March 2016 - Devel - libvirt List Archives

[libvirt] [PATCH] docs: Document NSS module
by Michal Privoznik 29 Mar '16

29 Mar '16

While we have a wiki page describing the feature [1] since the feature is distributed in our .tar.gz we ought to document it. So I went ahead, copied the wiki page and reformatted so it fits our docs coding style. 1: http://wiki.libvirt.org/page/NSS_module Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- docs/nss.html.in | 141 +++++++++++++++++++++++++++++++++++++++++++++++++++ docs/sitemap.html.in | 4 ++ 2 files changed, 145 insertions(+) create mode 100644 docs/nss.html.in diff --git a/docs/nss.html.in b/docs/nss.html.in new file mode 100644 index 0000000..84ea8df --- /dev/null +++ b/docs/nss.html.in @@ -0,0 +1,141 @@ +<?xml version="1.0" encoding="UTF-8"?> +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> +<html xmlns="http://www.w3.org/1999/xhtml"> + <body> + <h1>Libvirt NSS module</h1> + + <ul id="toc"></ul> + + <p> + When it comes to managing guests and executing commands inside them, logging + into guest operating system and doing the job is convenient. Users are used + to ssh in this case. Ideally: + </p> + + <code>ssh user@virtualMachine</code> + + <p> + would be nice. But depending on virtual network configuration it might not + be always possible. For instance, when using libvirt NATed network it's + dnsmasq (spawned by libvirt) who assigns IP addresses to domains. But by + default, the dnsmasq process is then not consulted when it comes to host + name translation. Users work around this problem by configuring their + libvirt network to assign static IP addresses and maintaining + <code>/etc/hosts</code> file in sync. But this puts needless burden onto + users. This is where NSS module comes handy. + </p> + + <h2><a name="Installation">Installation</a></h2> + + <p> + Installing the module is really easy: + </p> + +<pre> +# yum install libvirt-nss +</pre> + + <h2><a name="Configuration">Configuration</a></h2> + + <p> + Enabling the module is really easy. Just add <b>libvirt</b> into + <code>/etc/nsswitch.conf</code> file. For instance: + </p> + +<pre> +$ cat /etc/nsswitch.conf +# /etc/nsswitch.conf: +passwd: compat +shadow: compat +group: compat +hosts: files libvirt dns +# ... +</pre> + + <p> + So, in this specific case, whenever ssh program is looking up the host user + is trying to connect to, <b>files</b> module is consulted first (which + boils down to looking up the host name in <code>/etc/hosts</code> file), if + not found <b>libvirt</b> module is consulted then. The DNS is the last + effort then, if none of the previous modules matched the host in question. + Therefore users should consider the order in which they want the modules to + lookup given host name. + </p> + + <h2><a name="Internals">How does it work?</a></h2> + + <p> + Whenever an Unix process wants to do a host name translation + <a href="http://linux.die.net/man/3/gethostbyname"><code>gethostbyname()</code></a> + or some variant of it is called. This is a glibc function that takes a + string containing the host name, crunch it and produces a list of IP + addresses assigned to that host. Now, glibc developers made a really good + decision when implementing the internals of the function when they decided + to make the function pluggable. Since there can be several sources for the + records (e.g. <code>/etc/hosts</code> file, DNS, LDAP, etc.) it would not + make much sense to create one big implementation containing all possible + cases. What they have done instead is this pluggable mechanism. Small + plugins implementing nothing but specific technology for lookup process are + provided and the function then calls those plugins. There is just one + configuration file that instructs the lookup function in which order should + the plugins be called and which plugins should be loaded. For more info + reading <a href="https://en.wikipedia.org/wiki/Name_Service_Switch">wiki + page</a> is recommended. + </p> + + <p> + And this is point where libvirt comes in. Libvirt provides plugin for the + NSS ecosystem. For some time now libvirt keeps a list of assigned IP + addresses for libvirt networks. The NSS plugin does no more than search the + list trying to find matching record for given host name. When found, + matching IP address is returned to the caller. If not found, translation + process continues with the next plugin configured. At this point it is + important to stress the order in which plugins are called. Users should be + aware that a hostname might match in multiple plugins and right after first + match, translation process is terminated and no other plugin is consulted. + Therefore, if there are two different records for the same host name users + should carefully chose the lookup order. + </p> + + <h2><a name="Limitations">Limitations</a></h2> + + <ol> + <li>The libvirt NSS module matches only hostnames provided by guest. If + the libvirt name and one advertised by guest differs, the latter is + matched.</li> + <li>The module works only in that cases where IP addresses are assigned by + dnsmasq spawned by libvirt. Libvirt NATed networks are typical + example.</li> + </ol> + + <p> + These limitation are result of libvirt's internal implementation. While + libvirt can report IP addresses regardless of their origin, a public API + must be used to obtain those. However, for the API a connection object is + required. Doing that for every name translation request would be too + costly. Fortunately, libvirt spawns dnsmasq for NATed networks. Not only + that, it provides small executable that on each IP address space change + updates an internal list of addresses thus keeping it in sync. The NSS + module then merely consults the list trying to find the match. Users can + view the list themselves: + </p> + +<pre> +virsh net-dhcp-leases $network +</pre> + + <p> + where <code>$network</code> iterates through all running networks. So the module + does merely the same as + </p> + +<pre> +virsh domifaddr --source lease $domain +</pre> + + <p> + If there's no record for either of the aforementioned commands, it's very + likely that NSS module won't find anything and vice versa. + </p> + </body> +</html> diff --git a/docs/sitemap.html.in b/docs/sitemap.html.in index 7e3746e..22e410c 100644 --- a/docs/sitemap.html.in +++ b/docs/sitemap.html.in @@ -120,6 +120,10 @@ <a href="hooks.html">Hooks</a> <span>Hooks for system specific management</span> </li> + <li> + <a href="nss.html">NSS module</a> + <span>Enable domain host name translation to IP addresses</span> + </li> </ul> </li> <li> -- 2.7.3

2 1

[libvirt] [PATCH v3 0/2] qemu: Support for QXL heads
by Martin Kletzander 29 Mar '16

29 Mar '16

v3: - rebase on top of current master in order to cleanly apply v2: - don't change heads when migrating - https://www.redhat.com/archives/libvir-list/2016-March/msg00811.html v1: - https://www.redhat.com/archives/libvir-list/2016-March/msg00410.html Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1283207 Martin Kletzander (2): qemu: Check for qxl's max_outputs parameter qemu: Add support to QXL's max_outputs parameter src/qemu/qemu_capabilities.c | 4 ++ src/qemu/qemu_capabilities.h | 2 + src/qemu/qemu_command.c | 8 +++ src/qemu/qemu_migration.c | 64 ++++++++++++++++++++-- .../qemuxml2argv-video-qxl-heads.args | 28 ++++++++++ .../qemuxml2argv-video-qxl-heads.xml | 47 ++++++++++++++++ tests/qemuxml2argvtest.c | 8 +++ .../qemuxml2xmlout-video-qxl-heads.xml | 47 ++++++++++++++++ tests/qemuxml2xmltest.c | 2 + 9 files changed, 204 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-video-qxl-heads.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-video-qxl-heads.xml create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-video-qxl-heads.xml -- 2.7.4

2 4

[libvirt] Publishing libvirt 'patches' JSON database?
by Kashyap Chamarthy 29 Mar '16

29 Mar '16

I'd imagine most developers on this list have their own workflows to track/apply/test patches from the mailing list. I normally just save patches in Mutt's Maildir and apply them manually to a Git branch. I recently began using the 'patches'[1] tool, that some of you on this list must be familiar with, to test arbitrary QEMU patches from the mailing list, and found it relatively easy large patch series from mailing list. I'm writing this to check if the libvirt upstream community finds it desirable to setup such a patches database. What is 'patches'? ------------------ Quoting the README[2]: "patches is a patch tracking system. It consists of two parts: a set of commands that build a database of patches from a mailing list and then a set of commands that can search that database. It supports the following features: - Tracking patch status by determining which patches are already committed, have newer versions posted, are RFC, etc. - Applying patches or pull requests with a single command. - Searching for patches using a rich query language." A quick workflow to test patches from mailing list --------------------------------------------------- FWIW, this is my rudimentary use-case. Clone the repo, and fetch the patches database (may take a few minutes): $ git clone https://github.com/stefanha/patches.git $ cd patches $ ./patches fetch http://wiki.qemu.org/patches/patches.json And, suppose, I want to test Dan's LUKS encryption support (26) patch series, I go find the message ID: $ ./patches list | grep -B5 "Support LUKS encryption" Message-id: 1456747261-22032-1-git-send-email-berrange(a)redhat.com From: Daniel P. Berrange <berrange(a)redhat.com> Date: 2016-02-29 Tags: v4 [0/26] Support LUKS encryption in block devices Trivially apply it (assuming you have a test branch): $ ./patches apply --git-dir ~/src/qemu/ \ id:1456747261-22032-1-git-send-email-berrange@redhat.com Applying: crypto: add cryptographic random byte source Applying: crypto: add support for PBKDF2 algorithm Applying: crypto: add support for generating initialization vectors Applying: crypto: add support for anti-forensic split algorithm Applying: crypto: skip testing of unsupported cipher algorithms [...] Test away! Rinse, repeat: "The `fetch` command should be run whenever you want to refresh the patch database; it will download any new patches that have arrived since the last time you did a fetch." Here's[3] the documentation to setup the patches database. [1] https://github.com/stefanha/patches/tree/stefanha-tweaks [2] https://github.com/stefanha/patches/blob/stefanha-tweaks/README.md [3] https://github.com/stefanha/patches/blob/stefanha-tweaks/README-server.md -- /kashyap

3 7

[libvirt] Request for testing: migration of libvirt wiki site
by Daniel P. Berrange 29 Mar '16

29 Mar '16

I'm preparing for a migration of the libvirt wiki off the host it is currently running on, to OpenShift hosting. This isn't entirely straightforward, as I'm also upgrading the version of mediawiki and as a result had to re-write the CSS customization from scratch. I may well have messed something up, so would appreciate if people could have a quick look at this testing instance of the wiki and confirm that it looks / works reasonably for them. http://wiki-libvirt.rhcloud.com NB, any edits you make in this test instance will be discarded when I load a fresh database dump from the current live wiki. If all looks good, I'll look at getting DNS updated to switch over later this week. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|

5 12

[libvirt] [PATCH 00/10] Introduce worker tuning APIs
by Erik Skultety 29 Mar '16

29 Mar '16

Erik Skultety (10): libvirt-host: Move virTypedParam* to libvirt-common admin: Enable usage of typed parameters admin: Introduce virAdmConnectServerLookupByName util: Refactor thread creation by introducing virThreadPoolExpand util: Report system error when virThreadCreateFull fails util: Add more getters to threadpool parameters admin: Prepare admin protocol for future worker related procedures admin: Introduce virAdmServerGethreadPoolParameters admin: Introduce virAdmServerSetThreadPoolParameters virt-admin: Introduce srv-workertune command cfg.mk | 2 +- daemon/admin.c | 129 ++++++++++++++++++++++++ daemon/admin_server.c | 129 ++++++++++++++++++++++++ daemon/admin_server.h | 15 +++ include/libvirt/libvirt-admin.h | 74 ++++++++++++++ include/libvirt/libvirt-common.h.in | 185 ++++++++++++++++++++++++++++++++++ include/libvirt/libvirt-host.h | 186 ---------------------------------- include/libvirt/virterror.h | 1 + po/POTFILES.in | 4 +- src/admin/admin_protocol.x | 67 +++++++++++- src/admin/admin_remote.c | 104 +++++++++++++++++++ src/admin_protocol-structs | 52 ++++++++++ src/libvirt-admin.c | 114 +++++++++++++++++++++ src/libvirt_admin_private.syms | 5 + src/libvirt_admin_public.syms | 3 + src/libvirt_private.syms | 4 + src/rpc/virnetserver.c | 37 +++++++ src/rpc/virnetserver.h | 13 +++ src/util/virerror.c | 6 ++ src/util/virthreadpool.c | 196 ++++++++++++++++++++++++------------ src/util/virthreadpool.h | 8 ++ tools/virt-admin.c | 132 ++++++++++++++++++++++++ 22 files changed, 1212 insertions(+), 254 deletions(-) -- 2.4.3

3 13

[libvirt] [PATCH] rbd: use unsigned long long instead of uint64_t
by Shanzhi Yu 29 Mar '16

29 Mar '16

Fix below error: storage/storage_backend_rbd.c: In function 'virStorageBackendRBDSetAllocation': storage/storage_backend_rbd.c:337:5: error: format '%llu' expects argument of type 'long long unsigned int', but argument 8 has type 'uint64_t' [-Werror=format=] VIR_DEBUG("Found %llu bytes allocated for RBD image %s", Signed-off-by: Shanzhi Yu <shyu(a)redhat.com> --- src/storage/storage_backend_rbd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/storage/storage_backend_rbd.c b/src/storage/storage_backend_rbd.c index c9b47e2..dec0905 100644 --- a/src/storage/storage_backend_rbd.c +++ b/src/storage/storage_backend_rbd.c @@ -324,7 +324,7 @@ virStorageBackendRBDSetAllocation(virStorageVolDefPtr vol, rbd_image_info_t *info) { int r, ret = -1; - uint64_t allocation = 0; + unsigned long long allocation = 0; if ((r = rbd_diff_iterate2(image, NULL, 0, info->size, 0, 1, &virStorageBackendRBDRefreshVolInfoCb, -- 1.8.3.1

2 1

[libvirt] [PATCH] storage: rbd: Fix build
by Peter Krempa 29 Mar '16

29 Mar '16

After the recent commits the build didn't work for me. Fix it by using size_t as the callback argument is using and the correct formatter. The attempted fixup to use %llu as a formatter was wrong. --- Pushed now. src/storage/storage_backend_rbd.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/storage/storage_backend_rbd.c b/src/storage/storage_backend_rbd.c index c9b47e2..6e92ff7 100644 --- a/src/storage/storage_backend_rbd.c +++ b/src/storage/storage_backend_rbd.c @@ -311,7 +311,7 @@ virStorageBackendRBDRefreshVolInfoCb(uint64_t offset ATTRIBUTE_UNUSED, int exists, void *arg) { - uint64_t *used_size = (uint64_t *)(arg); + size_t *used_size = (size_t *)(arg); if (exists) (*used_size) += len; @@ -324,7 +324,7 @@ virStorageBackendRBDSetAllocation(virStorageVolDefPtr vol, rbd_image_info_t *info) { int r, ret = -1; - uint64_t allocation = 0; + size_t allocation = 0; if ((r = rbd_diff_iterate2(image, NULL, 0, info->size, 0, 1, &virStorageBackendRBDRefreshVolInfoCb, @@ -334,7 +334,7 @@ virStorageBackendRBDSetAllocation(virStorageVolDefPtr vol, goto cleanup; } - VIR_DEBUG("Found %llu bytes allocated for RBD image %s", + VIR_DEBUG("Found %zu bytes allocated for RBD image %s", allocation, vol->name); vol->target.allocation = allocation; -- 2.7.3

1 0

[libvirt] [PATCH] util: generate correct macvtap name
by Shanzhi Yu 29 Mar '16

29 Mar '16

in commit 370608b, bitmap of in-used macvtap devices was introduced. if there is already macvtap device created not by libvirt, virNetDevMacVLanCreateWithVPortProfile will failed after try MACVLAN_MAX_ID times call virNetDevMacVLanReleaseID Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1321546 Signed-off-by: Shanzhi Yu <shyu(a)redhat.com> --- src/util/virnetdevmacvlan.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/src/util/virnetdevmacvlan.c b/src/util/virnetdevmacvlan.c index 2409113..973363e 100644 --- a/src/util/virnetdevmacvlan.c +++ b/src/util/virnetdevmacvlan.c @@ -988,7 +988,8 @@ virNetDevMacVLanCreateWithVPortProfile(const char *ifnameRequested, MACVTAP_NAME_PREFIX : MACVLAN_NAME_PREFIX; const char *pattern = (flags & VIR_NETDEV_MACVLAN_CREATE_WITH_TAP) ? MACVTAP_NAME_PATTERN : MACVLAN_NAME_PATTERN; - int rc, reservedID = -1; + int rc = -1; + int reservedID = 0; char ifname[IFNAMSIZ]; int retries, do_retry = 0; uint32_t macvtapMode; @@ -1072,16 +1073,17 @@ virNetDevMacVLanCreateWithVPortProfile(const char *ifnameRequested, retries = MACVLAN_MAX_ID; while (!ifnameCreated && retries) { virMutexLock(&virNetDevMacVLanCreateMutex); - reservedID = virNetDevMacVLanReserveID(reservedID, flags, false, true); - if (reservedID < 0) { - virMutexUnlock(&virNetDevMacVLanCreateMutex); - return -1; - } snprintf(ifname, sizeof(ifname), pattern, reservedID); rc = virNetDevMacVLanCreate(ifname, type, macaddress, linkdev, macvtapMode, &do_retry); if (rc < 0) { - virNetDevMacVLanReleaseID(reservedID, flags); + reservedID++; + reservedID = virNetDevMacVLanReserveID(reservedID, flags, false, true); + if (reservedID < 0) { + virMutexUnlock(&virNetDevMacVLanCreateMutex); + return -1; + } + virMutexUnlock(&virNetDevMacVLanCreateMutex); if (!do_retry) return -1; -- 1.8.3.1

2 3

[libvirt] [PATCH 0/3] libxl: improve libxlDomainStart error handling
by Jim Fehlig 29 Mar '16

29 Mar '16

libxlDomainStart allocates and reserves resources, which should be cleaned up on error paths. This small series is based on Chunyan's patch to do the same [0]. It breaks the patch into three patches to ease review and better understand the changes. [0] https://www.redhat.com/archives/libvir-list/2016-March/msg00929.html Chunyan Liu (1): libxl: fix resource leaks in libxlDomainStart error paths Jim Fehlig (2): libxl: rename cleanup_dom label libxl: only disable domain death events in libxlDomainCleanup src/libxl/libxl_domain.c | 44 ++++++++++++++++++++------------------------ 1 file changed, 20 insertions(+), 24 deletions(-) -- 2.1.4

2 6

[libvirt] [PATCH] libxl: remove reference to non-existent out label
by Jim Fehlig 28 Mar '16

28 Mar '16

Commit e6336442 changed the 'out:' label to 'cleanup' in libxlDomainAttachNetDevice(), but missed a comment referencing the 'out:' label. Remove it from the comment since it is no longer accurate anyhow. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- Pushing as trivial. src/libxl/libxl_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c index cfc4555..5103495 100644 --- a/src/libxl/libxl_driver.c +++ b/src/libxl/libxl_driver.c @@ -3159,7 +3159,7 @@ libxlDomainAttachNetDevice(libxlDriverPrivatePtr driver, /* This is really a "smart hostdev", so it should be attached * as a hostdev (the hostdev code will reach over into the * netdev-specific code as appropriate), then also added to - * the nets list (see out:) if successful. + * the nets list if successful. */ ret = libxlDomainAttachHostDevice(driver, vm, hostdev); goto cleanup; -- 2.1.4

1 0