto cleanup; } + if (volumes && remove_all_storage) { + vshError(ctl, _("Specified both --storage and --remove-all-storage")); + goto cleanup; + } + if (!(def = virDomainGetXMLDesc(dom, 0))) { vshError(ctl, _("Could not retrieve domain XML description")); goto cleanup; } + + if (!(doc = virXMLParseStringCtxt(def, _("(domain_definition)"), + &ctxt))) + goto error; + + /* tokenize the string from user and save it's parts into an array */ + if (volumes) { + /* count the delimiters */ + volume_tok = volumes; + nvolume_tokens = 1; /* we need at least one member */ + while (*volume_tok) { + if (*(volume_tok++) == ',') + nvolume_tokens++; + } + + volume_tokens = vshCalloc(ctl, nvolume_tokens, sizeof(char *)); + + /* tokenize the input string */ + nvolume_tokens = 0; + volume_tok = volumes; + do { + volume_tokens[nvolume_tokens] = strsep(&volume_tok, ","); + nvolume_tokens++; + } while (volume_tok); + } + + if ((nvolumes = virXPathNodeSet("./devices/disk", ctxt, + &vol_nodes)) < 0) + goto error; + + if (nvolumes > 0) + vlist = vshCalloc(ctl, nvolumes, sizeof(*vlist)); + + for (vol_i = 0; vol_i < nvolumes; vol_i++) { + ctxt->node = vol_nodes[vol_i]; + + /* get volume source and target paths */ + if (!(target = virXPathString("string(./target/@dev)", ctxt))) + goto error; + + if (!(source = virXPathString("string(" + "./source/@file|" + "./source/@dir|" + "./source/@name|" + "./source/@dev)", ctxt))) { + if (last_error && last_error->code != VIR_ERR_OK) + goto error; + else + continue; + } + + /* lookup if volume was selected by user */ + if (volumes) { + volume_tok = NULL; + for (tok_i = 0; tok_i < nvolume_tokens; tok_i++) { + if (volume_tokens[tok_i] && + (STREQ(volume_tokens[tok_i], target) || + STREQ(volume_tokens[tok_i], source))) { + volume_tok = volume_tokens[tok_i]; + volume_tokens[tok_i] = NULL; + break; + } + } + if (!volume_tok) + continue; + } + + if (!(vlist[nvols].vol = virStorageVolLookupByPath(ctl->conn, + source))) { + vshPrint(ctl, + _("Storage volume '%s'(%s) is not managed by libvirt. " + "Remove it manually.\n"), target, source); + virFreeError(last_error); + last_error = NULL; + continue; + } + vlist[nvols].source = source; + vlist[nvols].target = target; + nvols++; + } + + /* print volumes specified by user that were not found in domain definition */ + if (volumes) { + for (tok_i = 0; tok_i < nvolume_tokens; tok_i++) { + if (volume_tokens[tok_i]) { + vshError(ctl, _("Volume '%s' was not found in domain's " + "definition.\n"), + volume_tokens[tok_i]); + vol_not_found = true; + } + } + + if (vol_not_found) + goto cleanup; + } } /* Generally we want to try the new API first. However, while @@ -3202,96 +3294,15 @@ out: } /* try to undefine storage volumes associated with this domain, if it's requested */ - if (remove_storage || remove_all_storage) { - ret = false; - - /* tokenize the string from user and save it's parts into an array */ - if (volumes) { - /* count the delimiters */ - volume_tok = volumes; - nvolume_tokens = 1; /* we need at least one member */ - while (*volume_tok) { - if (*volume_tok == ',') - nvolume_tokens++; - volume_tok++; - } - - volume_tokens = vshCalloc(ctl, nvolume_tokens, sizeof(char *)); - - /* tokenize the input string */ - nvolume_tokens = 0; - volume_tok = volumes; - do { - volume_tokens[nvolume_tokens] = strsep(&volume_tok, ","); - nvolume_tokens++; - } while (volume_tok); - } - - doc = virXMLParseStringCtxt(def, _("(domain_definition)"), &ctxt); - if (!doc) - goto cleanup; - - nvolumes = virXPathNodeSet("./devices/disk", ctxt, &vol_nodes); - - if (nvolumes < 0) - goto cleanup; - - for (vol_i = 0; vol_i < nvolumes; vol_i++) { - ctxt->node = vol_nodes[vol_i]; - VIR_FREE(target); - VIR_FREE(source); - if (vol) { - virStorageVolFree(vol); - vol = NULL; - } - - /* get volume source and target paths */ - if (!(target = virXPathString("string(./target/@dev)", ctxt))) { - vshError(ctl, _("Failed to enumerate devices")); - goto cleanup; - } - - if (!(source = virXPathString("string(" - "./source/@file|" - "./source/@dir|" - "./source/@name|" - "./source/@dev)", ctxt)) && - virGetLastError()) - goto cleanup; - - /* lookup if volume was selected by user */ - if (volumes) { - volume_tok = NULL; - for (tok_i = 0; tok_i < nvolume_tokens; tok_i++) { - if (volume_tokens[tok_i] && - (STREQ_NULLABLE(volume_tokens[tok_i], target) || - STREQ_NULLABLE(volume_tokens[tok_i], source))) { - volume_tok = volume_tokens[tok_i]; - volume_tokens[tok_i] = NULL; - break; - } - } - if (!volume_tok) - continue; - } - - if (!source) - continue; - - if (!(vol = virStorageVolLookupByPath(ctl->conn, source))) { - vshPrint(ctl, - _("Storage volume '%s'(%s) is not managed by libvirt. " - "Remove it manually.\n"), target, source); - virResetLastError(); - continue; - } - + if (nvols) { + for (vol_i = 0; vol_i < nvols; vol_i++) { if (wipe_storage) { - vshPrint(ctl, _("Wiping volume '%s'(%s) ... "), target, source); + vshPrint(ctl, _("Wiping volume '%s'(%s) ... "), + vlist[vol_i].target, vlist[vol_i].source); fflush(stdout); - if (virStorageVolWipe(vol, 0) < 0) { + if (virStorageVolWipe(vlist[vol_i].vol, 0) < 0) { vshError(ctl, _("Failed! Volume not removed.")); - vol_del_failed = true; + ret = false; continue; } else { vshPrint(ctl, _("Done.\n")); @@ -3299,41 +3310,38 @@ out: } /* delete the volume */ - if (virStorageVolDelete(vol, 0) < 0) { + if (virStorageVolDelete(vlist[vol_i].vol, 0) < 0) { vshError(ctl, _("Failed to remove storage volume '%s'(%s)"), - target, source); - vol_del_failed = true; - } - vshPrint(ctl, _("Volume '%s' removed.\n"), volume_tok?volume_tok:source); - } - - /* print volumes specified by user that were not found in domain definition */ - if (volumes) { - for (tok_i = 0; tok_i < nvolume_tokens; tok_i++) { - if (volume_tokens[tok_i]) - vshPrint(ctl, _("Volume '%s' was not found in domain's " - "definition.\n"), - volume_tokens[tok_i]); + vlist[vol_i].target, vlist[vol_i].source); + ret = false; + } else { + vshPrint(ctl, _("Volume '%s'(%s) removed.\n"), + vlist[vol_i].target, vlist[vol_i].source); } } - - if (!vol_del_failed) - ret = true; } cleanup: - VIR_FREE(source); - VIR_FREE(target); + for (vol_i = 0; vol_i < nvols; vol_i++) { + VIR_FREE(vlist[vol_i].source); + VIR_FREE(vlist[vol_i].target); + if (vlist[vol_i].vol) + virStorageVolFree(vlist[vol_i].vol); + } + VIR_FREE(vlist); + VIR_FREE(volumes); VIR_FREE(volume_tokens); VIR_FREE(def); VIR_FREE(vol_nodes); - if (vol) - virStorageVolFree(vol); xmlFreeDoc(doc); xmlXPathFreeContext(ctxt); virDomainFree(dom); return ret; + +error: + virshReportError(ctl); + goto cleanup; } -- 1.7.8.6

2 4

[libvirt] [PATCH v3] util: set minimum value of nodesuspend duration to 60 seconds
by Guannan Ren 26 Jul '12

26 Jul '12

Change the permissible minimum value of nodesuspend duration time to 60 seconds. If option is less than the value, reports error. Update virsh help and manpage the infomation. --- src/util/virnodesuspend.c | 2 +- tools/virsh.c | 2 +- tools/virsh.pod | 3 ++- 3 files changed, 4 insertions(+), 3 deletions(-) diff --git a/src/util/virnodesuspend.c b/src/util/virnodesuspend.c index 71beb3d..91dfdf6 100644 --- a/src/util/virnodesuspend.c +++ b/src/util/virnodesuspend.c @@ -90,7 +90,7 @@ static int virNodeSuspendSetNodeWakeup(unsigned long long alarmTime) virCommandPtr setAlarmCmd; int ret = -1; - if (alarmTime <= MIN_TIME_REQ_FOR_SUSPEND) { + if (alarmTime < MIN_TIME_REQ_FOR_SUSPEND) { virReportError(VIR_ERR_INVALID_ARG, "%s", _("Suspend duration is too short")); return -1; } diff --git a/tools/virsh.c b/tools/virsh.c index 6d65036..d236950 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -6986,7 +6986,7 @@ static const vshCmdInfo info_nodesuspend[] = { static const vshCmdOptDef opts_node_suspend[] = { {"target", VSH_OT_DATA, VSH_OFLAG_REQ, N_("mem(Suspend-to-RAM), " "disk(Suspend-to-Disk), hybrid(Hybrid-Suspend)")}, - {"duration", VSH_OT_INT, VSH_OFLAG_REQ, N_("Suspend duration in seconds")}, + {"duration", VSH_OT_INT, VSH_OFLAG_REQ, N_("Suspend duration in seconds, at least 60")}, {"flags", VSH_OT_INT, VSH_OFLAG_NONE, N_("Suspend flags, 0 for default")}, {NULL, 0, 0, NULL} }; diff --git a/tools/virsh.pod b/tools/virsh.pod index 6ecf6ce..dffd588 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -290,7 +290,8 @@ If I<cell> is specified, this will prints specified cell statistics only. Puts the node (host machine) into a system-wide sleep state such as Suspend-to-RAM, Suspend-to-Disk or Hybrid-Suspend and sets up a Real-Time-Clock interrupt to fire (to wake up the node) after a time delay -specified by the 'duration' parameter. +specified by the 'duration' parameter. The duration time should be +at least 60 seconds. =item B<capabilities> -- 1.7.7.5

2 2

[libvirt] [PATCH] doc: add more description on libvirtd option timeout
by Guannan Ren 26 Jul '12

26 Jul '12

Only when there is no running guest, virtual network, etc on each registered drivers and no client connections, the timeout take effect. --- daemon/libvirtd.pod.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/daemon/libvirtd.pod.in b/daemon/libvirtd.pod.in index ea6c37d..75c9040 100644 --- a/daemon/libvirtd.pod.in +++ b/daemon/libvirtd.pod.in @@ -54,7 +54,7 @@ Use this name for the PID file, overriding the default value. =item B<-t, --timeout> I<SECONDS> -Exit after timeout period (in seconds) expires. +When there is no any active resources on each registered driver and client connections, exit after timeout period (in seconds) expires. =item B<-v, --verbose> -- 1.7.10.4

2 5

[libvirt] [PATCH 0/4] Improve netlink to support all protocols.
by tangchen 26 Jul '12

26 Jul '12

Now, netlink in libvirt only suport NETLINK_ROUTE protocol. But we need other netlinks, especially to support cpu and memory hotplug. These patches did the following things: 1) Change global variable "server" to an array to store all the netlink sockets, and modified all the API prototypes. 2) Introduce a new API virNetlinkEventServiceStopAll() to stop all netlink servers. 3) Modify all the function calls. Tang Chen (4): Improve netlink to support all protocol. Introduce virNetlinkEventServiceStopAll() to stop all netlink services in libvirt. Modify all the function calls for API in virnetlink.c DEBUG: print out netlink message. daemon/libvirtd.c | 12 +- src/libvirt_private.syms | 1 + src/util/virnetdev.c | 6 +- src/util/virnetdevmacvlan.c | 12 +- src/util/virnetdevvportprofile.c | 7 +- src/util/virnetlink.c | 235 +++++++++++++++++++++++++++++++------- src/util/virnetlink.h | 22 ++-- 7 files changed, 232 insertions(+), 63 deletions(-) -- 1.7.10.2 -- Best Regards, Tang chen

1 5

Re: [libvirt] [PATCH] vl: add -libvirt-caps option for libvirt to stop parsing help output
by Eric Blake 26 Jul '12

26 Jul '12

[adding libvirt] On 07/25/2012 01:47 PM, Anthony Liguori wrote: > The help output is going to change dramatically for 0.13. We've spent too long 0.13? How long have you been sitting on this commit? :) > waiting for a perfect solution to capabilities handling and the end loser is > the direct consumer of QEMU because the help output is awful. > > I will not apply any patches that change help output until 0.13 development > opens up. This should give libvirt adequate time to implement support for > dealing with this new option. > > This capabilities set comes directly from libvirt's source code so it's entirely > adequate for libvirt's purposes. We can still explore more sophisticated > approaches that are more general purpose but the help output will be changing. We've gone a long way with things like newer libvirt requiring QMP, and being able to use query-commands and even 'qemu -device ?' to figure out which devices are present, which is already more reliable than parsing -help output. There may still be a few things to fix (for example, I already pointed out that libvirt 0.9.13 is bogus for refusing to even attempt 'qemu -device,?' unless it guesses from '-help' output that it will work; it would be better to just attempt it in the first place and deal with the fallout), but it should be easy to fix in time for libvirt 0.10.0, and certainly coordinate things so that new enough libvirt comes out close to the time of qemu 1.2. > +++ b/libvirt-caps.c > @@ -0,0 +1,166 @@ > +/* > + * Libvirt capabilities > + * > + * Copyright IBM, Corp. 2012 > + * > + * Authors: > + * Anthony Liguori <aliguori(a)us.ibm.com> > + * > + * This work is licensed under the terms of the GNU GPL, version 2 or later. > + * See the COPYING file in the top-level directory. > + * > + */ > + > +#include "libvirt-caps.h" > +#include "qemu-common.h" > + > +/* Make sure this stays in sync with libvirt/src/qemu/qemu_capabilities.c */ This is the part I'm most worried about - it says that any time libvirt decides to flag a new capability that it cares about within qemu, then both libvirt and qemu have to be upgraded in lockstep to coordinate the use of that capability. The problem is that sometimes libvirt doesn't care about a capability until after qemu has already been released (for example, we didn't realize the power of fd: migration until several releases after qemu had first implemented it, at which point when we added the capability bit, we were able to retroactively detect it for several qemu versions by parsing -help). > + > +static const char *supported_caps[] = { > +// "kqemu", > + "vnc-colon", > + "no-reboot", > + "drive", > + "drive-boot", > + ... > + "drive-iotune", > + "system_wakeup", > + "scsi-disk.channel", > + "scsi-block", > + "transaction", > + > + NULL Already incomplete. For example, libvirt has recently added block-job-sync, block-job-async, scsi-cd, and several others. Thankfully, these days most of the new feature bits being added in libvirt's qemu_capabilities.h are related to things that were probed from 'qemu -device ?' or from QMP 'query-commands', and not from something additionally scraped from '-help' for the first time. I had to go back to libvirt commit 63b4243 (May 15) to find a capability that libvirt learned by scraping -help output (namely, support for -no-user-config). But yet this is one of the important config bits that libvirt really needs to know, and not one that can easily be learned from machine-parseable '-device ?' I'm worried that an interface like this will let libvirt know about the existing features that libvirt is trying to use, but will hurt the ability of adding new feature detection in to libvirt (basically, a newer libvirt wouldn't be able to retroactively take advantage of an older feature already present in older qemu without parsing some back-channel, at which point what good was having this libvirt-specific channel?). There's also another problem if we decide to keep this file synced across projects: > +++ b/libvirt-caps.h > @@ -0,0 +1,19 @@ > +/* > + * Libvirt capabilities > + * > + * Copyright IBM, Corp. 2012 > + * > + * Authors: > + * Anthony Liguori <aliguori(a)us.ibm.com> > + * > + * This work is licensed under the terms of the GNU GPL, version 2 or later. > + * See the COPYING file in the top-level directory. Libvirt is LGPLv2+. You can obviously upgrade to GPLv2+ as part of importing the code from libvirt to qemu; but if you decide that a new feature is worth adding a bit to this file, then libvirt can't do the reverse sync unless you relax the license of this file. -- Eric Blake eblake(a)redhat.com +1-919-301-3266 Libvirt virtualization library http://libvirt.org

1 0

[libvirt] Changing qemu's -help output
by Luiz Capitulino 25 Jul '12

25 Jul '12

On Wed, 25 Jul 2012 20:02:53 +0100 Peter Maydell <peter.maydell(a)linaro.org> wrote: > On 25 July 2012 20:02, Luiz Capitulino <lcapitulino(a)redhat.com> wrote: > > On Wed, 25 Jul 2012 19:58:02 +0100 > > Peter Maydell <peter.maydell(a)linaro.org> wrote: > >> I think we should simply say "no, parsing -help is broken and wrong and it > >> was obviously broken and wrong and we are in fact going to change the > >> help output for QEMU 1.2, and you will need a new libvirt that can > >> cope with that". We can't be held hostage forever to really bad decisions > >> like that. > > > > We have to provide an alternative before doing that. > > Try whatever it is you wanted to try, see if it barfs. Or don't use it. Libvirt folks can answer if this is feasible (CC'ing them), I'd guess it's not.

2 1