December 2016 - Devel - libvirt List Archives

[libvirt] [PATCH 0/6] qemu: blockjob: make automatic disk locking less useless
by Peter Krempa 10 Jan '17

10 Jan '17

If automatic disk locking would be enabled with either lock manager the operations would most probably fail. This was mostly as libvirt did not bother to release images that were no longer written to by qemu, or tried re-lock the same image. This fixes the locking code for creation of external snapshots (both disk-only and with memory), block copy and active block commit. This series also kills the horrible "rollback" code in the external disk snapshot code. Regular block commit work will be tracked by: https://bugzilla.redhat.com/show_bug.cgi?id=1405537 Peter Krempa (6): qemu: blockcopy: Save monitor error prior to calling into lock manager locking: Fix documentation on how automatic sanlock leases are stored qemu: snapshot: Don't redetect backing chain after snapshot qemu: snapshot: Refactor snapshot rollback on failure qemu: snapshot: Properly handle image locking qemu: blockjob: Fix locking of block copy/active block commit src/locking/sanlock.conf | 2 +- src/qemu/qemu_blockjob.c | 26 ++-- src/qemu/qemu_driver.c | 389 +++++++++++++++++++++++++++-------------------- 3 files changed, 242 insertions(+), 175 deletions(-) -- 2.11.0

3 14

[libvirt] [PATCH v3 0/3] qemu: Use virtio-pci by default for mach-virt guests
by Andrea Bolognani 10 Jan '17

10 Jan '17

Changes from [v2]: * rename qemuDomainCountVirtioMMIODevices() to qemuDomainHasVirtioMMIODevices() and make it exit as soon as the first virtio-mmio device is encountered, as suggested by Laine * tweak test suite and note no new test cases are needed * add comments and user documentation * add release notes entry * can actually be merged now that all patches it builds on have been merged :) Changes from [v1]: * use virDomainDeviceInfoIterate(), as suggested by Martin and Laine, which results in cleaner and more robust code [v1] https://www.redhat.com/archives/libvir-list/2016-October/msg00988.html [v2] https://www.redhat.com/archives/libvir-list/2016-October/msg01042.html Andrea Bolognani (3): qemu: Use virtio-pci by default for mach-virt guests docs: Document virtio-mmio by default for mach-virt guests NEWS: Update for virtio-pci by default for mach-virt guests docs/formatdomain.html.in | 8 +++- docs/news.html.in | 6 +++ src/qemu/qemu_domain_address.c | 51 ++++++++++++++++++++-- ...l2argv-aarch64-virt-2.6-virtio-pci-default.args | 14 +++--- .../qemuxml2argv-aarch64-virtio-pci-default.args | 17 +++++--- .../qemuxml2argv-aarch64-virtio-pci-default.xml | 3 -- tests/qemuxml2argvtest.c | 1 + .../qemuxml2xmlout-aarch64-virtio-pci-default.xml | 40 ++++++++++++++--- tests/qemuxml2xmltest.c | 1 + 9 files changed, 119 insertions(+), 22 deletions(-) -- 2.7.4

2 8

[libvirt] [V2]RFC for support cache tune in libvirt
by Qiao, Liyong 10 Jan '17

10 Jan '17

Hi folks I’v send the V1 version RFC early this week, but no reply yet. Thanks Qiaowei for the comment, I’v change the RFC much more libvirt specify, please help to comments on. ##Propose Changes #Libvirtd configure changes Add a new configure option cache_allocation_ratio to libvirtd, which let libvirt to allocate how many cache to domains. Default is 0.5 Eg. On a host which has 55M cache, libvirt can allocate 55M * cache_allocation_ratio cache to domains ## Virsh command line changes: NAME cachetune - control or query domain cache allocation SYNOPSIS cachetune <domain> [--enabled true/false] [--type <type>][--size <number>] [--config] [--live] [--current] DESCRIPTION Allocate cache usage for domain. OPTIONS [--domain] <string> domain name, id or uuid --enabled <true/false> enable cache allocation --type <string> cache allocations type, support l3/l2 etc --size <number> the cache size in KB --config affect next boot --live affect running domain --current affect current domain This will allow libvirt to allocate specify type l3 cache for a domain Domain xml changes: <cachetune> <enabled=’yes’, size=4096, actual_size=4680,type=’l3’> <enabled=’no’, size=256, actual_size=0, type=’l2’> </cachetune> For more information about the detail design, please refer https://www.redhat.com/archives/libvir-list/2016-December/msg01011.html CAT intro: https://software.intel.com/en-us/articles/software-enabling-for-cache-alloc… Best Regards Eli Qiao(乔立勇）OpenStack Core team OTC Intel. --

2 2

[libvirt] [PATCH] util: fix domain object leaks on closecallbacks
by Wang King 10 Jan '17

10 Jan '17

From: wangjing <king.wang(a)huawei.com> The virCloseCallbacksSet method increase object reference for VM, and decrease object reference in virCloseCallbacksUnset. But VM UUID will be deleted from closecallbacks list in virCloseCallbacksRun when connection disconnected, and then object reference cannot be decreased by virCloseCallbacksUnset in callback functions. Signed-off-by: Wang King <king.wang(a)huawei.com> --- src/util/virclosecallbacks.c | 29 +++++++++++++++++------------ 1 file changed, 17 insertions(+), 12 deletions(-) diff --git a/src/util/virclosecallbacks.c b/src/util/virclosecallbacks.c index 891a92b..26d5075 100644 --- a/src/util/virclosecallbacks.c +++ b/src/util/virclosecallbacks.c @@ -300,7 +300,9 @@ virCloseCallbacksGetForConn(virCloseCallbacksPtr closeCallbacks, data.list = list; data.oom = false; + virObjectLock(closeCallbacks); virHashForEach(closeCallbacks->list, virCloseCallbacksGetOne, &data); + virObjectUnlock(closeCallbacks); if (data.oom) { VIR_FREE(list->entries); @@ -329,22 +331,15 @@ virCloseCallbacksRun(virCloseCallbacksPtr closeCallbacks, * them all from the hash. At that point we can release * the lock and run the callbacks safely. */ - virObjectLock(closeCallbacks); list = virCloseCallbacksGetForConn(closeCallbacks, conn); if (!list) return; for (i = 0; i < list->nentries; i++) { - char uuidstr[VIR_UUID_STRING_BUFLEN]; - virUUIDFormat(list->entries[i].uuid, uuidstr); - virHashRemoveEntry(closeCallbacks->list, uuidstr); - } - virObjectUnlock(closeCallbacks); - - for (i = 0; i < list->nentries; i++) { virDomainObjPtr vm; + virDomainObjPtr dom; - if (!(vm = virDomainObjListFindByUUID(domains, + if (!(vm = virDomainObjListFindByUUIDRef(domains, list->entries[i].uuid))) { char uuidstr[VIR_UUID_STRING_BUFLEN]; virUUIDFormat(list->entries[i].uuid, uuidstr); @@ -352,10 +347,20 @@ virCloseCallbacksRun(virCloseCallbacksPtr closeCallbacks, continue; } - vm = list->entries[i].callback(vm, conn, opaque); - if (vm) - virObjectUnlock(vm); + dom = list->entries[i].callback(vm, conn, opaque); + if (dom) + virObjectUnlock(dom); + virObjectUnref(vm); } + + virObjectLock(closeCallbacks); + for (i = 0; i < list->nentries; i++) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(list->entries[i].uuid, uuidstr); + virHashRemoveEntry(closeCallbacks->list, uuidstr); + } + virObjectUnlock(closeCallbacks); + VIR_FREE(list->entries); VIR_FREE(list); } -- 2.8.3

2 2

[libvirt] [PATCH v2] cmdPerf: Display enabled/disabled message on perf event enable/disable
by Nitesh Konkar 09 Jan '17

09 Jan '17

Currently no message is displayed on successful perf event enable or disable. This patch gives the enabled/disabled status message on successful enabling/disabling of the perf event. Eg:virsh perf Domain --enable instructions --disable cache_misses instructions : enabled cache_misses : disabled Signed-off-by: Nitesh Konkar <nitkon12(a)linux.vnet.ibm.com> --- tools/virsh-domain.c | 27 ++++++++++++++++++--------- 1 file changed, 18 insertions(+), 9 deletions(-) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index 3a6fa5c..287ca28 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -8848,13 +8848,27 @@ virshParseEventStr(const char *event, return ret; } +static void +virshPrintPerfStatus(vshControl *ctl, virTypedParameterPtr params, int nparams) +{ + size_t i; + + for (i = 0; i < nparams; i++) { + if (params[i].type == VIR_TYPED_PARAM_BOOLEAN && + params[i].value.b) { + vshPrint(ctl, "%-15s: %s\n", params[i].field, _("enabled")); + } else { + vshPrint(ctl, "%-15s: %s\n", params[i].field, _("disabled")); + } + } +} + static bool cmdPerf(vshControl *ctl, const vshCmd *cmd) { virDomainPtr dom; int nparams = 0; int maxparams = 0; - size_t i; virTypedParameterPtr params = NULL; bool ret = false; const char *enable = NULL, *disable = NULL; @@ -8891,18 +8905,13 @@ cmdPerf(vshControl *ctl, const vshCmd *cmd) vshError(ctl, "%s", _("Unable to get perf events")); goto cleanup; } - for (i = 0; i < nparams; i++) { - if (params[i].type == VIR_TYPED_PARAM_BOOLEAN && - params[i].value.b) { - vshPrint(ctl, "%-15s: %s\n", params[i].field, _("enabled")); - } else { - vshPrint(ctl, "%-15s: %s\n", params[i].field, _("disabled")); - } - } + virshPrintPerfStatus(ctl, params, nparams); } else { if (virDomainSetPerfEvents(dom, params, nparams, flags) != 0) { vshError(ctl, "%s", _("Unable to enable/disable perf events")); goto cleanup; + } else { + virshPrintPerfStatus(ctl, params, nparams); } } -- 1.9.3

2 1

[libvirt] [PATCH 1/2] perf: Refactor perf code
by Nitesh Konkar 09 Jan '17

09 Jan '17

Avoid unnecessary calling of function vshCommandOptStringReq. In the current code the function vshCommandOptStringReq is called irrespective of whether --enable and/or --disable is present in the command line. Eg: 'virsh perf domainName' also results in calling this function twice. This patch fixes this. Signed-off-by: Nitesh Konkar <nitkon12(a)linux.vnet.ibm.com> --- tools/virsh-domain.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index 3a6fa5c..91de532 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -8862,6 +8862,8 @@ cmdPerf(vshControl *ctl, const vshCmd *cmd) bool current = vshCommandOptBool(cmd, "current"); bool config = vshCommandOptBool(cmd, "config"); bool live = vshCommandOptBool(cmd, "live"); + bool shouldEnable = vshCommandOptBool(cmd, "enable"); + bool shouldDisable = vshCommandOptBool(cmd, "disable"); VSH_EXCLUSIVE_OPTIONS_VAR(current, live); VSH_EXCLUSIVE_OPTIONS_VAR(current, config); @@ -8871,9 +8873,15 @@ cmdPerf(vshControl *ctl, const vshCmd *cmd) if (live) flags |= VIR_DOMAIN_AFFECT_LIVE; - if (vshCommandOptStringReq(ctl, cmd, "enable", &enable) < 0 || - vshCommandOptStringReq(ctl, cmd, "disable", &disable) < 0) - return false; + if (shouldEnable) { + if (vshCommandOptStringReq(ctl, cmd, "enable", &enable) < 0) + return false; + } + + if (shouldDisable) { + if (vshCommandOptStringReq(ctl, cmd, "disable", &disable) < 0) + return false; + } if (!(dom = virshCommandOptDomain(ctl, cmd, NULL))) return false; -- 1.9.3

3 3

[libvirt] [PATCH v2 0/2] Detect misconfiguration between disk bus and disk address
by Marc Hartmayer 09 Jan '17

09 Jan '17

This patch series adds the functionality to detect a misconfiguration between disk bus type and disk address type for disks that are using the address type virDomainDeviceDriveAddress. It also adds a test for it. A check for other bus types may be needed. This may require a driver specific function, as it is already implemented in virDomainDeviceDefPostParse(), for example. Changelog: - v1 -> v2: + Use full enumeration of the bus types + Add warning for unexpected bus type Marc Hartmayer (2): conf: Detect misconfiguration between disk bus and disk address tests: Add tests for disk configuration validation src/conf/domain_conf.c | 56 ++++++++++++++++++++++ .../qemuxml2argv-disk-fdc-incompatible-address.xml | 22 +++++++++ .../qemuxml2argv-disk-ide-incompatible-address.xml | 23 +++++++++ ...qemuxml2argv-disk-sata-incompatible-address.xml | 23 +++++++++ ...qemuxml2argv-disk-scsi-incompatible-address.xml | 24 ++++++++++ tests/qemuxml2argvtest.c | 8 ++++ 6 files changed, 156 insertions(+) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-fdc-incompatible-address.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-ide-incompatible-address.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-sata-incompatible-address.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-scsi-incompatible-address.xml -- 2.5.5

4 7

[libvirt] RFC for support Intel RDT/CAT in libvirt
by Qiao, Liyong 09 Jan '17

09 Jan '17

Hi folks I would like to start a discussion about how to support a new cpu feature in libvirt. CAT support is not fully merged into linux kernel yet, the target release is 4.10, and all patches has been merged into linux tip branch. So there won’t be interface/design changes. ## Background Intel RDT is a toolkit to do resource Qos for cpu such as llc(l3) cache, memory bandwidth usage, these fine granularity resource control features are very useful in a cloud environment which will run logs of noisy instances. Currently, Libvirt has supported CAT/MBMT/MBML already, they are only for resource usage monitor, propose to supporting CAT to control VM’s l3 cache quota. ## CAT interface in kernel In kernel, a new resource interface has been introduced under /sys/fs/resctrl, it’s used for resource control, for more information, refer Intel_rdt_ui [ https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/tree/Documentation… ] Kernel requires to provide schemata for l3 cache before add a task to a new partition, these interface is too much detail for a virtual machine user, so propose to let Libvirt manage schemata on the host. ## What will libvirt do? ### Questions: To enable CAT support in libvirt, we need to think about follow questions 1. Only set CAT when an VM has CPU pin, which is to say, l3 cache is per cpu socket resources. On a host which has 2 cpu sockets, each cpu socket has it own cache, and can not be shared.. 2. What the cache allocation policy should be used, this will be looks like: a. VM has it’s own dedicated l3 cache and also can share other l3 cache. b. VM can only use the caches which allocated to it. c. Has some pre-defined policies and priority for a VM Like COB [1] 1. Should reserve some l3 cache for host’s system usage (related to 2) 2. What’s the unit for l3 cache allocation? (related to 2) ### Propose Changes XML domain user interface changes: Option 1: explicit specify cache allocation for a VM 1 work with numa node Some cloud orchestration software use numa + vcpu pin together, so we can enable cat supporting with numa infra. Expose how many l3 cache a VM want to reserved and we require that the l3 cache should be bind on some specify cpu socket, just like what we did for numa node. This is an domain xml example which is generated by OpenStack Nova for allocate llc(l3 cache) when booting a new VM <domain> … <cputune> <vcpupin vcpu='0' cpuset='19'/> <vcpupin vcpu='1' cpuset='63'/> <vcpupin vcpu='2' cpuset='83'/> <vcpupin vcpu='3' cpuset='39'/> <vcpupin vcpu='4' cpuset='40'/> <vcpupin vcpu='5' cpuset='84'/> <emulatorpin cpuset='19,39-40,63,83-84'/> </cputune> ... <cpu mode='host-model'> <model fallback='allow'/> <topology sockets='3' cores='1' threads='2'/> <numa> <cell id='0' cpus='0-1' memory='2097152' l3cache='1408' unit='KiB'/> <cell id='1' cpus='2-5' memory='4194304' l3cache='5632' unit='KiB'/> </numa> </cpu> ... </domain> Refer to [http://libvirt.org/formatdomain.html#elementsCPUTuning] So finally we can calculate on which CPU socket(cell) we need to allocate how may l3cache for a VM. 2. work with vcpu pin Forget numa part, CAT setting should have relationship with cpu core setting, we can apply CAT policy if VM has set cpu pin setting (only VM won’t be schedule to another CPU sockets) Cache allocation on which CPU socket can be calculate as just as 1. We may need to enable both 1 and 2. There are several policy for cache allocation: Let’ take some examples: For intel e5 v4 2699(Signal socket), there are 55M l3 cache on the chip , the default of L3 schemata is L3:0=ffffff , it represents to use 20 bit to control l3 cache, each bit will represent 2.75M, which will be the minimal unit on this host. The allocation policy could be 3 policies : 1. One priority VM: A priority import VM can be allocated a dedicated amount l3 cache (let’s say 2.75 * 4 = 11M) and it can also reach the left 44 M cache which will be shared with other process and VM on the same host. So that we need to create a new ‘Partition’ n-20371 root@s2600wt:/sys/fs/resctrl# ls cpus info n-20371 schemata tasks Inside of n-20371 directory: root@s2600wt:/sys/fs/resctrl# ls n-20371/ cpus schemata tasks The schemata content will be L3:0=fffff The tasks content will be the pids of that VM Along we need to change the default schemata of system: root@s2600wt:/sys/fs/resctrl# cat schemata L3:0=ffff # which can not use the highest 4 bits, only tasks in n-20371 can reach that. In this design , we can only get 1 priority VM. Let’s change it a bit to have 2 VMs The schemata of the 2 VMs could be: 1. L3:0=ffff0 # could not use the 4 low bits 11M l3 cache 2. L3:0=0ffff # could not use the 4 high bits 11M l3 cache Default schemata changes to L3:0=0fff0 # default system process could only use the middle 33M l3 cache 2. Isolated l3 cache dedicated allocation for each VM(if required) A VM can only use the cache allocated to it. For example VM 1 requires to allocate 11 M It’s schemata will be L3:0=f0000 # VM 2 requires to allocate 11M It’s schemata will be L3:0=f000 And default schemata will be L3:0=fff In this case, we can create multiple VMs which each of them can have dedicated l3 cache. The disadvantage is that we the allocated cache could be not be shared efficiently. 3. Isolated l3 cache shared allocation for each VM(if required by user) In this case, we will put some VMs (which consider as noisy neighbors) to a ‘Partition’, restrict them to use the only caches allocated to them, by do this, other much more priority VM can be ensure to have enough l3 cache Then we should decide how much cache the noisy group should have, and put all of their pids in that tasks file. Option 2: set cache priority and apply policies Don’t specify cache amount at all, only define cache usage priority when define a VM domain XML. Cache priority will decide how much the VM can use l3 cache on a host, it’s not a quantized. So user don’t need to think about how much cache it should have when define a domain XML. Libvirt will decide cache allocation by the priority of VM defined and policies using. Disadvantage is that caches ability on different host may be different. Same VM domain XML on different host may have vary caches allocation amount. # Support CAT in libvirt itself or leverage other software COB COB is Intel Cache Orchestrator Balancer (COB). please refer http://clrgitlab.intel.com/rdt/cob/tree/master COB supports some pre-defined policies, it will monitor cpu/cache/cache missing and do cache allocation based on policy using. If COB support monitor some specified process (VM process) and accept priority defined, it will be good to reuse. At last the question came out: * Support a fine-grained llc cache control , let user specify cache allocation * Support pre-defined policies and user specify llc allocation priority. Reference [1] COB http://clrgitlab.intel.com/rdt/cob/tree/master [2] CAT intro: https://software.intel.com/en-us/articles/software-enabling-for-cache-alloc… [3] kernel Intel_rdt_ui [ https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/tree/Documentation… ] Best Regards Eli Qiao(乔立勇）OpenStack Core team OTC Intel. --

2 1

[libvirt] [PATCH 0/6] Don't run whole sec driver in namespace
by Michal Privoznik 09 Jan '17

09 Jan '17

In eadaa97548 I've tried to solve the issue of setting seclabels on private /dev/* entries. While my approach works, it has tiny flaw - anything that happens in the namespace stays in the namespace. I mean, if there's a internal state change occurring on relabel operation (it should not, and it doesn't nowadays, but it's no guarantee), this change is not reflected in the daemon. This is because when entering the namespace, the daemon forks, enters the namespace and then executes the RelabelAll() function. This imperfection is: a) very easy to forget b) very hard to debug Therefore, we may have transaction APIs as suggested here [1]. On transactionBegin() the sec driver will record [path. seclabel] somewhere instead of applying the label. Then on transactionCommit() new process is forked, enters the namespace and perform previously recorded changes. This way it is only the minimal code that runs in the namespace. Moreover, it runs over constant data thus there can be no internal state transition. 1: https://www.redhat.com/archives/libvir-list/2016-December/msg00254.html Michal Privoznik (6): security_selinux: s/virSecuritySELinuxSecurity/virSecuritySELinux/ security_dac: Resolve virSecurityDACSetOwnershipInternal const correctness security driver: Introduce transaction APIs security_dac: Implement transaction APIs security_selinux: Implement transaction APIs qemu: Use transactions from security driver src/libvirt_private.syms | 3 + src/qemu/qemu_driver.c | 28 +++-- src/qemu/qemu_security.c | 98 +++++---------- src/security/security_dac.c | 197 +++++++++++++++++++++++++++++- src/security/security_driver.h | 9 ++ src/security/security_manager.c | 38 ++++++ src/security/security_manager.h | 7 +- src/security/security_selinux.c | 219 +++++++++++++++++++++++++++++++--- src/security/security_stack.c | 49 ++++++++ src/storage/storage_backend.h | 2 +- src/storage/storage_backend_fs.c | 2 +- src/storage/storage_backend_gluster.c | 2 +- src/storage/storage_driver.c | 6 +- src/storage/storage_driver.h | 4 +- src/util/virstoragefile.c | 2 +- src/util/virstoragefile.h | 2 +- 16 files changed, 561 insertions(+), 107 deletions(-) -- 2.11.0

2 12

[libvirt] memory-ballooning side-channel attack
by bancfc＠openmailbox.org 08 Jan '17

08 Jan '17

Hello and Happy Holidays, In the past few years many serious attacks against the memory deduplication (KSM) feature of all hypervisors have been shown. [1] Even allowing attackers to modify/steal APT keys and source lists on the host. [2] Since its not enabled by default the fall out is relatively low and easily mitigated. New side-channel attacks against memory-ballon enabled VMs are beginning to surface. Please consider documenting this and disabling this feature for newly created VMs to have safe defaults. [1] https://staff.aist.go.jp/c.artho/papers/EuroSec2011-suzaki.pdf [2] https://www.usenix.org/system/files/conference/usenixsecurity16/sec16_paper… [3] http://ieeexplore.ieee.org/document/7562068/ *Hint: If you can't see the IEEE paper use sci-hub.

1 1