- Devel - libvirt List Archives

[libvirt] [PATCH] news: Add entry for soft reset support in Xen
by Jim Fehlig 08 Nov '18

08 Nov '18

Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- I'm torn on whether this is a new feature, improvement, or even simply a bug fix... docs/news.xml | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/docs/news.xml b/docs/news.xml index 9d98c34df2..e1d10cf7f1 100644 --- a/docs/news.xml +++ b/docs/news.xml @@ -35,6 +35,17 @@ <libvirt> <release version="v4.10.0" date="unreleased"> <section title="New features"> + <change> + <summary> + Xen: Add support for soft reset shutdown event + </summary> + <description> + The pvops Linux kernel uses soft reset to handle the crash + machine operation. The libxl driver now supports the soft + reset shutdown event, allowing proper crash handling of + pvops-based HVM domains. + </description> + </change> </section> <section title="Improvements"> </section> -- 2.18.0

2 2

[libvirt] [PATCH v2 0/3] Add "memfd" memory backing type
by marcandre.lureau＠redhat.com 08 Nov '18

08 Nov '18

From: Marc-André Lureau <marcandre.lureau(a)redhat.com> Hi, This is an alternative series from "[PATCH 0/5] Use memfd if possible". Instead of automatically using memfd for anonymous memory when available (as suggested by Daniel), it introduces the "memfd" memory backing type. Although using memfd transparently when possible is a good idea, it is a source of various complications for migration & save/restore. This could eventually be challenged in a different series. The first two patches have been modified and reviewed by John Ferlan. Hopefully they can be merged early, regardless of the last patch outcome, to avoid the painful rebase conflicts due to capabilities checks introduction. Thanks Marc-André Lureau (3): qemu: add memory-backend-memfd capability check qemu: check memory-backend-memfd.hugetlb capability qemu: add memfd source type docs/formatdomain.html.in | 9 +- docs/schemas/domaincommon.rng | 1 + src/conf/domain_conf.c | 3 +- src/conf/domain_conf.h | 1 + src/qemu/qemu_capabilities.c | 10 ++ src/qemu/qemu_capabilities.h | 2 + src/qemu/qemu_command.c | 69 +++++++---- src/qemu/qemu_domain.c | 12 +- .../caps_2.12.0.aarch64.replies | 94 ++++++++++++--- .../caps_2.12.0.aarch64.xml | 4 +- .../caps_2.12.0.ppc64.replies | 90 +++++++++++--- .../caps_2.12.0.ppc64.xml | 4 +- .../caps_2.12.0.s390x.replies | 98 ++++++++++++---- .../caps_2.12.0.s390x.xml | 4 +- .../caps_2.12.0.x86_64.replies | 110 +++++++++++++----- .../caps_2.12.0.x86_64.xml | 4 +- .../caps_3.0.0.ppc64.replies | 90 +++++++++++--- .../qemucapabilitiesdata/caps_3.0.0.ppc64.xml | 4 +- .../caps_3.0.0.riscv32.replies | 86 +++++++++++--- .../caps_3.0.0.riscv32.xml | 2 + .../caps_3.0.0.riscv64.replies | 86 +++++++++++--- .../caps_3.0.0.riscv64.xml | 2 + .../caps_3.0.0.x86_64.replies | 110 +++++++++++++----- .../caps_3.0.0.x86_64.xml | 4 +- .../memfd-memory-numa.x86_64-latest.args | 34 ++++++ tests/qemuxml2argvdata/memfd-memory-numa.xml | 36 ++++++ tests/qemuxml2argvtest.c | 2 + 27 files changed, 788 insertions(+), 183 deletions(-) create mode 100644 tests/qemuxml2argvdata/memfd-memory-numa.x86_64-latest.args create mode 100644 tests/qemuxml2argvdata/memfd-memory-numa.xml -- 2.19.0

8 17

[libvirt] [PATCH] spec: Drop support for Fedora 27
by Andrea Bolognani 08 Nov '18

08 Nov '18

In accordance with our platform support policy, now that Fedora 29 is out we no longer support building on Fedora 27. This allows us to remove a few version checks. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- libvirt.spec.in | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 90daefe6f1..71cd45c603 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -4,7 +4,7 @@ # that's still supported by the vendor. It may work on other distros # or versions, but no effort will be made to ensure that going forward. %define min_rhel 7 -%define min_fedora 27 +%define min_fedora 28 %if (0%{?fedora} && 0%{?fedora} >= %{min_fedora}) || (0%{?rhel} && 0%{?rhel} >= %{min_rhel}) %define supported_platform 1 @@ -72,7 +72,7 @@ %endif # We need a recent enough libiscsi (>= 1.18.0) -%if 0%{?fedora} >= 28 || 0%{?rhel} > 7 +%if 0%{?fedora} || 0%{?rhel} > 7 %define with_storage_iscsi_direct 0%{!?_without_storage_iscsi_direct:1} %else %define with_storage_iscsi_direct 0 @@ -258,7 +258,7 @@ BuildRequires: /usr/bin/pod2man %endif BuildRequires: gcc BuildRequires: git -%if 0%{?fedora} >= 27 || 0%{?rhel} > 7 +%if 0%{?fedora} || 0%{?rhel} > 7 BuildRequires: perl-interpreter %else BuildRequires: perl @@ -386,7 +386,7 @@ BuildRequires: wireshark-devel >= 2.1.0 BuildRequires: libssh-devel >= 0.7.0 %endif -%if 0%{?fedora} > 27 || 0%{?rhel} > 7 +%if 0%{?fedora} || 0%{?rhel} > 7 BuildRequires: rpcgen BuildRequires: libtirpc-devel %endif -- 2.19.1

3 2

[libvirt] [PATCH v2] lxc: Clang is complaining about possible NULL pointer.
by Julio Faracco 08 Nov '18

08 Nov '18

The array "mount" inside lxc_container is not being checked before for loop. Clang syntax scan is complaining about this segmentation fault. Signed-off-by: Julio Faracco <jcfaracco(a)gmail.com> --- src/lxc/lxc_container.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index 918194dacd..d834bf01d7 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -867,9 +867,13 @@ static int lxcContainerSetReadOnly(void) } } - if (mounts) - qsort(mounts, nmounts, sizeof(mounts[0]), - virStringSortRevCompare); + if (!mounts) { + ret = 0; + goto cleanup; + } + + qsort(mounts, nmounts, sizeof(mounts[0]), + virStringSortRevCompare); for (i = 0; i < nmounts; i++) { VIR_DEBUG("Bind readonly %s", mounts[i]); @@ -883,9 +887,7 @@ static int lxcContainerSetReadOnly(void) ret = 0; cleanup: - for (i = 0; i < nmounts; i++) - VIR_FREE(mounts[i]); - VIR_FREE(mounts); + virStringListFreeCount(mounts, nmounts); endmntent(procmnt); return ret; -- 2.17.1

3 2

[libvirt] [PATCH 0/3] add disk driver metadata_cache_size option
by Nikolay Shirokovskiy 08 Nov '18

08 Nov '18

Hi, all. This is a patch series after offlist agreement on introducing metadata-cache-size option for disks. The options itself is described in 2nd patch of the series. There is a plenty of attempts to add option to set qcow2 metadata cache sizes in past, see [1] [2] to name recent that has comments I tried to address or has work I reused to some extent. I would like to address Peter's comment in [1] that this option is image's option rather then disk's here. While this makes sense if we set specific cache value that covers disk only partially, here when we have maximum policy that covers whole disk it makes sense to set same value for all images. The setted value is only upper limit on cache size and thus cache for every image will grow proportionally to image data size "visible from top" eventually I guess. And these sizes are changing during guest lifetime - thus we need set whole disk limit for every image for 'maxium' effect. On the other hand if we add policies different from 'maximum' it may require per image option. Well I can't imagine name for element for backing chain that will have cache size attribute better then 'driver'). And driver is already used for top image so I leave it this way. KNOWN ISSUES 1. when using -driver to configure disks in command line cache size does not get propagated thru backing chain 2. when making external disk snapshot cache size sneak into backing file filename and thus cache size for backing image became remembered there 3. blockcommit after such snapshot will not work because libvirt is not ready for backing file name is reported as json sometimes Looks like 1 can be addressed in qemu. The reason for behaviour in 2 I do not understand honestly. And 3 will be naturally addessed after blockjobs starts working with blockdev configuration I guess. LINKS [1] [PATCH] qemu: Added support L2 table cache for qcow2 disk https://www.redhat.com/archives/libvir-list/2018-July/msg00166.html [2] [PATCH v6 0/2] Add support for qcow2 cache https://www.redhat.com/archives/libvir-list/2017-September/msg00553.html Nikolay Shirokovskiy (3): qemu: caps: add drive.qcow2.l2-cache-size xml: add disk driver metadata_cache_size option qemu: support metadata-cache-size for blockdev docs/formatdomain.html.in | 7 ++++ docs/schemas/domaincommon.rng | 10 +++++ src/conf/domain_conf.c | 17 ++++++++ src/conf/domain_conf.h | 9 ++++ src/qemu/qemu_block.c | 5 ++- src/qemu/qemu_capabilities.c | 3 ++ src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 26 ++++++++++++ src/qemu/qemu_domain.c | 2 + src/util/virstoragefile.c | 1 + src/util/virstoragefile.h | 1 + tests/qemucapabilitiesdata/caps_2.10.0.aarch64.xml | 1 + tests/qemucapabilitiesdata/caps_2.10.0.ppc64.xml | 1 + tests/qemucapabilitiesdata/caps_2.10.0.s390x.xml | 1 + tests/qemucapabilitiesdata/caps_2.10.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_2.11.0.s390x.xml | 1 + tests/qemucapabilitiesdata/caps_2.11.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_2.12.0.aarch64.xml | 1 + tests/qemucapabilitiesdata/caps_2.12.0.ppc64.xml | 1 + tests/qemucapabilitiesdata/caps_2.12.0.s390x.xml | 1 + tests/qemucapabilitiesdata/caps_2.12.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_2.5.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_2.6.0.aarch64.xml | 1 + tests/qemucapabilitiesdata/caps_2.6.0.ppc64.xml | 1 + tests/qemucapabilitiesdata/caps_2.6.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_2.7.0.s390x.xml | 1 + tests/qemucapabilitiesdata/caps_2.7.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_2.8.0.s390x.xml | 1 + tests/qemucapabilitiesdata/caps_2.8.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_2.9.0.ppc64.xml | 1 + tests/qemucapabilitiesdata/caps_2.9.0.s390x.xml | 1 + tests/qemucapabilitiesdata/caps_2.9.0.x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_3.0.0.ppc64.xml | 1 + tests/qemucapabilitiesdata/caps_3.0.0.riscv32.xml | 1 + tests/qemucapabilitiesdata/caps_3.0.0.riscv64.xml | 1 + tests/qemucapabilitiesdata/caps_3.0.0.s390x.xml | 1 + tests/qemucapabilitiesdata/caps_3.0.0.x86_64.xml | 1 + .../qemuxml2argvdata/disk-metadata_cache_size.args | 34 +++++++++++++++ .../qemuxml2argvdata/disk-metadata_cache_size.xml | 42 +++++++++++++++++++ tests/qemuxml2argvtest.c | 2 + .../disk-metadata_cache_size.xml | 48 ++++++++++++++++++++++ tests/qemuxml2xmltest.c | 2 + 42 files changed, 235 insertions(+), 1 deletion(-) create mode 100644 tests/qemuxml2argvdata/disk-metadata_cache_size.args create mode 100644 tests/qemuxml2argvdata/disk-metadata_cache_size.xml create mode 100644 tests/qemuxml2xmloutdata/disk-metadata_cache_size.xml -- 1.8.3.1

3 12

[libvirt] [PATCH] virNetworkDefUpdateDNSHost: Require both IP and a hostname to match
by W. Trevor King 08 Nov '18

08 Nov '18

Since fc19a0059 (network: backend functions for updating network dns host/srv/txt, 2012-11-12), the matching logic for various network components has been: 1) for HOST records, it's considered a match if the IP address or any of the hostnames of an existing record matches. 2) for SRV records, it's a match if all of domain+service+protocol+target *which have been specified* are matched. 3) for TXT records, there is only a single field to match - name (value can be the same for multiple records, and isn't considered a search term), so by definition there can be no ambiguous matches. But HOST records can have the same hostname for multiple records (similar to TXT records with the same value). The value that needs to be distinct for HOST records is the IP address. This commit updates the matching logic to only consider the IP address. Compared to the previous HOST logic: 1. You can now delete entries from an existing network like: <dns> <host ip="192.168.1.1"> <hostname>example</hostname> </host> <host ip="192.168.1.2"> <hostname>example</hostname> </host> </dns> with input like: <host ip="192.168.1.1"> </host> or: <host ip="192.168.1.1"> <hostname>example</hostname> </host> Previously, only the former would work (the latter used to raise "multiple matching DNS HOST records were found in network"). 2. You can no longer remove entries by hostname alone. Previously, you may have been able to remove an entry from an existing network like: <dns> <host ip="192.168.1.1"> <hostname>example-1</hostname> </host> <host ip="192.168.1.2"> <hostname>example-2</hostname> </host> </dns> with input like: <host name="example"> <hostname>example-1</hostname> </host> using the 'name' property to get through the partialOkay check in virNetworkDHCPHostDefParseXML. Now that input will raise "Missing IP address in network '%s' DNS HOST record". 3. You can now add multiple entries with a common hostname (as long as they have distinct IP addresses). Previously, adding: <host ip="192.168.1.1"> <hostname>example</hostname> </host> to an existing: <host ip="192.168.1.2"> <hostname>example</hostname> </host> would have raised "there is already at least one DNS HOST record with a matching field in network". --- I'm actually not clear on whether the 'ip' attribute is required to be unique or not. If not, maybe the logic should be: * Deletes with just an IP remove all <host> entries that match that IP. * Deletes with just a hostname remove all <hostname> entries that match that hostname. * Deletes with an IP and a hostname remove matching <hostname> entries from <host> entries which match the IP. * If <hostname> removal completely empties a <host>, the <host> is also removed. Thoughts? src/conf/network_conf.c | 31 ++++++++++--------------------- 1 file changed, 10 insertions(+), 21 deletions(-) diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index 39a13b4..8ed62ac 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -587,14 +587,14 @@ virNetworkDNSHostDefParseXML(const char *networkName, xmlNodePtr cur; char *ip; - if (!(ip = virXMLPropString(node, "ip")) && !partialOkay) { + if (!(ip = virXMLPropString(node, "ip"))) { virReportError(VIR_ERR_XML_DETAIL, _("Missing IP address in network '%s' DNS HOST record"), networkName); goto error; } - if (ip && (virSocketAddrParse(&def->ip, ip, AF_UNSPEC) < 0)) { + if (virSocketAddrParse(&def->ip, ip, AF_UNSPEC) < 0) { virReportError(VIR_ERR_XML_DETAIL, _("Invalid IP address in network '%s' DNS HOST record"), networkName); @@ -603,6 +603,13 @@ virNetworkDNSHostDefParseXML(const char *networkName, } VIR_FREE(ip); + if (!VIR_SOCKET_ADDR_VALID(&def->ip)) { + virReportError(VIR_ERR_XML_DETAIL, + _("Invalid IP address in network '%s' DNS HOST record"), + networkName); + goto error; + } + cur = node->children; while (cur != NULL) { if (cur->type == XML_ELEMENT_NODE && @@ -631,13 +638,6 @@ virNetworkDNSHostDefParseXML(const char *networkName, goto error; } - if (!VIR_SOCKET_ADDR_VALID(&def->ip) && def->nnames == 0) { - virReportError(VIR_ERR_XML_DETAIL, - _("Missing ip and hostname in network '%s' DNS HOST record"), - networkName); - goto error; - } - return 0; error: @@ -3334,18 +3334,7 @@ virNetworkDefUpdateDNSHost(virNetworkDefPtr def, goto cleanup; for (i = 0; i < dns->nhosts; i++) { - bool foundThisTime = false; - - if (virSocketAddrEqual(&host.ip, &dns->hosts[i].ip)) - foundThisTime = true; - - for (j = 0; j < host.nnames && !foundThisTime; j++) { - for (k = 0; k < dns->hosts[i].nnames && !foundThisTime; k++) { - if (STREQ(host.names[j], dns->hosts[i].names[k])) - foundThisTime = true; - } - } - if (foundThisTime) { + if virSocketAddrEqual(&host.ip, &dns->hosts[i].ip) { foundCt++; foundIdx = i; } -- 1.8.3.1

2 4

[libvirt] Parsing of memory.stat in libvirt
by Peter.Chubb＠data61.csiro.au 08 Nov '18

08 Nov '18

Hi Folks, libvirt currently spams the logs with error : virCgroupGetMemoryStat:2490 : internal error: Cannot parse 'memory.stat' cgroup file whenever someone does ps in a container. This is because the parser for memory/stat is incorrect: the `line' variable is never updated, so each time through the loop, the same start-of-line is compared for the token; and as all spaces are eventually replaced with NUL the error exit is taken instead of ending the loop properly. Here is a strawman patch to fix the problem. Please note, I'm not subscribed to this list; I reported the bug as https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=913023 and was asked to send the patch upstream. src/util/vircgroup.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) --- libvirt.orig/src/util/vircgroup.c +++ libvirt/src/util/vircgroup.c @@ -2477,7 +2477,7 @@ virCgroupGetMemoryStat(virCgroupPtr grou line = stat; - while (line) { + while (*line) { char *newLine = strchr(line, '\n'); char *valueStr = strchr(line, ' '); unsigned long long value; @@ -2507,6 +2507,11 @@ virCgroupGetMemoryStat(virCgroupPtr grou inactiveFileVal = value >> 10; else if (STREQ(line, "unevictable")) unevictableVal = value >> 10; + + if (newLine) + line = newLine + 1; + else + break; } *cache = cacheVal; -- Dr Peter Chubb Tel: +61 2 9490 5852 http://ts.data61.csiro.au/ Trustworthy Systems Group Data61 (formerly NICTA)

2 2

[libvirt] [PATCH] tests: Use correct function name in error path
by John Ferlan 08 Nov '18

08 Nov '18

Commit id 5eb61e6846 neglected to change the name in the wrong value output to virCgroupGetPercpuStats from virCgroupGetMemoryUsage. Signed-off-by: John Ferlan <jferlan(a)redhat.com> --- Pushed under the trivial rule. tests/vircgrouptest.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/vircgrouptest.c b/tests/vircgrouptest.c index 8fcee21bb2..310e1fb6a2 100644 --- a/tests/vircgrouptest.c +++ b/tests/vircgrouptest.c @@ -756,7 +756,7 @@ static int testCgroupGetPercpuStats(const void *args ATTRIBUTE_UNUSED) if (params[i].value.ul != expected[i]) { fprintf(stderr, - "Wrong value from virCgroupGetMemoryUsage at %zu (expected %llu)\n", + "Wrong value from virCgroupGetPercpuStats at %zu (expected %llu)\n", i, params[i].value.ul); goto cleanup; } -- 2.17.2

1 0

[libvirt] [PATCH 0/3] qemu: don't duplicate suspended events and state changes
by Nikolay Shirokovskiy 07 Nov '18

07 Nov '18

This patch series is similar in implementation to [1] but the motivation is simply to avoid suspended event duplication. [1] https://www.redhat.com/archives/libvir-list/2018-September/msg01304.html Nikolay Shirokovskiy (3): qemu: pass stop reason from qemuProcessStopCPUs to stop handler qemu: map suspended state reason to suspended event detail qemu: don't duplicate suspended events and state changes src/qemu/qemu_domain.c | 34 ++++++++++++++++++++++++++++++++++ src/qemu/qemu_domain.h | 7 +++++++ src/qemu/qemu_driver.c | 26 +++----------------------- src/qemu/qemu_migration.c | 42 ++++++------------------------------------ src/qemu/qemu_migration.h | 4 ---- src/qemu/qemu_process.c | 37 +++++++++++++++++++++++++------------ 6 files changed, 75 insertions(+), 75 deletions(-) -- 1.8.3.1

2 12

[libvirt] [PATCH 0/2] conf: Forbid some nonsensical combinations under <memoryBackend>
by Michal Privoznik 07 Nov '18

07 Nov '18

*** BLURB HERE *** Michal Prívozník (2): conf: Forbid hugepages and <source type='file'/> conf: Forbid hugepages and <allocation mode="ondemand"/> src/conf/domain_conf.c | 25 +++++++++++++++++-- .../qemuxml2argvdata/hugepages-memaccess2.xml | 1 - 2 files changed, 23 insertions(+), 3 deletions(-) -- 2.18.1

3 5