- Devel - libvirt List Archives

[libvirt] [BUG] Two different connections to the same host through different transport
by Емил Гелев 08 Aug '14

08 Aug '14

Hi. I am using libvirt for some time and I have found the following problem. When I try opening two connections with the same credentials to the same host but using different transport(TCP and TLS) I get the following error when trying to free the TLS one: "*** Error in `./test': double free or corruption (!prev): 0x00000000014a7a60 *** Aborted (core dumped)" Example code: #include <stdio.h> #include <stdlib.h> #include <string.h> #include <libvirt/libvirt.h> struct CredentialsPack { const char* username; const char* password; }; struct CredentialsPack* newCredentialsPack(const char* username, const char* password) { struct CredentialsPack* ptr = malloc(sizeof(struct CredentialsPack)); ptr->username = username; ptr->password = password; return ptr; } static int authCredTypes[] = { VIR_CRED_AUTHNAME, VIR_CRED_PASSPHRASE, }; static int authCb(virConnectCredentialPtr cred, unsigned int ncred, void* cbdata) { struct CredentialsPack* inputCreds = (struct CredentialsPack*) cbdata; int i; for (i = 0; i < ncred; ++i) { if (cred[i].type == VIR_CRED_AUTHNAME) { cred[i].result = strdup(inputCreds->username); if (cred[i].result == NULL) { return -1; } cred[i].resultlen = strlen(cred[i].result); } else if (cred[i].type == VIR_CRED_PASSPHRASE) { cred[i].result = strdup(inputCreds->password); if (cred[i].result == NULL) { return -1; } cred[i].resultlen = strlen(cred[i].result); } } return 0; } int main(){ struct CredentialsPack* credentials = newCredentialsPack("user", "pass"); virConnectAuth auth; auth.credtype = authCredTypes; auth.ncredtype = sizeof(authCredTypes) / sizeof(int); auth.cb = authCb; auth.cbdata = credentials; virConnectPtr conn; conn = virConnectOpenAuth("qemu+tcp://host1/system", &auth, 0); struct CredentialsPack* credentials2 = newCredentialsPack("user", "pass"); virConnectAuth auth2; auth2.credtype = authCredTypes; auth2.ncredtype = sizeof(authCredTypes) / sizeof(int); auth2.cb = authCb; auth2.cbdata = credentials2; virConnectPtr conn2; conn2 = virConnectOpenAuth("qemu://host1/system", &auth2, 0); // virConnectClose(conn); // this does not change the behaviour virConnectClose(conn2); // here the error occurs } If I try to close "conn" then everything works until I try to close "conn2". I can even invoke virConnectListAllDomains using "conn2" and use the domain pointers when "conn" is closed. The problem is when I try to close "conn2". The platform running this code is Ubuntu 14.04. The version of libvirt is 1.2.2 (default configuration). The certificate of the CA which signed the server certificate is under /etc/pki/CA/cacert.pem The target machine is RHEL 6.5. The version of libvirt is 0.10.2. It is configured this way: listen_tls = 1 listen_tcp = 1 auth_tcp = sasl auth_tls = sasl tls_no_verify_certificate = 1 Everything else is set to default. The certificate of the CA which signed the server certificate is under /etc/pki/CA/cacert.pem The user "user" is added to the DB( using $ saslpasswd2 -a libvirt user). I can successfully execute the following on the command line using the "user"/"pass" credentials: virsh -c qemu://host1/system list virsh -c qemu+tcp://host1/system list I hope this will be useful. Thank you.

1 0

[libvirt] virsh migrate: sometimes command line cannot exit until manually press "ENTER" key
by Chun Yan Liu 08 Aug '14

08 Aug '14

Hi, List, Do you meet the same problem? Host is kvm, doing: (e.g.) #virsh migrate sles11 qemu+ssh://147.2.207.55/system --live --unsafe password: Sometimes, VM is already running on the target host and disappears from source host, but the command line still hangs there, if pressing "ENTER", it will exit. I debugged the code, but found the result is weird. The code hangs at tools/virsh-domain.c -> cmdMigrate ->vshWatchJob->poll(): poll() is trying to select pipe_fd, which is used to receive message from doMigrate thread. In debugging, found that doMigrate finishes and at the end it does call safewrite() to write the retval ('0' or '1') to pipe_fd, and the write is completed. But cmdMigrate poll() cannot get the event. Then after pressing "ENTER" key, poll() can get the event and select pipe_fd, then command line can exit. Any ideas about the possible reason? Thanks, Chunyan

2 2

[libvirt] [PATCH 0/2] active commit fixes
by Eric Blake 08 Aug '14

08 Aug '14

In response to the valgrind trace I posted to the list last night. Valgrind is awesome :) I had a very hard time reproducing the race when run under gdb or at normal speeds, but valgrind was slow enough that I was hitting it more frequently, as well as getting information about the bug and where to go to fix it. Eric Blake (2): blockjob: avoid memory leak during block pivot blockjob: fix use-after-free in blockcopy src/qemu/qemu_driver.c | 40 +++++++++++++++++++++++++--------------- src/util/virstoragefile.c | 1 + 2 files changed, 26 insertions(+), 15 deletions(-) -- 1.9.3

2 7

Re: [libvirt] [PATCH] audit: Fix some comments
by Wang Rui 08 Aug '14

08 Aug '14

On 2014/8/7 17:07, Ján Tomko wrote: > On 08/07/2014 10:12 AM, Wang Rui wrote: >> Fix a comment in virDomainAuditNetDevice. >> Fix a typo in comment of qemuPhysIfaceConnect which is >> the caller of virDomainAuditNetDevice. >> >> Signed-off-by: Wang Rui <moon.wangrui(a)huawei.com> >> --- >> src/conf/domain_audit.c | 4 ++-- >> src/qemu/qemu_command.c | 2 +- >> 2 files changed, 3 insertions(+), 3 deletions(-) >> > > ACK and pushed. Thanks. I have a question about comment style. I see there are two comment styles in our libvirt code. One is /* XXXX * XXXX */ The other is /* XXXX * XXXX */ Are both of them fine ? Or which one is recommended ? (I'm not sure about my comments when coding.)

2 1

[libvirt] [PATCH v3 0/3] Add iSCSI hostdev pass-through device
by John Ferlan 07 Aug '14

07 Aug '14

Patches 1 & 2 are related to: Update to http://www.redhat.com/archives/libvir-list/2014-July/msg01098.html Patch 1/3 is already ACK'd, but felt it was easier to push together. Patch 2/3 change since last: - Fix typos as called in review of 8/8 - Adjust the rng file to - Change the order of things to be <source>, then <auth> to follow how <disk> (and storage pools) do things. That resulted in code motion in domain_conf.c and adjustment to the *-iscsi-auth.xml files Patch 3 is related to http://www.redhat.com/archives/libvir-list/2014-July/msg01268.html John Ferlan (3): domain_conf: Common routine to handle network storage host xml def hostdev: Add iSCSI hostdev XML qemu: Remove extraneous space in function prototypes docs/formatdomain.html.in | 144 ++++++---- docs/schemas/domaincommon.rng | 48 +++- src/conf/domain_conf.c | 289 +++++++++++++++------ src/qemu/qemu_capabilities.h | 2 +- src/qemu/qemu_command.h | 136 +++++----- src/qemu/qemu_conf.h | 2 +- .../qemuxml2argv-hostdev-scsi-lsi-iscsi-auth.args | 14 + .../qemuxml2argv-hostdev-scsi-lsi-iscsi-auth.xml | 46 ++++ .../qemuxml2argv-hostdev-scsi-lsi-iscsi.args | 14 + .../qemuxml2argv-hostdev-scsi-lsi-iscsi.xml | 40 +++ ...emuxml2argv-hostdev-scsi-virtio-iscsi-auth.args | 16 ++ ...qemuxml2argv-hostdev-scsi-virtio-iscsi-auth.xml | 46 ++++ .../qemuxml2argv-hostdev-scsi-virtio-iscsi.args | 16 ++ .../qemuxml2argv-hostdev-scsi-virtio-iscsi.xml | 40 +++ tests/qemuxml2argvtest.c | 16 ++ tests/qemuxml2xmltest.c | 5 + 16 files changed, 672 insertions(+), 202 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-lsi-iscsi-auth.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-lsi-iscsi-auth.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-lsi-iscsi.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-lsi-iscsi.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-virtio-iscsi-auth.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-virtio-iscsi-auth.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-virtio-iscsi.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hostdev-scsi-virtio-iscsi.xml -- 1.9.3

2 8

[libvirt] File permissions on VNC unix sockets
by Landon Gilbert-Bland 07 Aug '14

07 Aug '14

I am using kvm libvirt/qemu, and am trying to get VNC working with unix sockets. I'm using the following in my XML: <graphics type='vnc' socket='/tmp/kvmtest'/> This works, it creates the unix socket and I can use it for VNC. But it creates the socket with 755 permissions, and owned by libvirt-qemu:kvm. Ideally, I would like it to be 770 root:libvirtd, but could also work with 775 libvirt-qemu:kvm. Basically, I would like the group to be read/write/execute, so that anyone in this group can use virt-manager (or whatever) to get into the vm with vnc. I haven't found a way to change this in the settins. It doesn't seem to be honoring the unix socket settings in /etc/libvirt/libvirdd.conf (it looks like those are only for libvirt-sock and libvirt-sock-ro), and I haven't been able to find any documentation about changing these permissions in the XML. I can manually chmod/chown the socket after it is created (manually or with a cron), but that is far from an ideal solution. I don't suppose anyone has run across a configuration option for this that I have missed? I have only tested this in ubuntu 14.01 and debian sid, for what it's worth. Thanks, --landon

1 1

[libvirt] Error
by ahmad jan 07 Aug '14

07 Aug '14

I am using Libvirt-1.0.0 on ubuntu 14.02 LTS and found the following error "checking linux/if_bridge.h usability... no checking linux/if_bridge.h presence... yes configure: WARNING: linux/if_bridge.h: present but cannot be compiled configure: WARNING: linux/if_bridge.h: check for missing prerequisite headers? configure: WARNING: linux/if_bridge.h: see the Autoconf documentation configure: WARNING: linux/if_bridge.h: section "Present But Cannot Be Compiled" configure: WARNING: linux/if_bridge.h: proceeding with the compiler's result configure: WARNING: ## ------------------------------------- ## configure: WARNING: ## Report this to libvir-list <at> redhat.com ## configure: WARNING: ## ------------------------------------- ## checking for linux/if_bridge.h... no configure: error: You must install kernel-headers in order to compile libvirt with QEMU or LXC support" Please Guide, how do I recover it thanks

3 2

[libvirt] Help Required
by ahmad jan 07 Aug '14

07 Aug '14

Respected Sir i am using libvirt-1.2.6 API and facing the following error please guide libvirtd: /usr/local/lib/libvirt.so.0: version `LIBVIRT_PRIVATE_1.2.6' not found (required by libvirtd) best regrads

2 1

[libvirt] [PATCH] virsh: clean up attach-interface paragraph in man page
by Laine Stump 07 Aug '14

07 Aug '14

This makes the paragaph about attach-interface more descriptive and correct, adding in a few bits of information that were previously missing, e.g. --script is only allowed for bridge interfaces of Xen domains, target name is regenerated if it starts with vnet, mac address will be autogenerated if not specified. (I did this in response to an email asking why a script couldn't be specified for a bridge interface of a qemu domain, and why an interface of type='ethernet' couldn't be created with attach-interface) --- tools/virsh.pod | 36 +++++++++++++++++++++++------------- 1 file changed, 23 insertions(+), 13 deletions(-) diff --git a/tools/virsh.pod b/tools/virsh.pod index 849ae31..abb8743 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -2202,19 +2202,29 @@ Likewise, I<--shareable> is an alias for I<--mode shareable>. [I<--target target>] [I<--mac mac>] [I<--script script>] [I<--model model>] [I<--config>] [I<--inbound average,peak,burst>] [I<--outbound average,peak,burst>] -Attach a new network interface to the domain. I<type> can be either I<network> -to indicate a physical network device or I<bridge> to indicate a bridge to a -device. I<source> indicates the source device. I<target> allows to indicate -the target device in the guest. Names starting with 'vnet' are considered as -auto-generated an hence blanked out. I<mac> allows to specify the MAC address -of the network interface. I<script> allows to specify a path to a script -handling a bridge instead of the default one. I<model> allows to specify the -model type. I<inbound> and I<outbound> control the bandwidth of the interface. -I<peak> and I<burst> are optional, so "average,peak", "average,,burst" and -"average" are also legal. Values for I<average> and I<peak> are -expressed in kilobytes per second, while I<burst> is expressed in kilobytes -in a single burst at -I<peak> speed as described in the Network XML -documentation at L<http://libvirt.org/formatnetwork.html#elementQoS>. +Attach a new network interface to the domain. I<type> can be either +I<network> to indicate connection via a libvirt virtual network or +I<bridge> to indicate connection via a bridge device on the host. +I<source> indicates the source of the connection (either the name of a +network, or of a bridge device). I<target> is used to specify the +tap/macvtap device to be used to connect the domain to the +source. Names starting with 'vnet' are considered as auto-generated +and are blanked out/regenerated each time the interface is attached. +I<mac> specifies the MAC address of the network interface; if a MAC +address is not given, a new address will be automatically generated +(and stored in the persistent configuration if "--config" is given on +the commandline). I<script> is used to specify a path to a custom +script to be called while attaching to a bridge - this will be called +instead of the default script not in addition to it; --script is valid +only for interfaces of type I<bridge> and only for Xen domains. +I<model> specifies the network device model to be presented to the +domain. I<inbound> and I<outbound> control the bandwidth of the +interface. I<peak> and I<burst> are optional, so "average,peak", +"average,,burst" and "average" are also legal. Values for I<average> +and I<peak> are expressed in kilobytes per second, while I<burst> is +expressed in kilobytes in a single burst at -I<peak> speed as +described in the Network XML documentation at +L<http://libvirt.org/formatnetwork.html#elementQoS>. If I<--live> is specified, affect a running domain. If I<--config> is specified, affect the next startup of a persistent domain. -- 1.9.3

2 1

[libvirt] Can't run vbox's network driver and storage driver.
by Taowei Luo 07 Aug '14

07 Aug '14

Hi, I just can't find a way running the network driver and storage driver provided by vbox. When I use vbox:///session as an url to run vbox driver and call network (or storage) API, the vbox's driver didn't execute. Instead, it would run drivers in src/network and src/storage. The problem only comes with the network driver and storage driver. Which means the vbox's domain driver works well with libvirt. I know that actually libvirt communicate with vbox driver though the remote driver, but I don't know how to make remote driver call vbox's network driver(or storage driver). BTW, I need vbox's driver because I have to do some test on it. Thanks, Taowei

1 0