- Devel - libvirt List Archives

[libvirt] [PATCHv2 0/3] tighten up integer parsing
by Eric Blake 20 Apr '12

20 Apr '12

v1: https://www.redhat.com/archives/libvir-list/2012-April/msg00954.html changes in v2: fix regression in patch 1, incomplete change in patch 2 Eric Blake (3): virsh: avoid strtol conf: tighten up XML integer parsing build: avoid strtol and strtod cfg.mk | 14 ++++++++++ src/conf/domain_conf.c | 24 ++++++++++------- src/conf/storage_conf.c | 7 +++-- src/util/util.c | 14 +++++----- src/util/virmacaddr.c | 2 +- src/util/xml.c | 36 +++---------------------- tools/virsh.c | 67 ++++++++++++++--------------------------------- 7 files changed, 64 insertions(+), 100 deletions(-) -- 1.7.7.6

2 7

[libvirt] autogen.sh broken on RHEL5
by Daniel P. Berrange 20 Apr '12

20 Apr '12

$ ./autogen.sh --system Running ./configure with --prefix=/usr --sysconfdir=/etc --localstatedir=/var --libdir=/usr/lib64 running bootstrap... ./bootstrap: Bootstrapping from checked-out http://libvirt.org sources... ./bootstrap: consider installing git-merge-changelog from gnulib ./bootstrap: getting gnulib files... ./bootstrap: autopoint --force Copying file build-aux/config.rpath running: libtoolize --copy --install libtoolize: unrecognized option `--install' Try `libtoolize --help' for more information. There is no such '--install' option - is that even needed - AFAIK that is the default behaviour ? 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 :|

2 2

[libvirt] [PATCH] Improve on virAtomic implementation
by Stefan Berger 19 Apr '12

19 Apr '12

This patch improves on the previously added virAtomicInt operations by testing for the compiler and if GCC >= 4.1 (not found in docs prior to that) is used on Linux and has the appropriate processor (that I have access to) then use the implementation based on the gcc-builtins. I also did not look at other systems (cygwin, win32) that do not need access to virAtomic right now. --- src/util/viratomic.h | 88 +++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 72 insertions(+), 16 deletions(-) Index: libvirt-acl/src/util/viratomic.h =================================================================== --- libvirt-acl.orig/src/util/viratomic.h +++ libvirt-acl/src/util/viratomic.h @@ -30,6 +30,22 @@ typedef struct _virAtomicInt virAtomicInt; typedef virAtomicInt *virAtomicIntPtr; +# define __VIR_ATOMIC_USES_LOCK + +# if defined(__GNUC__) +# if ((__GNUC__ == 4) && (__GNUC_MINOR__ >= 1)) || (__GNUC__ > 4) +# if defined(__linux__) +# if defined(__i386__) || defined(__x86_64__) || \ + defined(__powerpc64__) || defined(__powerpc__) +# undef __VIR_ATOMIC_USES_LOCK +# endif +# endif +# endif +# endif + + +# ifdef __VIR_ATOMIC_USES_LOCK + struct _virAtomicInt { virMutex lock; int value; @@ -42,22 +58,6 @@ virAtomicIntInit(virAtomicIntPtr vaip) return virMutexInit(&vaip->lock); } -static inline void -virAtomicIntSet(virAtomicIntPtr vaip, int value) -{ - virMutexLock(&vaip->lock); - - vaip->value = value; - - virMutexUnlock(&vaip->lock); -} - -static inline int -virAtomicIntRead(virAtomicIntPtr vaip) -{ - return vaip->value; -} - static inline int virAtomicIntAdd(virAtomicIntPtr vaip, int add) { @@ -88,4 +88,60 @@ virAtomicIntSub(virAtomicIntPtr vaip, in return ret; } +# else /* __VIR_ATOMIC_USES_LOCK */ + +struct _virAtomicInt { + int value; +}; + +static inline int +virAtomicIntInit(virAtomicIntPtr vaip) +{ + vaip->value = 0; + return 0; +} + +static inline int +virAtomicIntAdd(virAtomicIntPtr vaip, int add) +{ + return __sync_add_and_fetch(&vaip->value, add); +} + +static inline int +virAtomicIntSub(virAtomicIntPtr vaip, int sub) +{ + return __sync_sub_and_fetch(&vaip->value, sub); +} + +# endif /* __VIR_ATOMIC_USES_LOCK */ + + + +/* common operations that need no locking or build on others */ + + +static inline void +virAtomicIntSet(virAtomicIntPtr vaip, int value) +{ + vaip->value = value; +} + +static inline int +virAtomicIntRead(virAtomicIntPtr vaip) +{ + return *(volatile int *)&vaip->value; +} + +static inline int +virAtomicIntInc(virAtomicIntPtr vaip) +{ + return virAtomicIntAdd(vaip, 1); +} + +static inline int +virAtomicIntDec(virAtomicIntPtr vaip) +{ + return virAtomicIntSub(vaip, 1); +} + #endif /* __VIR_ATOMIC_H */

2 1

[libvirt] [PATCH 0/2 v3 use qemu's dump-guest-meory when vm uses host device
by Wen Congyang 19 Apr '12

19 Apr '12

Currently, we use migrate to dump guest's memory. There is one restriction in migrate command: the device's status should be stored in qemu because the device's status should be passed to target machine. If we passthrough a host device to guest, the device's status is stored in the real device. So migrate command will fail. We usually use dump when guest is panicked. So there is no need to store device's status in the vmcore. qemu will have a new monitor command dump-guest-memory to dump guest memory, but it doesn't support async now(it will support later when the common async API is implemented). So I use dump-guest-memory only when the guest uses host device in this patchset. Note: the patchset for qemu is still queued. Luiz has acked, but he waits an ACK from Jan and/or Anthony. They are too busy, and donot reply. Changes from v2 to v3: 1. qemu supports the fd that is associated with a pipe, socket, or FIFO. So pass a pipe fd to qemu and O_DIRECT can work now. Change from v1 to v2: 1. remove the implemention for text mode. Wen Congyang (2): qemu: implement qemu's dump-guest-memory qemu: try to use qemu's dump-guest-meory when vm uses host device src/qemu/qemu_domain.c | 1 + src/qemu/qemu_domain.h | 1 + src/qemu/qemu_driver.c | 39 +++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_monitor.c | 36 ++++++++++++++++++++++++++++++++++++ src/qemu/qemu_monitor.h | 13 +++++++++++++ src/qemu/qemu_monitor_json.c | 42 ++++++++++++++++++++++++++++++++++++++++++ src/qemu/qemu_monitor_json.h | 7 +++++++ 7 files changed, 137 insertions(+), 2 deletions(-)

3 5

[libvirt] [PATCH V11 0/7] Add DHCP snooping support to nwfilter
by Stefan Berger 19 Apr '12

19 Apr '12

This series of patches adds DHCP snooping support to libvirt's nwfilter subsystem. DHCP snooping detects DHCP leases obtained by a VM and automatically adjusts the network traffic filters to reflect the IP addresses with which a VM may send its traffic, thus for example preventing IP address spoofing. Once leases on IP addresses expire or if a VM gives up on a lease on an IP address, the filters are also adjusted. All leases are persisted and automatically applied upon a VM's restart. Leases are associated with the tuple of VM-UUID and interface MAC address. The following interface XML activates and uses the DHCP snooping: <interface type='bridge'> <source bridge='virbr0'/> <filterref filter='clean-traffic'> <parameter name='ip_learning' value='dhcp'/> </filterref> </interface> Once an IP address has been detected on an interface, 'virsh dumpxml <vm>' would show the IP address lease in the format <IP address>,<lease timeout in seconds>: <interface type='bridge'> <source bridge='virbr0'/> <filterref filter='clean-traffic'> <parameter name='ip_learning' value='dhcp'/> <parameter name='IP_LEASE' value='192.168.122.210,180'/> </filterref> </interface> Regards, David and Stefan

3 20

[libvirt] [PATCH v3 0/9] Add basic driver for Parallels Virtuozzo Server
by Dmitry Guryanov 19 Apr '12

19 Apr '12

Parallels Virtuozzo Server is a cloud-ready virtualization solution that allows users to simultaneously run multiple virtual machines and containers on the same physical server. Current name of this product is Parallels Server Bare Metal and more information about it can be found here - http://www.parallels.com/products/server/baremetal/sp/. This driver will work with PVS version 6.0 , beta version scheduled at 2012 Q2. Dmitry Guryanov (9): pvs: add driver skeleton util: add functions for interating over json object pvs: add functions to list domains and get info pvs: implement functions for domain life cycle management pvs: get info about serial ports pvs: add support of VNC remote display pvs: implement virDomainDefineXML operation for existing domains pvs: add storage driver pvs: implement VM creation cfg.mk | 1 + configure.ac | 23 + docs/drvpvs.html.in | 28 + include/libvirt/virterror.h | 1 + libvirt.spec.in | 7 + mingw32-libvirt.spec.in | 6 + po/POTFILES.in | 1 + src/Makefile.am | 23 + src/conf/domain_conf.c | 3 +- src/conf/domain_conf.h | 1 + src/driver.h | 1 + src/libvirt.c | 12 + src/pvs/pvs_driver.c | 1281 +++++++++++++++++++++++++++++++++++++ src/pvs/pvs_driver.h | 75 +++ src/pvs/pvs_storage.c | 1458 +++++++++++++++++++++++++++++++++++++++++++ src/pvs/pvs_utils.c | 143 +++++ src/util/json.c | 30 + src/util/json.h | 4 + src/util/virterror.c | 3 + 19 files changed, 3100 insertions(+), 1 deletions(-) create mode 100644 docs/drvpvs.html.in create mode 100644 src/pvs/pvs_driver.c create mode 100644 src/pvs/pvs_driver.h create mode 100644 src/pvs/pvs_storage.c create mode 100644 src/pvs/pvs_utils.c

3 12

[libvirt] [PATCH 0/3] tighten up integer parsing
by Eric Blake 19 Apr '12

19 Apr '12

I thought my recent improvements to <memory> parsing were complete, until I stumbled upon https://bugzilla.redhat.com/show_bug.cgi?id=617711, which shows a file that fails 'virt-xml-validate' but was silently accepted, and worse, misintepreted the user's intentions. I didn't have time to scrub the entire code base, but I can at least prevent the clean files from getting worse. Eric Blake (3): virsh: avoid strtol conf: tighten up XML integer parsing build: avoid strtol and strtod cfg.mk | 14 ++++++++++ src/conf/domain_conf.c | 12 +++----- src/conf/storage_conf.c | 6 ++-- src/util/util.c | 14 +++++----- src/util/virmacaddr.c | 2 +- src/util/xml.c | 36 +++---------------------- tools/virsh.c | 66 +++++++++++++--------------------------------- 7 files changed, 53 insertions(+), 97 deletions(-) -- 1.7.7.6

2 8

[libvirt] Problem with USB pass-through
by Carlos Rodrigues 19 Apr '12

19 Apr '12

Hello, I have a problem with USB pass-through. I can see the USB device in host machine, i see it in libvirt but when attach to VM i get an error: # lsusb ... Bus 002 Device 014: ID 0529:0001 Aladdin Knowledge Systems HASP v0.06 ... # virsh nodedev-list ... usb_device_529_1_noserial usb_device_529_1_noserial_if0 usb_device_529_1_noserial_usbraw ... # virsh dumpxml vm ... <hostdev mode='subsystem' type='usb' managed='no'> <source> <vendor id='0x0529'/> <product id='0x0001'/> </source> </hostdev> ... # cat /var/log/libvirt/qemu/vm.log ... LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin HOME=/ QEMU_AUDIO_DRV=none /usr/bin/qemu-kvm -S -M pc -m 3800 -smp 4 -name vm -uuid 056dad8b-22a4-4472-cb5e-8cf60215a41b -monitor unix:/var/lib/libvirt/qemu/v,.monitor,server,nowait -boot c -drive file=/dev/vg_sata2/vol_vm,if=virtio,boot=on,format=raw -net nic,macaddr=00:00:00:0b:5b:ec,vlan=0,model=virtio,name=net0 -net tap,fd=21,vlan=0,name=hostnet0 -serial none -parallel none -usb -usbdevice tablet -vnc 0.0.0.0:4 -k pt -vga cirrus -usbdevice host:002.014 -balloon virtio husb: open device 2.12 /proc/bus/usb/002/012: No such file or directory Warning: could not add USB device host:002.014 ... I'm using libvirt 0.8.3 with kvm 88. Anyone have see this problem? Regards, -- Carlos Rodrigues <cmar(a)eurotux.com> Investigação e Desenvolvimento Eurotux Informática, S.A. [http://eurotux.com]

2 2

[libvirt] [PATCH 0/2 v2] use qemu's dump-guest-meory when vm uses host device
by Wen Congyang 19 Apr '12

19 Apr '12

Currently, we use migrate to dump guest's memory. There is one restriction in migrate command: the device's status should be stored in qemu because the device's status should be passed to target machine. If we passthrough a host device to guest, the device's status is stored in the real device. So migrate command will fail. We usually use dump when guest is panicked. So there is no need to store device's status in the vmcore. qemu will have a new monitor command dump-guest-memory to dump guest memory, but it doesn't support async now(it will support later when the common async API is implemented). So I use dump-guest-memory only when the guest uses host device in this patchset. Note: the patchset for qemu is still queued. Luiz has acked, but he waits an ACK from Jan and/or Anthony. They are too busy, and donot reply. Change from v1 to v2: 1. remove the implemention for text mode. Wen Congyang (2): qemu: implement qemu's dump-guest-memory qemu: try to use qemu's dump-guest-meory when vm uses host device src/qemu/qemu_domain.c | 1 + src/qemu/qemu_domain.h | 1 + src/qemu/qemu_driver.c | 42 ++++++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_monitor.c | 36 ++++++++++++++++++++++++++++++++++++ src/qemu/qemu_monitor.h | 13 +++++++++++++ src/qemu/qemu_monitor_json.c | 42 ++++++++++++++++++++++++++++++++++++++++++ src/qemu/qemu_monitor_json.h | 7 +++++++ 7 files changed, 140 insertions(+), 2 deletions(-)

2 4

[libvirt] Making the Thread-safety issues with vbox driver ?
by Jean-Baptiste Rouault 19 Apr '12

19 Apr '12

On Wednesday 04 April 2012 17:56:12 Jean-Baptiste Rouault wrote: > On Monday 16 January 2012 11:34:53 Matthias Bolte wrote: > > Okay, without looking deeper into this here are some ideas: > > > > The XPCOM API libvirt uses might not be threadsafe, or needs to be > > initialized for every thread that wants to use it. Currently its only > > initialized for the thread that opens the driver. I know that this is > > the case on Windows were VirtualBox uses MSCOM for its API and you > > need to call CoInitialize on every thread. This is currently not done > > for the MSCOM glue in libvirt, so I know that on Windows the > > VirtualBox driver is not threadsafe currently. Also I didn't look into > > a solution for this yet. Maybe we need a thread local variable that > > holds whether (MS/XP)COM was already initialized for this thread and > > add a check to every driver function to initialize it when needed. > > > > Did you try to open a connection for each thread instead of trying to > > share one? If that works reliable it might indicate that there is an > > VirtualBox API initialization problem. > > I tried today with one connection for each thread and it works. > > I changed the vbox driver so that the pfnComInitialize function is called > only when the first connection is opened : it breaks the test, even with > one connection per thread. > > I guess we'll have to use a thread local variable as you suggested, unless > someone has a better idea to handle this problem. Hi, I looked deeper into these thread-safety issues, once a new connection is opened for each thread, everything works well. However, opening and closing connections isn't thread-safe at all for two reasons : - VirtualBox C bindings initialization and uninitialization functions aren't thread-safe. I talked about it with upstream on IRC and they are probably not going to fix it, but would accept a patch fixing the issue. I'm going to contact upstream again to get some advices so I can write a patch. - In the libvirt vbox driver, for each new connection, modification of the global variable g_pVBoxGlobalData isn't protected (see line 1040 of vbox_tmpl.c). First of all, is it really necessary to replace it on each new connection, or would it be ok to initialize it only when the first connection is opened ? A global mutex is needed in the vbox driver to protect access to g_pVBoxGlobalData, the vboxRegister() function seems to be the best place to initialize such a mutex unless there is another entry point to do this ? -- Jean-Baptiste ROUAULT Ingénieur R&D - diateam : Architectes de l'information Phone : +33 (0)2 98 050 050 Fax : +33 (0)2 98 050 051

3 3