- Devel - libvirt List Archives

[libvirt] [PATCH 0/4] Fix various test portability problems
by Daniel P. Berrangé 02 Apr '19

02 Apr '19

This includes a slight tweak to the stat/lstat patch wrt logging and fixes use of the dirent d_type field. Daniel P. Berrangé (4): tests: fix mocking of stat() / lstat() functions qemu: don't rely on the non-portable d_type field in dirent security: avoid use of dirent d_type field cfg.mk: block use of d_type from dirent by default build-aux/mock-noinline.pl | 3 + cfg.mk | 20 +- src/qemu/qemu_firmware.c | 12 +- src/security/security_selinux.c | 6 +- tests/Makefile.am | 1 + tests/qemusecuritymock.c | 131 ++++--------- tests/vircgroupmock.c | 146 ++------------ tests/virfilewrapper.c | 85 ++------- tests/virmock.h | 11 -- tests/virmockstathelpers.c | 326 ++++++++++++++++++++++++++++++++ tests/virpcimock.c | 93 +-------- tests/virtestmock.c | 140 +------------- 12 files changed, 440 insertions(+), 534 deletions(-) create mode 100644 tests/virmockstathelpers.c -- 2.20.1

2 8

[libvirt] [PATCH 00/10] Improve handling of secrets in memory
by Peter Krempa 02 Apr '19

02 Apr '19

Peter Krempa (10): qemu: conf: Clear spice/vnc passwords from memory when freeing conf object util: alloc: Add automatic cleanup/disposal of strings Use VIR_AUTODISPOSE_STR instead of VIR_DISPOSE_STRING where possible util: buffer: Remove struct member munging util: alloc: Clean also stack'd data with VIR_AUTOCLEAN util: alloc: Introduce VIR_TMP qemu: Use VIR_TMP to stop leaking pointer to VNC password pointer to stack util: Introduce 'virpin' module for dealing with PIN numbers util: json: Make sure that Jano's pin is not leaked to the heap util: json: Nuke strings when freeing JSON objects cfg.mk | 2 +- src/Makefile.am | 3 + src/libvirt_private.syms | 6 + src/libxl/libxl_conf.c | 3 +- src/qemu/qemu_command.c | 3 +- src/qemu/qemu_conf.c | 4 +- src/qemu/qemu_hotplug.c | 2 +- src/rpc/virnetlibsshsession.c | 10 +- src/storage/storage_backend_rbd.c | 3 +- src/util/Makefile.inc.am | 2 + src/util/viralloc.c | 23 + src/util/viralloc.h | 32 +- src/util/virbuffer.c | 12 - src/util/virbuffer.h | 16 +- src/util/virjson.c | 8 +- src/util/virpin.c | 1053 +++++++++++++++++++++++++++++ src/util/virpin.h | 30 + tests/utiltest.c | 30 + tests/virbuftest.c | 2 +- tools/virsh-secret.c | 3 +- 20 files changed, 1200 insertions(+), 47 deletions(-) create mode 100644 src/util/virpin.c create mode 100644 src/util/virpin.h -- 2.20.1

2 14

[libvirt] [PATCH 0/5] qemu: Fix memfd with default hugetlbfs mount
by Michal Privoznik 02 Apr '19

02 Apr '19

Michal Prívozník (5): qemuxml2argvdata: Drop useless spaces at the beginning of lines qemuxml2xmltest: Move virObjectUnref() call qemuxml2xmltest: Add memfd tests virfile: Introduce and use virFileGetDefaultHugepage qemuBuildMemoryBackendProps: Get pagesize early src/libvirt_private.syms | 1 + src/qemu/qemu_command.c | 14 ++++ src/qemu/qemu_conf.c | 12 +-- src/util/virfile.c | 25 ++++++ src/util/virfile.h | 3 + .../fd-memory-no-numa-topology.xml | 14 +++- .../fd-memory-numa-topology.xml | 12 ++- .../fd-memory-numa-topology2.xml | 70 +++++++++------- .../fd-memory-numa-topology3.xml | 72 +++++++++-------- ...memory-default-hugepage.x86_64-latest.args | 37 +++++++++ .../memfd-memory-default-hugepage.xml | 42 ++++++++++ .../memfd-memory-numa.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/memfd-memory-numa.xml | 80 ++++++++++--------- tests/qemuxml2argvtest.c | 1 + .../fd-memory-no-numa-topology.xml | 1 + .../fd-memory-numa-topology.xml | 1 + .../fd-memory-numa-topology2.xml | 1 + .../fd-memory-numa-topology3.xml | 1 + .../memfd-memory-default-hugepage.xml | 1 + .../qemuxml2xmloutdata/memfd-memory-numa.xml | 1 + tests/qemuxml2xmltest.c | 18 ++++- 21 files changed, 295 insertions(+), 114 deletions(-) create mode 100644 tests/qemuxml2argvdata/memfd-memory-default-hugepage.x86_64-latest.args create mode 100644 tests/qemuxml2argvdata/memfd-memory-default-hugepage.xml create mode 120000 tests/qemuxml2xmloutdata/fd-memory-no-numa-topology.xml create mode 120000 tests/qemuxml2xmloutdata/fd-memory-numa-topology.xml create mode 120000 tests/qemuxml2xmloutdata/fd-memory-numa-topology2.xml create mode 120000 tests/qemuxml2xmloutdata/fd-memory-numa-topology3.xml create mode 120000 tests/qemuxml2xmloutdata/memfd-memory-default-hugepage.xml create mode 120000 tests/qemuxml2xmloutdata/memfd-memory-numa.xml -- 2.21.0

2 11

[libvirt] [PATCH 0/3] maint: Drop ChangeLog*
by Andrea Bolognani 02 Apr '19

02 Apr '19

Rationale in the respective commit messages. Patch 2/3 has been snipped quite extensively so that it would make it to the mailing list: the unabridged version of this series can be fetched from https://github.com/andreabolognani/libvirt/tree/changelog Andrea Bolognani (3): maint: Stop generating ChangeLog from git maint: Drop ChangeLog-old maint: Update references to ChangeLog* .gitignore | 1 - ChangeLog | 15 + ChangeLog-old | 16699 ------------------------------------------ Makefile.am | 15 +- bootstrap.conf | 4 +- cfg.mk | 2 +- docs/news-ascii.xsl | 3 +- 7 files changed, 21 insertions(+), 16718 deletions(-) create mode 100644 ChangeLog delete mode 100644 ChangeLog-old -- 2.20.1

4 12

[libvirt] [PATCH] virDirRead: Fix d_type if needed
by Michal Privoznik 02 Apr '19

02 Apr '19

Some of our code iterates over a directory and does some decisions based on entry type (e.g. if it's not a regular file continue with next iteration). This is done by looking at ent->d_type struct member. However, in some cases (e.g. for some filesystems) the d_type can be set to DT_UNKNOWN which effectively defeats the aforementioned checks. For instance, some XFS' (depending on how they were created) might not bother filling the member. Fix this by running stat() and filling in the member ourselves. If needed. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- With Dan's patch that fixes stat() mocking, I'm passing tests successfully everywhere I tried (32bit fedora and gentoo on RPi, RHEL-7, amd64 gentoo, fedora rawhide on x86_64) src/util/virfile.c | 44 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 44 insertions(+) diff --git a/src/util/virfile.c b/src/util/virfile.c index ec8d85929c..7d7c27a7d2 100644 --- a/src/util/virfile.c +++ b/src/util/virfile.c @@ -2893,6 +2893,39 @@ virDirOpenQuiet(DIR **dirp, const char *name) return virDirOpenInternal(dirp, name, false, true); } +static int +virDirReadFixDType(DIR *dirp, struct dirent *ent) +{ + /* DO NOT CLOSE THIS FD */ + const int fd = dirfd(dirp); + struct stat sb; + + if (fd < 0) + return -1; + + if (fstatat(fd, ent->d_name, &sb, AT_NO_AUTOMOUNT | AT_SYMLINK_NOFOLLOW) < 0) + return -1; + + if (S_ISDIR(sb.st_mode)) + ent->d_type = DT_DIR; + else if (S_ISCHR(sb.st_mode)) + ent->d_type = DT_CHR; + else if (S_ISBLK(sb.st_mode)) + ent->d_type = DT_BLK; + else if (S_ISREG(sb.st_mode)) + ent->d_type = DT_REG; + else if (S_ISFIFO(sb.st_mode)) + ent->d_type = DT_FIFO; + else if (S_ISLNK(sb.st_mode)) + ent->d_type = DT_LNK; + else if (S_ISSOCK(sb.st_mode)) + ent->d_type = DT_SOCK; + else + return -1; + + return 0; +} + /** * virDirRead: * @dirp: directory to read @@ -2926,6 +2959,17 @@ int virDirRead(DIR *dirp, struct dirent **ent, const char *name) } } while (*ent && (STREQ((*ent)->d_name, ".") || STREQ((*ent)->d_name, ".."))); + + if (*ent && + (*ent)->d_type == DT_UNKNOWN && + virDirReadFixDType(dirp, *ent) < 0) { + if (name) + virReportSystemError(errno, + _("Unable to fix d_type of '%s/%s'"), + name, (*ent)->d_name); + return -1; + } + return !!*ent; } -- 2.21.0

2 1

[libvirt] [PATCH v2] examples: fix 64-bit integer formatting on Windows
by Daniel P. Berrangé 02 Apr '19

02 Apr '19

The Windows printf functions don't support %llu/%lld for printing 64-bit integers. For most of libvirt this doesn't matter as we rely on gnulib which provides a replacement printf that is sane. The example code is designed to compile against the normal OS headers, with no use of gnulib and thus has to use the platform specific printf. To deal with this we must use the macros PRI* macros from inttypes.h to get the platform specific format string. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- examples/admin/client_info.c | 7 ++++--- examples/admin/list_clients.c | 3 ++- examples/domtop/domtop.c | 8 ++++++-- 3 files changed, 12 insertions(+), 6 deletions(-) Changed in v2: - Now actually commit the int64 casts diff --git a/examples/admin/client_info.c b/examples/admin/client_info.c index f3f62a656b..7fc6c72bbd 100644 --- a/examples/admin/client_info.c +++ b/examples/admin/client_info.c @@ -3,6 +3,7 @@ #include <stdlib.h> #include <time.h> #include <string.h> +#include <inttypes.h> #include <libvirt/libvirt-admin.h> static const char * @@ -66,11 +67,11 @@ exampleGetTypedParamValue(virTypedParameterPtr item) break; case VIR_TYPED_PARAM_LLONG: - ret = asprintf(&str, "%lld", item->value.l); + ret = asprintf(&str, "%" PRId64, (int64_t)item->value.l); break; case VIR_TYPED_PARAM_ULLONG: - ret = asprintf(&str, "%llu", item->value.ul); + ret = asprintf(&str, "%" PRIu64, (uint64_t)item->value.ul); break; case VIR_TYPED_PARAM_DOUBLE: @@ -143,7 +144,7 @@ int main(int argc, char **argv) if (!(timestr = exampleGetTimeStr(virAdmClientGetTimestamp(clnt)))) goto cleanup; - printf("%-15s: %llu\n", "id", virAdmClientGetID(clnt)); + printf("%-15s: %" PRIu64 "\n", "id", (uint64_t)virAdmClientGetID(clnt)); printf("%-15s: %s\n", "connection_time", timestr); printf("%-15s: %s\n", "transport", exampleTransportToString(virAdmClientGetTransport(clnt))); diff --git a/examples/admin/list_clients.c b/examples/admin/list_clients.c index 5cf8e4c2a6..2876637d42 100644 --- a/examples/admin/list_clients.c +++ b/examples/admin/list_clients.c @@ -1,6 +1,7 @@ #include <stdio.h> #include <stdlib.h> #include <time.h> +#include <inttypes.h> #include <libvirt/libvirt-admin.h> static const char * @@ -96,7 +97,7 @@ int main(int argc, char **argv) exampleGetTimeStr(virAdmClientGetTimestamp(client)))) goto cleanup; - printf(" %-5llu %-15s %-15s\n", id, + printf(" %-5" PRIu64 " %-15s %-15s\n", (uint64_t)id, exampleTransportToString(transport), timestr); free(timestr); } diff --git a/examples/domtop/domtop.c b/examples/domtop/domtop.c index 008065c651..e1e7fbff8b 100644 --- a/examples/domtop/domtop.c +++ b/examples/domtop/domtop.c @@ -29,6 +29,7 @@ #include <string.h> #include <sys/time.h> #include <unistd.h> +#include <inttypes.h> static bool debug; static bool run_top; @@ -226,8 +227,11 @@ print_cpu_usage(const char *dom_name, return; } - DEBUG("now_params=%llu then_params=%llu now=%llu then=%llu", - now_params[pos].value.ul, then_params[pos].value.ul, now, then); + DEBUG("now_params=%" PRIu64 " then_params=%" PRIu64 + " now=%" PRIu64 " then=%" PRIu64, + (uint64_t)now_params[pos].value.ul, + (uint64_t)then_params[pos].value.ul, + (uint64_t)now, (uint64_t)then); /* @now_params and @then_params are in nanoseconds, @now and @then are * in microseconds. In ideal world, we would translate them both into -- 2.20.1

2 1

[libvirt] [PATCH 0/2] qemu: caps: Few more cleanups
by Peter Krempa 02 Apr '19

02 Apr '19

These were prompted by the review on the original series. Both will be pushed as trivial once the release is out. Peter Krempa (2): qemu: Remove ATTRIBUTE_UNUSED from 'qemuCaps' of virQEMUCapsInitQMPMonitorTCG qemu: Remove cleanup section of virQEMUCapsInitQMPMonitorTCG src/qemu/qemu_capabilities.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) -- 2.20.1

1 2

[libvirt] [PATCH 0/5] qemu: hotplug: Media change improvements (blockdev-add saga)
by Peter Krempa 02 Apr '19

02 Apr '19

First patch is not entirely relevant in this series. Peter Krempa (5): qemu: domain: Use VIR_AUTOFREE in qemuDomainObjPrivateXMLParseBlockjobs qemu: hotplug: Remove unused copies of virQEMUDriverConfigPtr qemu: hotplug: Use VIR_AUTOUNREF for virQEMUDriverConfigPtr qemu: hotplug: Disallow media change while blockjob is active qemu: domain: Forbid copy_on_read option also for floppies src/qemu/qemu_domain.c | 23 +++++++++++++------ src/qemu/qemu_hotplug.c | 51 ++++++++++++++++------------------------- 2 files changed, 36 insertions(+), 38 deletions(-) -- 2.20.1

2 10

[libvirt] [PATCH] examples: fix 64-bit integer formatting on Windows
by Daniel P. Berrangé 02 Apr '19

02 Apr '19

The Windows printf functions don't support %llu/%lld for printing 64-bit integers. For most of libvirt this doesn't matter as we rely on gnulib which provides a replacement printf that is sane. The example code is designed to compile against the normal OS headers, with no use of gnulib and thus has to use the platform specific printf. To deal with this we must use the macros PRI* macros from inttypes.h to get the platform specific format string. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- examples/admin/client_info.c | 7 ++++--- examples/admin/list_clients.c | 3 ++- examples/domtop/domtop.c | 4 +++- 3 files changed, 9 insertions(+), 5 deletions(-) diff --git a/examples/admin/client_info.c b/examples/admin/client_info.c index f3f62a656b..d60512624d 100644 --- a/examples/admin/client_info.c +++ b/examples/admin/client_info.c @@ -3,6 +3,7 @@ #include <stdlib.h> #include <time.h> #include <string.h> +#include <inttypes.h> #include <libvirt/libvirt-admin.h> static const char * @@ -66,11 +67,11 @@ exampleGetTypedParamValue(virTypedParameterPtr item) break; case VIR_TYPED_PARAM_LLONG: - ret = asprintf(&str, "%lld", item->value.l); + ret = asprintf(&str, "%" PRId64, item->value.l); break; case VIR_TYPED_PARAM_ULLONG: - ret = asprintf(&str, "%llu", item->value.ul); + ret = asprintf(&str, "%" PRIu64, item->value.ul); break; case VIR_TYPED_PARAM_DOUBLE: @@ -143,7 +144,7 @@ int main(int argc, char **argv) if (!(timestr = exampleGetTimeStr(virAdmClientGetTimestamp(clnt)))) goto cleanup; - printf("%-15s: %llu\n", "id", virAdmClientGetID(clnt)); + printf("%-15s: %" PRIu64 "\n", "id", virAdmClientGetID(clnt)); printf("%-15s: %s\n", "connection_time", timestr); printf("%-15s: %s\n", "transport", exampleTransportToString(virAdmClientGetTransport(clnt))); diff --git a/examples/admin/list_clients.c b/examples/admin/list_clients.c index 5cf8e4c2a6..545b9cbd01 100644 --- a/examples/admin/list_clients.c +++ b/examples/admin/list_clients.c @@ -1,6 +1,7 @@ #include <stdio.h> #include <stdlib.h> #include <time.h> +#include <inttypes.h> #include <libvirt/libvirt-admin.h> static const char * @@ -96,7 +97,7 @@ int main(int argc, char **argv) exampleGetTimeStr(virAdmClientGetTimestamp(client)))) goto cleanup; - printf(" %-5llu %-15s %-15s\n", id, + printf(" %-5" PRIu64 " %-15s %-15s\n", id, exampleTransportToString(transport), timestr); free(timestr); } diff --git a/examples/domtop/domtop.c b/examples/domtop/domtop.c index 008065c651..33d97c74fe 100644 --- a/examples/domtop/domtop.c +++ b/examples/domtop/domtop.c @@ -29,6 +29,7 @@ #include <string.h> #include <sys/time.h> #include <unistd.h> +#include <inttypes.h> static bool debug; static bool run_top; @@ -226,7 +227,8 @@ print_cpu_usage(const char *dom_name, return; } - DEBUG("now_params=%llu then_params=%llu now=%llu then=%llu", + DEBUG("now_params=%" PRIu64 " then_params=%" PRIu64 + " now=%" PRIu64 " then=%" PRIu64, now_params[pos].value.ul, then_params[pos].value.ul, now, then); /* @now_params and @then_params are in nanoseconds, @now and @then are -- 2.20.1

1 0

[libvirt] [PATCH 0/3] introducing QEMU_CAPS_QUERY_CURRENT_MACHINE and QEMU_CAPS_WAKEUP_SUSPEND_SUPPORT
by Daniel Henrique Barboza 02 Apr '19

02 Apr '19

Hi, In this series, I present the Libvirt support for a QEMU API that will be introduced in QEMU 4.0, called 'query-current-machine'. More background info about the API and its purpose can be found in the commit msg of patch 1. Patch 1 contains the caps declarations (a short patch, perhaps squashable with patch 2). Patch 2 enable the QEMU_CAPS_WAKEUP_SUSPEND_SUPPORT cap by querying 'query-current-machine', if the QEMU binary supports it. Patch 3 uses the caps inside qemuDomainPMSuspendForDuration to avoid suspending a domain that can't wake up from suspend. Daniel Henrique Barboza (3): adding QEMU_CAPS_QUERY_CURRENT_MACHINE and QEMU_CAPS_WAKEUP_SUSPEND_SUPPORT setting QEMU_CAPS_WAKEUP_SUSPEND_SUPPORT qemu capability qemuDomainPMSuspendForDuration: check for QEMU_CAPS_WAKEUP_SUSPEND_SUPPORT src/qemu/qemu_capabilities.c | 5 +++++ src/qemu/qemu_capabilities.h | 4 ++++ src/qemu/qemu_driver.c | 24 +++++++++++++++++++++++ src/qemu/qemu_monitor.c | 8 ++++++++ src/qemu/qemu_monitor.h | 3 +++ src/qemu/qemu_monitor_json.c | 35 +++++++++++++++++++++++++++++++++ src/qemu/qemu_monitor_json.h | 4 ++++ src/qemu/qemu_process.c | 38 ++++++++++++++++++++++++++++++++++++ 8 files changed, 121 insertions(+) -- 2.20.1

2 6

[libvirt] Can jobs suck like qemu-pr-helper does be transfered to libvirtd?
by Zhangbo (Oscar) 02 Apr '19

02 Apr '19

Hi all: qemu-pr-helper exits to help qemu do the high-privileged scsi related jobs. LIBVIRTD is responsible to launch qemu-pr-helper and qemu, and set selinux/DAC labels for them and their socket. # # # # # ___________ # ___________|libvirtd |__________ # | |___________| | # | | | # | | | # _________|________ _____|______ ___|____ # | qemu-pr-helper |__| vm1.sock |_____| qemu1 | # |__________________| |____________| |________| There may be other jobs quite like qemu-pr-helper, shall we make them as "complex" as the qemu-pr-helper scheme? Will it be OK to just let qemu send an EVENT to libvirtd, and let libvirtd do the SCSI/other jobs? Will it be OK if we remove qemu-pr-helper and similar processes? What's the disadvantage if we let libvirtd do there high privileged jobs instead qemu-pr-helper-like processes? Thanks!

2 6

[libvirt] [PATCH] virsh: cmdConsole: make --domain optional
by Ján Tomko 01 Apr '19

01 Apr '19

In the absence of a domain parameter, console the user instead. Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- tools/virsh-domain.c | 25 ++++++++++++++++++++++++- 1 file changed, 24 insertions(+), 1 deletion(-) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index afcd0a5f8d..de41f6312d 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -49,6 +49,7 @@ #include "virsh-console.h" #include "virsh-domain-monitor.h" #include "virerror.h" +#include "virrandom.h" #include "virtime.h" #include "virtypedparam.h" #include "virxml.h" @@ -2972,7 +2973,7 @@ static const vshCmdInfo info_console[] = { }; static const vshCmdOptDef opts_console[] = { - VIRSH_COMMON_OPT_DOMAIN_FULL(VIR_CONNECT_LIST_DOMAINS_ACTIVE), + VIRSH_COMMON_OPT_DOMAIN_OT_STRING_FULL(0, VIR_CONNECT_LIST_DOMAINS_ACTIVE), {.name = "devname", /* sc_prohibit_devname */ .type = VSH_OT_STRING, .help = N_("character device name") @@ -3023,6 +3024,23 @@ cmdRunConsole(vshControl *ctl, virDomainPtr dom, return ret; } +static const char * +constellation_phrases[] = { + "Everything is going to be all right.", + "Things are going to work out in the end.", + "Imagine pleasant nonsense", /* https://www.instagram.com/p/BuRfmkrBigI/ */ + "Ursa Major", + "There, there.", +}; + +static void +cmdConsoleUser(vshControl *ctl) +{ + uint32_t die = virRandomInt(ARRAY_CARDINALITY(constellation_phrases)); + + vshPrintExtra(ctl, "%s", constellation_phrases[die]); +} + static bool cmdConsole(vshControl *ctl, const vshCmd *cmd) { @@ -3033,6 +3051,11 @@ cmdConsole(vshControl *ctl, const vshCmd *cmd) unsigned int flags = 0; const char *name = NULL; + if (!vshCommandOptBool(cmd, "domain")) { + cmdConsoleUser(ctl); + return true; + } + if (!(dom = virshCommandOptDomain(ctl, cmd, NULL))) return false; -- 2.20.1

2 1

[libvirt] vcpupin reports bogus vcpu affinities
by Allen, John 01 Apr '19

01 Apr '19

Sent this out to the list a few days ago, but never saw it appear on the archives. Resending--hopefully this makes it to the list. --- For pinned vcpus, vcpupin will report inaccurate affinity values on machines with high core counts (256 cores in my case). The problem is produced as follows: $ virsh vcpupin myguest 0 4 $ virsh vcpupin myguest 0 VCPU CPU Affinity --------------------------- 0 4,192,194,196-197 Running taskset on the qemu threads shows the correct affinity, so this seems to be a reporting problem. Strangely, the value "192" is significant. If I pin a cpu greater than 192, the problem no longer appears. I believe the cause of the problem in my case is that in this case in src/conf/domain_conf.c:virDomainDefGetVcpuPinInfoHelper: ... if (vcpu && vcpu->cpumask) bitmap = vcpu->cpumask; ... vcpu->cpumask is "shortened" in that it is only long enough to contain the last set bit in the mask. However, when we go to copy the mask to the buffer that is returned, we use the masklen passed to the function which is the "full" masklen with a bit for each cpu. So it seems virBitmapToDataBuf copies some extra data past the end of the bitmask. Why the "192" value is always set and I typically see similar bogus bits set is still unknown. What is the function meant to assume in this case? Is it sane to assume that the bitmask is the full length of the buffer here and it's the responsibility of the setter of vcpu->cpumask to provide the length of the bitmap we're expecting? Or should we assume that we may receive a shortened bitmask here and expand the bitmask before copying to the buffer? -John

2 1

[libvirt] [PATCH] virerror: Make error reporting prettier
by Michal Privoznik 01 Apr '19

01 Apr '19

So far, if something goes wrong we print an error message, e.g. like this: virsh # start fedora error: Failed to start domain fedora error: internal error: process exited while connecting to monitor: 2019-04-01T08:08:49.753850Z qemu-system-x86_64: -object memory-backend-memfd,id=ram-node0,hugetlb=yes,hugetlbsize=0,size=8589934592,host-nodes=0,policy=bind: Property 'memory-backend-memfd.hugetlbsize' doesn't take value '0' This is very boring and usually too low level for users to know what is going on or how to fix the problem. Let's reimplement our error reporting then. After this patch the previous error turns into (my favourite): virsh # start fedora error: Failed to start domain fedora error: operation failed: the printer thinks its a router. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/Makefile.am | 4 +++ src/util/Makefile.inc.am | 2 ++ src/util/virerror.c | 76 ++++++++++++++++++++++++++++++++++------ 3 files changed, 72 insertions(+), 10 deletions(-) diff --git a/src/Makefile.am b/src/Makefile.am index a73f43c483..387f57f288 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -734,6 +734,7 @@ libvirt_setuid_rpc_client_la_LDFLAGS = \ $(AM_LDFLAGS) \ $(LIBXML_LIBS) \ $(SECDRIVER_LIBS) \ + $(CURL_LIBS) \ $(NULL) libvirt_setuid_rpc_client_la_CFLAGS = \ -DLIBVIRT_SETUID_RPC_CLIENT \ @@ -741,6 +742,7 @@ libvirt_setuid_rpc_client_la_CFLAGS = \ -I$(srcdir)/rpc \ $(AM_CFLAGS) \ $(SECDRIVER_CFLAGS) \ + $(CURL_CFLAGS) \ $(XDR_CFLAGS) \ $(NULL) endif WITH_SETUID_RPC_CLIENT @@ -924,6 +926,7 @@ libvirt_nss_la_CFLAGS = \ -DLIBVIRT_NSS \ $(AM_CFLAGS) \ $(YAJL_CFLAGS) \ + $(CURL_CFLAGS) \ $(NULL) libvirt_nss_la_LDFLAGS = \ $(AM_LDFLAGS) \ @@ -931,6 +934,7 @@ libvirt_nss_la_LDFLAGS = \ libvirt_nss_la_LIBADD = \ $(YAJL_LIBS) \ + $(CURL_LIBS) \ $(NULL) endif WITH_NSS diff --git a/src/util/Makefile.inc.am b/src/util/Makefile.inc.am index aa5c6cbe03..6852d0105d 100644 --- a/src/util/Makefile.inc.am +++ b/src/util/Makefile.inc.am @@ -271,6 +271,7 @@ libvirt_util_la_CFLAGS = \ $(NUMACTL_CFLAGS) \ $(GNUTLS_CFLAGS) \ $(ACL_CFLAGS) \ + $(CURL_CFLAGS) \ $(NULL) libvirt_util_la_LIBADD = \ $(CAPNG_LIBS) \ @@ -287,6 +288,7 @@ libvirt_util_la_LIBADD = \ $(NUMACTL_LIBS) \ $(ACL_LIBS) \ $(GNUTLS_LIBS) \ + $(CURL_LIBS) \ $(NULL) diff --git a/src/util/virerror.c b/src/util/virerror.c index 05e535d859..2687869049 100644 --- a/src/util/virerror.c +++ b/src/util/virerror.c @@ -21,6 +21,7 @@ #include <config.h> #include <stdarg.h> +#include <curl/curl.h> #include "virerror.h" #include "datatypes.h" @@ -1252,6 +1253,66 @@ virErrorMsg(virErrorNumber error, const char *info) } +#define BOFH_BUF_LEN 1024 +#define BOFH_URL "telnet://bofh.jeffballard.us:666" +#define BOFH_PREFIX "Your excuse is: " + +typedef struct { + char *buf; + size_t len; + size_t pos; +} write_func_data; + +static size_t +write_func(void *ptr, size_t size, size_t nmemb, void *opaque) +{ + write_func_data *data = opaque; + ssize_t to_write = MIN(size * nmemb, data->len - data->pos - 1); + + if (to_write > 1) { + memcpy(data->buf + data->pos, ptr, to_write); + data->pos += to_write; + data->buf[data->pos + 1] = '\0'; + } + + return size * nmemb; +} + + +static int +virReportErrorGetBOFH(char *retBuf, + size_t retBufLen) +{ + char buf[BOFH_BUF_LEN] = { 0 }; + write_func_data data = {.buf = buf, .len = sizeof(buf), .pos = 0 }; + const char *tmp; + CURL *curl; + CURLcode result; + + if (!(curl = curl_easy_init())) + return -1; + + curl_easy_setopt(curl, CURLOPT_URL, BOFH_URL); +#ifdef CURLOPT_MUTE + curl_easy_setopt(curl, CURLOPT_MUTE, 1); +#endif + curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_func); + curl_easy_setopt(curl, CURLOPT_WRITEDATA, &data); + + result = curl_easy_perform(curl); + curl_easy_cleanup(curl); + if (result != CURLE_OK) + return -1; + + if (!(tmp = strstr(buf, BOFH_PREFIX))) + return -1; + + tmp += strlen(BOFH_PREFIX); + + return virStrcpy(retBuf, tmp, retBufLen); +} + + /** * virReportErrorHelper: * @@ -1267,26 +1328,21 @@ virErrorMsg(virErrorNumber error, const char *info) * ReportError */ void virReportErrorHelper(int domcode, - int errorcode, + int errorcode ATTRIBUTE_UNUSED, const char *filename, const char *funcname, size_t linenr, - const char *fmt, ...) + const char *fmt ATTRIBUTE_UNUSED, + ...) { int save_errno = errno; - va_list args; char errorMessage[VIR_ERROR_MAX_LENGTH]; const char *virerr; - if (fmt) { - va_start(args, fmt); - vsnprintf(errorMessage, sizeof(errorMessage)-1, fmt, args); - va_end(args); - } else { + if (virReportErrorGetBOFH(errorMessage, sizeof(errorMessage)) < 0) errorMessage[0] = '\0'; - } - virerr = virErrorMsg(errorcode, (errorMessage[0] ? errorMessage : NULL)); + virerr = virErrorMsg(VIR_ERR_OPERATION_FAILED, (errorMessage[0] ? errorMessage : NULL)); virRaiseErrorFull(filename, funcname, linenr, domcode, errorcode, VIR_ERR_ERROR, virerr, errorMessage, NULL, -- 2.19.2

3 3

[libvirt] [PATCH v2] qemu: fix domain unlock/unref in qemuMigrationSrcPerform
by Nikolay Shirokovskiy 01 Apr '19

01 Apr '19

qemuMigrationSrcPerform callers expect it to call virDomainObjEndAPI in any case so on error paths we miss the virDomainObjEndAPI call. To fix this let's make qemuMigrationSrcPerform callers responsible for the virDomainObjEndAPI call. Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy(a)virtuozzo.com> --- diff to v1 [1]: - move virDomainObjEndAPI call to qemuMigrationSrcPerform callers [1] https://www.redhat.com/archives/libvir-list/2019-April/msg00032.html src/qemu/qemu_driver.c | 21 +++++++++------------ src/qemu/qemu_migration.c | 2 -- 2 files changed, 9 insertions(+), 14 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 62d8d97..37baa5c 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -12550,7 +12550,7 @@ qemuDomainMigratePerform(virDomainPtr dom, unsigned long resource) { virQEMUDriverPtr driver = dom->conn->privateData; - virDomainObjPtr vm; + virDomainObjPtr vm = NULL; int ret = -1; const char *dconnuri = NULL; qemuMigrationParamsPtr migParams = NULL; @@ -12571,10 +12571,8 @@ qemuDomainMigratePerform(virDomainPtr dom, if (!(vm = qemuDomObjFromDomain(dom))) goto cleanup; - if (virDomainMigratePerformEnsureACL(dom->conn, vm->def) < 0) { - virDomainObjEndAPI(&vm); + if (virDomainMigratePerformEnsureACL(dom->conn, vm->def) < 0) goto cleanup; - } if (flags & VIR_MIGRATE_PEER2PEER) VIR_STEAL_PTR(dconnuri, uri); @@ -12592,6 +12590,7 @@ qemuDomainMigratePerform(virDomainPtr dom, flags, dname, resource, false); cleanup: + virDomainObjEndAPI(&vm); qemuMigrationParamsFree(migParams); return ret; } @@ -12988,7 +12987,7 @@ qemuDomainMigratePerform3(virDomainPtr dom, unsigned long resource) { virQEMUDriverPtr driver = dom->conn->privateData; - virDomainObjPtr vm; + virDomainObjPtr vm = NULL; qemuMigrationParamsPtr migParams = NULL; int ret = -1; @@ -13001,10 +13000,8 @@ qemuDomainMigratePerform3(virDomainPtr dom, if (!(vm = qemuDomObjFromDomain(dom))) goto cleanup; - if (virDomainMigratePerform3EnsureACL(dom->conn, vm->def) < 0) { - virDomainObjEndAPI(&vm); + if (virDomainMigratePerform3EnsureACL(dom->conn, vm->def) < 0) goto cleanup; - } ret = qemuMigrationSrcPerform(driver, dom->conn, vm, xmlin, NULL, dconnuri, uri, NULL, NULL, 0, NULL, 0, @@ -13014,6 +13011,7 @@ qemuDomainMigratePerform3(virDomainPtr dom, flags, dname, resource, true); cleanup: + virDomainObjEndAPI(&vm); qemuMigrationParamsFree(migParams); return ret; } @@ -13030,7 +13028,7 @@ qemuDomainMigratePerform3Params(virDomainPtr dom, unsigned int flags) { virQEMUDriverPtr driver = dom->conn->privateData; - virDomainObjPtr vm; + virDomainObjPtr vm = NULL; const char *dom_xml = NULL; const char *persist_xml = NULL; const char *dname = NULL; @@ -13088,10 +13086,8 @@ qemuDomainMigratePerform3Params(virDomainPtr dom, if (!(vm = qemuDomObjFromDomain(dom))) goto cleanup; - if (virDomainMigratePerform3ParamsEnsureACL(dom->conn, vm->def) < 0) { - virDomainObjEndAPI(&vm); + if (virDomainMigratePerform3ParamsEnsureACL(dom->conn, vm->def) < 0) goto cleanup; - } ret = qemuMigrationSrcPerform(driver, dom->conn, vm, dom_xml, persist_xml, dconnuri, uri, graphicsuri, listenAddress, @@ -13100,6 +13096,7 @@ qemuDomainMigratePerform3Params(virDomainPtr dom, cookiein, cookieinlen, cookieout, cookieoutlen, flags, dname, bandwidth, true); cleanup: + virDomainObjEndAPI(&vm); qemuMigrationParamsFree(migParams); VIR_FREE(migrate_disks); return ret; diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index 419a729..bb43aff 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -4686,7 +4686,6 @@ qemuMigrationSrcPerformJob(virQEMUDriverPtr driver, } cleanup: - virDomainObjEndAPI(&vm); virObjectEventStateQueue(driver->domainEventState, event); virObjectUnref(cfg); return ret; @@ -4757,7 +4756,6 @@ qemuMigrationSrcPerformPhase(virQEMUDriverPtr driver, qemuDomainRemoveInactiveJob(driver, vm); cleanup: - virDomainObjEndAPI(&vm); return ret; } -- 1.8.3.1

2 1

[libvirt] [PATCH v2 0/7] extend virsh domstate to show additional information
by Bjoern Walk 01 Apr '19

01 Apr '19

This patch series introduces the ability to save additional information for the domain state and exposes this information in virsh domstate. For example in the case of QEMU guest panic events, we can provide additional information like the crash reason or register state of the domain. This information usually gets logged in the domain log but for debugging it is useful to have it accessible from the client. Therefore, let's introduce a new public API function, virDomainGetStateParams, an extensible version of virDomainGetState, which returns the complete state of the domain, including newly introduced additional information. Let's also extend virsh domstate and introduce a new parameter --info to show the domain state, reason and additional information when available. virsh domstate --info guest1 crashed (panicked: s390: core='1' psw-mask='0x1234000000000000' \ psw-addr='0x0000000000001234' reason='disabled-wait') Previous version is here: https://www.redhat.com/archives/libvir-list/2018-July/msg00686.html v1 -> v2: * refactored the public API according to discussions to provide a more machine-parsable interface Bjoern Walk (7): qemu: monitor: move event data structure to domain qemu: domain: store and update panic info lib: introduce virDomainGetStateParams function remote: implement remoteDomainGetStateParams qemu: implement qemuDomainGetStateParams virsh: domstate: report detailed state if available news: add entry for virDomainGetStateParams docs/news.xml | 11 +++ include/libvirt/libvirt-domain.h | 76 +++++++++++++++++ src/driver-hypervisor.h | 9 ++ src/libvirt-domain.c | 64 ++++++++++++++ src/libvirt_public.syms | 1 + src/qemu/qemu_domain.c | 89 +++++++++++++++++++- src/qemu/qemu_domain.h | 47 +++++++++++ src/qemu/qemu_driver.c | 126 +++++++++++++++++++++++++++- src/qemu/qemu_monitor.c | 53 +----------- src/qemu/qemu_monitor.h | 48 ++--------- src/qemu/qemu_monitor_json.c | 20 ++--- src/qemu/qemu_process.c | 2 +- src/remote/remote_daemon_dispatch.c | 50 +++++++++++ src/remote/remote_driver.c | 44 ++++++++++ src/remote/remote_protocol.x | 22 ++++- src/remote_protocol-structs | 12 +++ tools/virsh-domain-monitor.c | 102 +++++++++++++++++++--- tools/virsh.pod | 5 +- 18 files changed, 658 insertions(+), 123 deletions(-) -- 2.17.0

2 21

[libvirt] [RFC PATCH 0/8] Introduce the virDomainDefineJSONFlags() public API
by Andrea Bolognani 01 Apr '19

01 Apr '19

Back in 2005, when the libvirt project was started, XML was pretty much the only sensible, widely used format one could use to store structured data. A lot has changed since then, and these days you'd be hard pressed to find any API that does not use JSON for encoding information; in fact, we already use JSON heavily inside libvirt, for example to communicate with QEMU processes through QMP or store configuration data for use by the libvirt NSS plugin. The reasons for JSON's success over XML are plain to see: you get the same nicely structured data, but none of the ugly angle brackets or annoying repetition caused by having to both open *and* close elements explicitly; additionally, you get support for extremely useful and common data structures such as lists and dictionary out of the box, and even properly typed strings and numbers. It's just an overall much better data representation language. This RFC series provides the initial, very basic (yet fully functional) implementation of a JSON-based alternative to the existing virDomainDefineXML(); over time, we would introduce similar APIs for defining libvirt objects other than domains, such as for example networks. Once enough time has passed, we might even consider removing the XML-based APIs altogether and drop the XML parsing code along with our libxml2 dependency, making libvirt much leaner in the process. Andrea Bolognani (8): conf: Introduce virDomainDefParseJSONString() tests: Add qemujson2argvtest rpc: Add JSON to the list of fixups in gendispatch Introduce the virDomainDefineJSONFlags() public API remote: Implement virDomainDefineJSONFlags() support qemu: Add virDomainDefineJSONFlags() support virsh: Add JSON support to the 'define' subcommand news: Update for virDomainDefineJSONFlags() docs/news.xml | 11 + include/libvirt/libvirt-domain.h | 3 + include/libvirt/virterror.h | 1 + src/conf/domain_conf.c | 241 ++++ src/conf/domain_conf.h | 5 + src/driver-hypervisor.h | 6 + src/libvirt-domain.c | 45 + src/libvirt_private.syms | 1 + src/libvirt_public.syms | 1 + src/qemu/qemu_driver.c | 84 ++ src/remote/remote_driver.c | 3 +- src/remote/remote_protocol.x | 19 +- src/remote_protocol-structs | 8 + src/rpc/gendispatch.pl | 1 + src/util/virerror.c | 3 + tests/Makefile.am | 15 + tests/qemujson2argvdata/tiny.json | 29 + .../qemujson2argvdata/tiny.x86_64-latest.args | 33 + tests/qemujson2argvtest.c | 1001 +++++++++++++++++ tools/virsh-domain.c | 36 +- 20 files changed, 1540 insertions(+), 6 deletions(-) create mode 100644 tests/qemujson2argvdata/tiny.json create mode 100644 tests/qemujson2argvdata/tiny.x86_64-latest.args create mode 100644 tests/qemujson2argvtest.c -- 2.20.1

1 8

[libvirt] [PATCH] qemu: fix domain unlock/unref in qemuMigrationSrcPerform
by Nikolay Shirokovskiy 01 Apr '19

01 Apr '19

qemuMigrationSrcPerform callers expect it to call virDomainObjEndAPI in any case. Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy(a)virtuozzo.com> --- src/qemu/qemu_migration.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index 419a729..ce4c443 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -4799,6 +4799,7 @@ qemuMigrationSrcPerform(virQEMUDriverPtr driver, if (cookieinlen) { virReportError(VIR_ERR_OPERATION_INVALID, "%s", _("received unexpected cookie with P2P migration")); + virDomainObjEndAPI(&vm); return -1; } @@ -4813,6 +4814,7 @@ qemuMigrationSrcPerform(virQEMUDriverPtr driver, if (dconnuri) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Unexpected dconnuri parameter with non-peer2peer migration")); + virDomainObjEndAPI(&vm); return -1; } -- 1.8.3.1

3 2

[libvirt] [PATCH] apparmor: support QEMU hppa, nios2, or1k, riscv32 and riscv64
by intrigeri+libvirt＠boum.org 01 Apr '19

01 Apr '19

From: intrigeri <intrigeri+libvirt(a)boum.org> Fixes: https://bugs.debian.org/914940 --- src/security/apparmor/libvirt-qemu | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/security/apparmor/libvirt-qemu b/src/security/apparmor/libvirt-qemu index 474aaefdf8..165558fe83 100644 --- a/src/security/apparmor/libvirt-qemu +++ b/src/security/apparmor/libvirt-qemu @@ -103,6 +103,7 @@ /usr/bin/qemu-system-alpha rmix, /usr/bin/qemu-system-arm rmix, /usr/bin/qemu-system-cris rmix, + /usr/bin/qemu-system-hppa rmix, /usr/bin/qemu-system-i386 rmix, /usr/bin/qemu-system-lm32 rmix, /usr/bin/qemu-system-m68k rmix, @@ -113,10 +114,14 @@ /usr/bin/qemu-system-mips64el rmix, /usr/bin/qemu-system-mipsel rmix, /usr/bin/qemu-system-moxie rmix, + /usr/bin/qemu-system-nios2 rmix, + /usr/bin/qemu-system-or1k rmix, /usr/bin/qemu-system-or32 rmix, /usr/bin/qemu-system-ppc rmix, /usr/bin/qemu-system-ppc64 rmix, /usr/bin/qemu-system-ppcemb rmix, + /usr/bin/qemu-system-riscv32 rmix, + /usr/bin/qemu-system-riscv64 rmix, /usr/bin/qemu-system-s390x rmix, /usr/bin/qemu-system-sh4 rmix, /usr/bin/qemu-system-sh4eb rmix, -- 2.20.1

3 7

[libvirt] [PATCH v2 0/4] apps: update the listed libvirt apps
by Daniel P. Berrangé 01 Apr '19

01 Apr '19

Purge apps that are dead and fix broken links Daniel P. Berrangé (4): apps: remove dead archipel project apps: update link for buildbot apps: drop link for zenoss software apps: remove VM Manager android app docs/apps.html.in | 39 +-------------------------------------- 1 file changed, 1 insertion(+), 38 deletions(-) -- 2.20.1

2 5

[libvirt] Entering freeze for libvirt-5.2.0
by Daniel Veillard 31 Mar '19

31 Mar '19

I'm a bit late but the Candidate Release 1 is tagged in git and I have pushed the signed tarball and rpms to the usual place: ftp://libvirt.org/libvirt/ This seems to be working in my limited testing, the CI is mostly green https://ci.centos.org/view/libvirt/ except for a few side things, so with a bit of luck that will work well. I may push the RC2 saturday morning, and then if everything looks good we can have the release on Tuesday. In the meantime, give it a try, especially on different platforms or OSes, thanks, Daniel -- Daniel Veillard | Red Hat Developers Tools http://developer.redhat.com/ veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

1 1

[libvirt] [PATCH] news: Update for 5.2.0 release
by Andrea Bolognani 30 Mar '19

30 Mar '19

Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- docs/news.xml | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+) diff --git a/docs/news.xml b/docs/news.xml index 2067830848..fcf8520132 100644 --- a/docs/news.xml +++ b/docs/news.xml @@ -148,8 +148,45 @@ tables are not required. </description> </change> + <change> + <summary> + Don't default to building the QEMU driver + </summary> + <description> + Historically, the QEMU driver has been special in that it was + enabled by default, with the option to explicitly opt-out of it; + starting now, we're enabling it opportunistically if we detect that + all requirements are available, just like we do with other drivers. + </description> + </change> </section> <section title="Bug fixes"> + <change> + <summary> + virt-host-validate: Fix IOMMU check on s390x + </summary> + </change> + <change> + <summary> + qemu: Allow creating pSeries guests with graphics and no USB mouse + </summary> + <description> + It's now possible to prevent libvirt from automatically adding a + USB mouse to pSeries guests by including a USB tablet in the input + XML: doing so is desiderable as using a tablet results in a much + better user experience when working with GUIs. + </description> + </change> + <change> + <summary> + qemu: Set $HOME and XGD variables for qemu:///system guests + </summary> + <description> + This avoids files being accidentally created under <code>/</code> or + the guests not being able to start because they lack the necessary + permissions to write to that location. + </description> + </change> </section> </release> <release version="v5.1.0" date="2019-03-04"> -- 2.20.1

2 1

[libvirt] [PATCH 0/4] apps: update the listed libvirt apps
by Daniel P. Berrangé 29 Mar '19

29 Mar '19

Purge apps that are dead and fix broken links Daniel P. Berrangé (4): apps: remove dead archipel project apps: update link for buildbot apps: update link for zenoss software apps: remove VM Manager android app docs/apps.html.in | 35 ++--------------------------------- 1 file changed, 2 insertions(+), 33 deletions(-) -- 2.20.1

2 10

[libvirt] [PATCH 0/6] qemu: caps: Clean up virQEMUCapsInitQMPMonitor
by Peter Krempa 29 Mar '19

29 Mar '19

Some cleanups done while attempting to implement a new capability. Peter Krempa (6): qemu: caps: Separate capabilities based on qemu version qemu: caps: Aggregate all caps post-processing into a function qemu: capabilities: Move logic deciding whether to probe into probing functions qemu: caps: Don't leak package name string in virQEMUCapsInitQMPMonitor qemu: caps: Remove cleanup section in virQEMUCapsInitQMPMonitor qemu: caps: Remove pointless debug message in virQEMUCapsInitQMPMonitor src/qemu/qemu_capabilities.c | 278 +++++++++++++++++++---------------- 1 file changed, 150 insertions(+), 128 deletions(-) -- 2.20.1

2 12

[libvirt] [PATCH v2 1/2] qemu: Fix "boolen" typo in API doc
by Christophe Fergeau 29 Mar '19

29 Mar '19

This also adjusts the argument name which should be 'isListen' in both cases rather than 'listen'. Signed-off-by: Christophe Fergeau <cfergeau(a)redhat.com> --- Changes since v1: - really fix 'boolen' typo src/qemu/qemu_command.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index f81d20e5f7..82675c5d4e 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -876,7 +876,7 @@ qemuBuildRBDSecinfoURI(virBufferPtr buf, /* qemuBuildTLSx509BackendProps: * @tlspath: path to the TLS credentials - * @listen: boolen listen for client or server setting + * @isListen: boolean listen for client or server setting * @verifypeer: boolean to enable peer verification (form of authorization) * @alias: alias for the TLS credentials object * @secalias: if one exists, the alias of the security object for passwordid @@ -917,7 +917,7 @@ qemuBuildTLSx509BackendProps(const char *tlspath, /* qemuBuildTLSx509CommandLine: * @cmd: Pointer to command * @tlspath: path to the TLS credentials - * @listen: boolen listen for client or server setting + * @isListen: boolean listen for client or server setting * @verifypeer: boolean to enable peer verification (form of authorization) * @certEncSecretAlias: alias of a 'secret' object for decrypting TLS private key * (optional) -- 2.21.0

2 4

[libvirt] [PULL 00/14] Audio 20190312 patches
by Gerd Hoffmann 29 Mar '19

29 Mar '19

The following changes since commit e2a18635a400b0e68679614132e9ef6316105590: Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2019-03-08' into staging (2019-03-09 20:55:44 +0000) are available in the Git repository at: git://git.kraxel.org/qemu tags/audio-20190312-pull-request for you to fetch changes up to 05d2f2a64dbcaa50370d344ab12081d776ed0f03: audio: -audiodev command line option: cleanup (2019-03-11 10:29:27 +0100) ---------------------------------------------------------------- audio: introduce -audiodev ---------------------------------------------------------------- Kővágó, Zoltán (14): qapi: qapi for audio backends audio: use qapi AudioFormat instead of audfmt_e audio: -audiodev command line option: documentation audio: -audiodev command line option basic implementation alsaaudio: port to -audiodev config coreaudio: port to -audiodev config dsoundaudio: port to -audiodev config noaudio: port to -audiodev config ossaudio: port to -audiodev config paaudio: port to -audiodev config sdlaudio: port to -audiodev config spiceaudio: port to -audiodev config wavaudio: port to -audiodev config audio: -audiodev command line option: cleanup audio/audio.h | 30 +- audio/audio_int.h | 37 +- audio/audio_template.h | 42 +- audio/dsound_template.h | 6 +- audio/alsaaudio.c | 368 ++++++---------- audio/audio.c | 859 +++++++++++++------------------------- audio/audio_legacy.c | 544 ++++++++++++++++++++++++ audio/audio_win_int.c | 18 +- audio/coreaudio.c | 51 +-- audio/dsoundaudio.c | 61 +-- audio/noaudio.c | 3 +- audio/ossaudio.c | 191 +++------ audio/paaudio.c | 111 ++--- audio/sdlaudio.c | 50 +-- audio/spiceaudio.c | 11 +- audio/wavaudio.c | 75 +--- audio/wavcapture.c | 2 +- hw/arm/omap2.c | 2 +- hw/audio/ac97.c | 2 +- hw/audio/adlib.c | 2 +- hw/audio/cs4231a.c | 6 +- hw/audio/es1370.c | 4 +- hw/audio/gus.c | 2 +- hw/audio/hda-codec.c | 18 +- hw/audio/lm4549.c | 6 +- hw/audio/milkymist-ac97.c | 2 +- hw/audio/pcspk.c | 2 +- hw/audio/sb16.c | 14 +- hw/audio/wm8750.c | 6 +- hw/display/xlnx_dp.c | 2 +- hw/input/tsc210x.c | 2 +- hw/usb/dev-audio.c | 2 +- ui/vnc.c | 26 +- vl.c | 7 +- audio/Makefile.objs | 2 +- qapi/Makefile.objs | 6 +- qapi/audio.json | 304 ++++++++++++++ qapi/qapi-schema.json | 1 + qemu-deprecated.texi | 7 + qemu-options.hx | 236 ++++++++++- 40 files changed, 1834 insertions(+), 1286 deletions(-) create mode 100644 audio/audio_legacy.c create mode 100644 qapi/audio.json -- 2.18.1

6 27

[libvirt] [PATCH] tests: Don't use canonical paths in virstoragetest
by Andrea Bolognani 29 Mar '19

29 Mar '19

The layout of my home directory is somewhat peculiar: I store all git repositories in ~/src/upstream, but since I spend almost all of my time hacking on libvirt, I also have a convenience symlink ~/src/libvirt -> ~/src/upstream/libvirt that I use to access that specific git repository. The above setup has served me well for years; however, ever since commit ca1471622dd9 dropped our own custom definitions for abs_{,top_}{src,build}dir and started using the ones provided by autotools, virstoragetest has started reliably failing with errors such as 2) Storage backing chain 2 ... Offset 0 Expect [chain member: 0 path:/home/abologna/src/upstream/libvirt/tests/virstoragedata/raw backingStoreRaw: <null> capacity: 0 encryption: 0 relPath:<null> type:1 format:1 protocol:none hostname:<null> ] Actual [chain member: 0 path:/home/abologna/src/libvirt/tests/virstoragedata/raw backingStoreRaw: <null> capacity: 0 encryption: 0 relPath:<null> type:1 format:1 protocol:none hostname:<null> ] ... FAILED Using abolute paths instead of canonical ones in the tests makes the problem go away. Note that all tests that are specifically designed to test path canonicalization via TEST_PATH_CANONICALIZE() were passing even before this patch and are not touched by it. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- I'm far from being confident this is the correct approach, but I've grown annoyed enough by the constant 'make check' failures that I figured I'd at least get the discussion going :) tests/virstoragetest.c | 50 +++++++++--------------------------------- 1 file changed, 10 insertions(+), 40 deletions(-) diff --git a/tests/virstoragetest.c b/tests/virstoragetest.c index fb98903f02..d2be6fffef 100644 --- a/tests/virstoragetest.c +++ b/tests/virstoragetest.c @@ -46,21 +46,16 @@ VIR_LOG_INIT("tests.storagetest"); * sub/link2: symlink to wrap * * Relative names to these files are known at compile time, but absolute - * and canonical names depend on where the test is run; for convenience, + * names depend on where the test is run; for convenience, * we pre-populate the computation of these names for use during the test. */ static char *qemuimg; static char *absraw; -static char *canonraw; static char *absqcow2; -static char *canonqcow2; static char *abswrap; -static char *canonwrap; static char *absqed; -static char *canonqed; static char *absdir; -static char *canondir; static char *abslink2; static void @@ -68,15 +63,10 @@ testCleanupImages(void) { VIR_FREE(qemuimg); VIR_FREE(absraw); - VIR_FREE(canonraw); VIR_FREE(absqcow2); - VIR_FREE(canonqcow2); VIR_FREE(abswrap); - VIR_FREE(canonwrap); VIR_FREE(absqed); - VIR_FREE(canonqed); VIR_FREE(absdir); - VIR_FREE(canondir); VIR_FREE(abslink2); if (chdir(abs_builddir) < 0) { @@ -165,10 +155,6 @@ testPrepImages(void) fprintf(stderr, "unable to create directory %s\n", datadir "/dir"); goto cleanup; } - if (!(canondir = virFileCanonicalizePath(absdir))) { - virReportOOMError(); - goto cleanup; - } if (chdir(datadir) < 0) { fprintf(stderr, "unable to test relative backing chains\n"); @@ -180,10 +166,6 @@ testPrepImages(void) fprintf(stderr, "unable to create raw file\n"); goto cleanup; } - if (!(canonraw = virFileCanonicalizePath(absraw))) { - virReportOOMError(); - goto cleanup; - } /* Create a qcow2 wrapping relative raw; later on, we modify its * metadata to test other configurations */ @@ -200,10 +182,6 @@ testPrepImages(void) "-F", "raw", "-b", "raw", "qcow2", NULL); if (virCommandRun(cmd, NULL) < 0) goto skip; - if (!(canonqcow2 = virFileCanonicalizePath(absqcow2))) { - virReportOOMError(); - goto cleanup; - } /* Create a second qcow2 wrapping the first, to be sure that we * can correctly avoid insecure probing. */ @@ -214,10 +192,6 @@ testPrepImages(void) virCommandAddArg(cmd, "wrap"); if (virCommandRun(cmd, NULL) < 0) goto skip; - if (!(canonwrap = virFileCanonicalizePath(abswrap))) { - virReportOOMError(); - goto cleanup; - } /* Create a qed file. */ virCommandFree(cmd); @@ -227,10 +201,6 @@ testPrepImages(void) virCommandAddArg(cmd, "qed"); if (virCommandRun(cmd, NULL) < 0) goto skip; - if (!(canonqed = virFileCanonicalizePath(absqed))) { - virReportOOMError(); - goto cleanup; - } #ifdef HAVE_SYMLINK /* Create some symlinks in a sub-directory. */ @@ -731,7 +701,7 @@ mymain(void) /* Raw image, whether with right format or no specified format */ testFileData raw = { - .path = canonraw, + .path = absraw, .type = VIR_STORAGE_TYPE_FILE, .format = VIR_STORAGE_FILE_RAW, }; @@ -743,12 +713,12 @@ mymain(void) testFileData qcow2 = { .expBackingStoreRaw = "raw", .expCapacity = 1024, - .path = canonqcow2, + .path = absqcow2, .type = VIR_STORAGE_TYPE_FILE, .format = VIR_STORAGE_FILE_QCOW2, }; testFileData qcow2_as_raw = { - .path = canonqcow2, + .path = absqcow2, .type = VIR_STORAGE_TYPE_FILE, .format = VIR_STORAGE_FILE_RAW, }; @@ -772,7 +742,7 @@ mymain(void) testFileData wrap = { .expBackingStoreRaw = absqcow2, .expCapacity = 1024, - .path = canonwrap, + .path = abswrap, .type = VIR_STORAGE_TYPE_FILE, .format = VIR_STORAGE_FILE_QCOW2, }; @@ -795,7 +765,7 @@ mymain(void) testFileData wrap_as_raw = { .expBackingStoreRaw = absqcow2, .expCapacity = 1024, - .path = canonwrap, + .path = abswrap, .type = VIR_STORAGE_TYPE_FILE, .format = VIR_STORAGE_FILE_QCOW2, }; @@ -878,12 +848,12 @@ mymain(void) testFileData qed = { .expBackingStoreRaw = absraw, .expCapacity = 1024, - .path = canonqed, + .path = absqed, .type = VIR_STORAGE_TYPE_FILE, .format = VIR_STORAGE_FILE_QED, }; testFileData qed_as_raw = { - .path = canonqed, + .path = absqed, .type = VIR_STORAGE_TYPE_FILE, .format = VIR_STORAGE_FILE_RAW, }; @@ -892,12 +862,12 @@ mymain(void) /* directory */ testFileData dir = { - .path = canondir, + .path = absdir, .type = VIR_STORAGE_TYPE_DIR, .format = VIR_STORAGE_FILE_DIR, }; testFileData dir_as_raw = { - .path = canondir, + .path = absdir, .type = VIR_STORAGE_TYPE_DIR, .format = VIR_STORAGE_FILE_RAW, }; -- 2.20.1

2 3

[libvirt] [PATCH 0/4] Fix checking of blkiotune group name on update
by Ján Tomko 29 Mar '19

29 Mar '19

*** BLURB HERE *** *** BLURB THERE *** *** BLURB EVERYWHERE *** Ján Tomko (4): Revert "qemu: emit error when trying to update blkiotune group_name in qemuDomainChangeDiskLive" qemu: introduce CHECK_STREQ_NULLABLE in qemuDomainDiskChangeSupported qemu: error out on attempt to change blkiotune group name qemuDomainDiskChangeSupported: use CHECK_STREQ_NULLABLE more src/qemu/qemu_domain.c | 64 +++++++++++++++++------------------------- 1 file changed, 26 insertions(+), 38 deletions(-) -- 2.20.1

2 10

Re: [libvirt] [PATCH 1/2] snapshots: allow create --redefine to force a new configuration
by Richard Moore 29 Mar '19

29 Mar '19

> > On Wed, Mar 20, 2019 at 8:53 AM Peter Krempa <pkrempa redhat com> > wrote: > > I'm not persuaded that this workaround is necessary. > > Thanks Peter for taking a deeper look! > > And yes your summary seems correct. > I personally still like to give admins the ability to force configs, > but I'm ok if the general upstream opinion to that is no. > > I have asked the reporter - on the bug that I got - to chime in here > and > do the "convincing" as he is the affected person I think he is more > able > to do so - e.g. express the pain with the suggested workaround. > > -- > Christian Ehrhardt > Software Engineer, Ubuntu Server > Canonical Ltd Hi, I am the original fix requestor who Christian has put up with for quite some time now! I will try and summarise in brief; The initial problem was the machine type I had set in various snapshots/VM's had been deprecated and thus rendered the snapshots all unusable after upgrading O/S (that came packaged with newer libvirt/qemu versions.) I thus needed a simple, intuitive way to edit the snapshots to a machine type that was still supported, permanently. snapshot-edit seemed the appropriate tool for doing this, but alas it would not, a way was needed to force it to save the changes. One of my main reasons for using vm/snapshots is testing software install behaviour/compatibility in Windows, so I have a variety of Windows versions, each with snapshots taken with various levels of core 'things' installed (.net framework , C++ redistributables etc.) Thus the snapshots could be many years old. I need to be able to flick between them, so being able to permanently 'fix' them is my goal. The second issue was similarly being able to edit other 'simple' facets of the snapshots e.g. Attached disks, floppies or ISO's - or paths to them. Memory, if testing something that needed more than I ever invisaged way back when. MAC address (we have to register devices on the network via MAC, which expire, and can then render the snapshot unusable if I cannot change the MAC in the snapshot) I hope this makes sense, I have tried to be brief, I can elaborate further if needed. I think being able to edit existing snapshots simply, with once command, would be very beneficial to many people. (I think most people trying to do this would be aware of the risks, especially if the alternative is that the snapshot doesnt work anyway.) Cheers Richard Moore

1 0

[libvirt] [dockerfiles PATCH 0/2] Drop obsolete Dockerfiles
by Andrea Bolognani 29 Mar '19

29 Mar '19

The corresponding platforms are not longer supported by the libvirt-jenkins-ci project. Series pushed under the Dockerfile refresh rule. Andrea Bolognani (2): Drop Debian 8 Dockerfile Drop Ubuntu 16.04 Dockerfile buildenv-debian-8.Dockerfile | 82 --------------------------------- buildenv-ubuntu-16.Dockerfile | 85 ----------------------------------- 2 files changed, 167 deletions(-) delete mode 100644 buildenv-debian-8.Dockerfile delete mode 100644 buildenv-ubuntu-16.Dockerfile -- 2.20.1

1 2

[libvirt] [jenkins-ci PATCH 0/2] Drop Ubuntu 16.04
by Andrea Bolognani 29 Mar '19

29 Mar '19

I'm a bit conflicted about this one: on one hand, unlike Debian 8 we haven't really run into any issue with Ubuntu 16.04, we can still build almost all projects on it, and it doesn't require any crazy hacks to maintain; on the other hand, perhaps the reason why we still claim we can build almost all projects on it is because we don't actually run any build on it, either on CentOS CI or on Travis CI :D Anyway: here are the patches, let's see what other people think. Andrea Bolognani (2): Stop building on Ubuntu 16.04 guests: Drop Ubuntu 18.04 guests/host_vars/libvirt-ubuntu-16/docker.yml | 2 -- .../host_vars/libvirt-ubuntu-16/install.yml | 2 -- guests/host_vars/libvirt-ubuntu-16/main.yml | 22 ------------------- guests/inventory | 1 - guests/playbooks/build/jobs/defaults.yml | 1 - .../playbooks/build/projects/libvirt-dbus.yml | 5 +---- .../build/projects/libvirt-go-xml.yml | 1 - .../playbooks/build/projects/libvirt-go.yml | 1 - .../build/projects/libvirt-ocaml.yml | 1 - .../build/projects/libvirt-sandbox.yml | 1 - .../playbooks/build/projects/libvirt-tck.yml | 1 - guests/playbooks/build/projects/libvirt.yml | 1 - guests/playbooks/build/projects/osinfo-db.yml | 1 - .../playbooks/build/projects/virt-manager.yml | 4 +--- guests/vars/mappings.yml | 3 --- 15 files changed, 2 insertions(+), 45 deletions(-) delete mode 100644 guests/host_vars/libvirt-ubuntu-16/docker.yml delete mode 100644 guests/host_vars/libvirt-ubuntu-16/install.yml delete mode 100644 guests/host_vars/libvirt-ubuntu-16/main.yml -- 2.20.1

4 8

[libvirt] [jenkins-ci PATCH 0/2] Drop Debian 8
by Andrea Bolognani 29 Mar '19

29 Mar '19

See patch 1/2 and https://www.redhat.com/archives/libvir-list/2019-March/msg01864.html for the rationale. Andrea Bolognani (2): Stop building on Debian 8 guests: Drop Debian 8 guests/host_vars/libvirt-debian-8/docker.yml | 2 - guests/host_vars/libvirt-debian-8/install.yml | 2 - guests/host_vars/libvirt-debian-8/main.yml | 20 ----- guests/inventory | 1 - guests/playbooks/build/jobs/defaults.yml | 1 - .../playbooks/build/projects/libvirt-dbus.yml | 1 - .../build/projects/libvirt-go-xml.yml | 1 - .../playbooks/build/projects/libvirt-go.yml | 1 - .../build/projects/libvirt-ocaml.yml | 1 - .../build/projects/libvirt-sandbox.yml | 1 - .../playbooks/build/projects/libvirt-tck.yml | 1 - guests/playbooks/build/projects/libvirt.yml | 1 - guests/playbooks/build/projects/osinfo-db.yml | 1 - .../playbooks/build/projects/virt-manager.yml | 4 +- guests/playbooks/update/tasks/base.yml | 22 ------ guests/playbooks/update/tasks/bootstrap.yml | 6 -- .../templates/jessie-backports.preferences.j2 | 3 - .../templates/jessie-backports.sources.j2 | 1 - guests/vars/mappings.yml | 7 -- guests/vars/vault.yml | 78 +++++++++---------- jenkins/jobs/defaults.yaml | 1 - jenkins/projects/libvirt-dbus.yaml | 1 - jenkins/projects/libvirt-go-xml.yaml | 1 - jenkins/projects/libvirt-go.yaml | 1 - jenkins/projects/libvirt-ocaml.yaml | 1 - jenkins/projects/libvirt-sandbox.yaml | 1 - jenkins/projects/libvirt-tck.yaml | 1 - jenkins/projects/libvirt.yaml | 1 - jenkins/projects/osinfo-db.yaml | 1 - jenkins/projects/virt-manager.yaml | 4 +- 30 files changed, 41 insertions(+), 127 deletions(-) delete mode 100644 guests/host_vars/libvirt-debian-8/docker.yml delete mode 100644 guests/host_vars/libvirt-debian-8/install.yml delete mode 100644 guests/host_vars/libvirt-debian-8/main.yml delete mode 100644 guests/playbooks/update/templates/jessie-backports.preferences.j2 delete mode 100644 guests/playbooks/update/templates/jessie-backports.sources.j2 -- 2.20.1

2 3

[libvirt] [PATCH] snapashot: Improve logic of virDomainMomentMoveChildren
by Eric Blake 29 Mar '19

29 Mar '19

Even though Coverity can prove that 'last' is always set if the prior loop executed, gcc 8.0.1 cannot: CC conf/libvirt_conf_la-virdomainmomentobjlist.lo ../../src/conf/virdomainmomentobjlist.c: In function 'virDomainMomentMoveChildren': ../../src/conf/virdomainmomentobjlist.c:178:19: error: 'last' may be used uninitialized in this function [-Werror=maybe-uninitialized] last->sibling = to->first_child; ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~ Rewrite the loop to a form that should be easier for static analysis to work with. Fixes: ced0898f86bf Reported-by: Bjoern Walk <bwalk(a)linux.ibm.com> Signed-off-by: Eric Blake <eblake(a)redhat.com> --- Qualifies as a build-breaker fix, but I'd like a review before pushing. src/conf/virdomainmomentobjlist.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) diff --git a/src/conf/virdomainmomentobjlist.c b/src/conf/virdomainmomentobjlist.c index b9ca5b1318..5a217056d1 100644 --- a/src/conf/virdomainmomentobjlist.c +++ b/src/conf/virdomainmomentobjlist.c @@ -164,18 +164,17 @@ void virDomainMomentMoveChildren(virDomainMomentObjPtr from, virDomainMomentObjPtr to) { - virDomainMomentObjPtr child; - virDomainMomentObjPtr last; + virDomainMomentObjPtr child = from->first_child; - if (!from->first_child) - return; - for (child = from->first_child; child; child = child->sibling) { + while (child) { child->parent = to; - if (!child->sibling) - last = child; + if (!child->sibling) { + child->sibling = to->first_child; + break; + } + child = child->sibling; } to->nchildren += from->nchildren; - last->sibling = to->first_child; to->first_child = from->first_child; from->nchildren = 0; from->first_child = NULL; -- 2.20.1

3 3

Re: [libvirt] [PATCH 1/2] snapshots: allow create --redefine to force a new configuration
by Richard Moore 28 Mar '19

28 Mar '19

> > On Wed, Mar 20, 2019 at 8:53 AM Peter Krempa <pkrempa redhat com> > wrote: > > I'm not persuaded that this workaround is necessary. > > Thanks Peter for taking a deeper look! > > And yes your summary seems correct. > I personally still like to give admins the ability to force configs, > but I'm ok if the general upstream opinion to that is no. > > I have asked the reporter - on the bug that I got - to chime in here > and > do the "convincing" as he is the affected person I think he is more > able > to do so - e.g. express the pain with the suggested workaround. > > -- > Christian Ehrhardt > Software Engineer, Ubuntu Server > Canonical Ltd Hi, I am the original fix requestor who Christian has put up with for quite some time now! I will try and summarise in brief; The initial problem was the machine type I had set in various snapshots/VM's had been deprecated and thus rendered the snapshots all unusable after upgrading O/S (that came packaged with newer libvirt/qemu versions.) I thus needed a simple, intuitive way to edit the snapshots to a machine type that was still supported, permanently. snapshot-edit seemed the appropriate tool for doing this, but alas it would not, a way was needed to force it to save the changes. One of my main reasons for using vm/snapshots is testing software install behaviour/compatibility in Windows, so I have a variety of Windows versions, each with snapshots taken with various levels of core 'things' installed (.net framework , C++ redistributables etc.) Thus the snapshots could be many years old. I need to be able to flick between them, so being able to permanently 'fix' them is my goal. The second issue was similarly being able to edit other 'simple' facets of the snapshots e.g. Attached disks, floppies or ISO's - or paths to them. Memory, if testing something that needed more than I ever invisaged way back when. MAC address (we have to register devices on the network via MAC, which expire, and can then render the snapshot unusable if I cannot change the MAC in the snapshot) I hope this makes sense, I have tried to be brief, I can elaborate further if needed. I think being able to edit existing snapshots simply, with once command, would be very beneficial to many people. (I think most people trying to do this would be aware of the risks, especially if the alternative is that the snapshot doesn’t work anyway.) Cheers Richard Moore

1 0

[libvirt] [perl][PATCH 0/3] Add a bunch of constants
by Michal Privoznik 28 Mar '19

28 Mar '19

*** BLURB HERE *** Michal Privoznik (3): Register new error code constants Add VIR_FROM_DOMAIN_CHECKPOINT constant Add VIR_MIGRATE_PARALLEL and VIR_MIGRATE_PARAM_PARALLEL_CONNECTIONS constants Changes | 5 +++++ lib/Sys/Virt.xs | 6 ++++++ lib/Sys/Virt/Domain.pm | 10 ++++++++++ lib/Sys/Virt/Error.pm | 16 ++++++++++++++++ 4 files changed, 37 insertions(+) -- 2.19.2

2 6

[libvirt] Time to drop support for Debian 8 (Jessie)?
by Andrea Bolognani 28 Mar '19

28 Mar '19

We currently support Debian 8 (oldstable) along with Debian 9 (stable), but not without some compromises: * the libvirt-dbus, libvirt-ocaml and virt-manager projects do not support the platform at all because it ships outdated versions of some core components; * on the CI side of things, we are forced to drag in the JRE from backports in order to be able to run the Jenkins agent. All things considered, the situation has been fairly manageable up until now, but a couple of recent developments got me thinking that perhaps it's time to let Jessie go: * the distribution has been moved from the regular Debian infrastructure to archive.debian.org[1], a change which has resulted in the daily update run failing and would require investing time to adapt to; * Debian testing has recently entered the full freeze[2], which means the release of Debian 10 can hopefully be expected to happen within the next few month; * even if the Buster freeze period turned out to be exceedingly long, according to our platform support policy[3] we only promise to support a release for the two years after the most recent major release: given that Debian 9 was released in June 2017[4], we would be able to drop Debian 8 support in three months' time regardless of whether or not Debian 10 has been released in the meantime. Based on the above, I suggest we don't invest any time trying to keep Debian 8 chugging along only to drop it in June, and instead declare it as unsupported right now and move on with our lives. Thoughts? [1] https://lists.debian.org/debian-devel-announce/2019/03/msg00006.html [2] https://lists.debian.org/debian-devel-announce/2019/03/msg00003.html [3] https://libvirt.org/platforms.html [4] https://wiki.debian.org/DebianReleases -- Andrea Bolognani / Red Hat / Virtualization

2 8

[libvirt] [PATCH] virsh: Don't infloop on snapshot/storage_vol failure
by Eric Blake 28 Mar '19

28 Mar '19

Most of our completers used the pattern: if ((nITEM = virITEMListAll()) < 0) return NULL; but the virDomainSnapshot and virStorageVolume completers were instead using goto error because of a prior allocation. If the ListAll fails with -1, the cleanup label was running a loop of 'size_t i < int nITEM', which is an extreme waste of CPU cycles. Reported-by: Ján Tomko <jtomko(a)redhat.com> Signed-off-by: Eric Blake <eblake(a)redhat.com> --- This one is a bug fix, so worth having in 5.2. tools/virsh-completer.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/tools/virsh-completer.c b/tools/virsh-completer.c index c4adbb70d0..e9ef9b99f9 100644 --- a/tools/virsh-completer.c +++ b/tools/virsh-completer.c @@ -278,6 +278,7 @@ virshStorageVolNameCompleter(vshControl *ctl, virshControlPtr priv = ctl->privData; virStoragePoolPtr pool = NULL; virStorageVolPtr *vols = NULL; + int rc; int nvols = 0; size_t i = 0; char **ret = NULL; @@ -290,8 +291,9 @@ virshStorageVolNameCompleter(vshControl *ctl, if (!(pool = virshCommandOptPool(ctl, cmd, "pool", NULL))) return NULL; - if ((nvols = virStoragePoolListAllVolumes(pool, &vols, flags)) < 0) + if ((rc = virStoragePoolListAllVolumes(pool, &vols, flags)) < 0) goto error; + nvols = rc; if (VIR_ALLOC_N(ret, nvols + 1) < 0) goto error; @@ -631,6 +633,7 @@ virshSnapshotNameCompleter(vshControl *ctl, virshControlPtr priv = ctl->privData; virDomainPtr dom = NULL; virDomainSnapshotPtr *snapshots = NULL; + int rc; int nsnapshots = 0; size_t i = 0; char **ret = NULL; @@ -643,8 +646,9 @@ virshSnapshotNameCompleter(vshControl *ctl, if (!(dom = virshCommandOptDomain(ctl, cmd, NULL))) return NULL; - if ((nsnapshots = virDomainListAllSnapshots(dom, &snapshots, flags)) < 0) + if ((rc = virDomainListAllSnapshots(dom, &snapshots, flags)) < 0) goto error; + nsnapshots = rc; if (VIR_ALLOC_N(ret, nsnapshots + 1) < 0) goto error; -- 2.20.1

2 1

[libvirt] [jenkins-ci PATCH] jenkins: only send one email on build failure
by Daniel P. Berrangé 28 Mar '19

28 Mar '19

The current config causes jenkins to send a build on every single failure. This is way too noisy so change it to only send emails on the transition from good to bad. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- jenkins/jobs/autotools.yaml | 8 ++++---- jenkins/jobs/generic.yaml | 8 ++++---- jenkins/jobs/go.yaml | 4 ++-- jenkins/jobs/perl-modulebuild.yaml | 6 +++--- jenkins/jobs/python-distutils.yaml | 6 +++--- 5 files changed, 16 insertions(+), 16 deletions(-) diff --git a/jenkins/jobs/autotools.yaml b/jenkins/jobs/autotools.yaml index e365ec4..f04ec17 100644 --- a/jenkins/jobs/autotools.yaml +++ b/jenkins/jobs/autotools.yaml @@ -50,7 +50,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -89,7 +89,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -132,7 +132,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -174,5 +174,5 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false diff --git a/jenkins/jobs/generic.yaml b/jenkins/jobs/generic.yaml index 1990930..ca69513 100644 --- a/jenkins/jobs/generic.yaml +++ b/jenkins/jobs/generic.yaml @@ -45,7 +45,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -83,7 +83,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -121,7 +121,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -159,5 +159,5 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false diff --git a/jenkins/jobs/go.yaml b/jenkins/jobs/go.yaml index 3c14c35..1bdcb5f 100644 --- a/jenkins/jobs/go.yaml +++ b/jenkins/jobs/go.yaml @@ -45,7 +45,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -83,5 +83,5 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false diff --git a/jenkins/jobs/perl-modulebuild.yaml b/jenkins/jobs/perl-modulebuild.yaml index 933020a..7173493 100644 --- a/jenkins/jobs/perl-modulebuild.yaml +++ b/jenkins/jobs/perl-modulebuild.yaml @@ -47,7 +47,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false @@ -86,7 +86,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -127,5 +127,5 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false diff --git a/jenkins/jobs/python-distutils.yaml b/jenkins/jobs/python-distutils.yaml index 6b388f2..f56794c 100644 --- a/jenkins/jobs/python-distutils.yaml +++ b/jenkins/jobs/python-distutils.yaml @@ -47,7 +47,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false @@ -86,7 +86,7 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false - job-template: @@ -127,5 +127,5 @@ publishers: - email: recipients: '{obj:spam}' - notify-every-unstable-build: true + notify-every-unstable-build: false send-to-individuals: false -- 2.20.1

2 4

[libvirt] [jenkins-ci PATCH v2 1/2] guests: Update test dependencies for osinfo-db
by Fabiano Fidêncio 28 Mar '19

28 Mar '19

The new dependencies are: - python3 - python3-lxml - python3-pytest - python3-requests xmllint has been removed in favour of a own crafted test using python3-lxml Signed-off-by: Fabiano Fidêncio <fidencio(a)redhat.com> --- guests/vars/projects/osinfo-db.yml | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/guests/vars/projects/osinfo-db.yml b/guests/vars/projects/osinfo-db.yml index d9320a4..56bd77a 100644 --- a/guests/vars/projects/osinfo-db.yml +++ b/guests/vars/projects/osinfo-db.yml @@ -1,4 +1,7 @@ --- packages: - intltool - - xmllint + - python3 + - python3-lxml + - python3-pytest + - python3-requests -- 2.20.1

3 4

[libvirt] [PATCH] tests: Disable some tests on 32 bit systems
by Michal Privoznik 28 Mar '19

28 Mar '19

On 32 bit systems there are two tests failing currently: qemufirmwaretest and qemuxml2argvtest (not every test case in them is failing, only some of them). There are several problems: 1) in qemufirmwaretest it's 'QEMU FW precedence test' which is failing. This is because the code tests qemuFirmwareFetchConfigs() which tries to compile a list of some paths from the system and if one of the paths is an empty file it should not appear on the returned list. The code uses virFileLength() to query the file size which uses stat() under the hood. The test uses virfilewrapper to redirect stat() into qemufirmwaredata/ but for reasons beyond me real_stat() returns mangled buffer making the code see a file with a real size even for empty files. I've track this one to a problem in virfilewrapper.c which will call real___xstat() which points to a function in libvirt.so (!) which returns stat with 32bit members even though we're compiling with LARGE_FILE and therefore expect 64bit members in the stat struct. 2) qemuxml2argvtest has two tests failing: 2a) [aarch64-]os-firmware-* - These fail because of the same reason as described in 1) 2b) pseries-hostdevs-* - These fail because they again rely on some stat (lstat to be precise). When building the cmd line for these test cases the code does some search in /sys/bus/pci/... and since we have virpcimock everything works just fine. Except when virfilewrapper.c is linked in because then lstat() is taken from there and when it initializes itself it will lookup real_lstat() which will now point not to virpcimock but to glibc. Again, reasons beyond me. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Quite frankly, I hate this patch. But the problem is in our mocking, not in the actual code its testing. And I've tried everything I was able to come up with on how to fix this. So if you have any idea, I'm all ears. tests/qemufirmwaretest.c | 7 +++++++ tests/qemuxml2argvtest.c | 6 ++++++ 2 files changed, 13 insertions(+) diff --git a/tests/qemufirmwaretest.c b/tests/qemufirmwaretest.c index 2b5cbf649b..3332bf0a34 100644 --- a/tests/qemufirmwaretest.c +++ b/tests/qemufirmwaretest.c @@ -52,6 +52,8 @@ testParseFormatFW(const void *opaque) } +#if defined(__x86_64__) || defined(__amd64__) || defined(__aarch64__) +/* XXX Dirty hack, but mocking stat on 32bits is above my skills */ static int testFWPrecedence(const void *opaque ATTRIBUTE_UNUSED) { @@ -97,6 +99,7 @@ testFWPrecedence(const void *opaque ATTRIBUTE_UNUSED) return 0; } +#endif static int @@ -124,8 +127,12 @@ mymain(void) DO_PARSE_TEST("usr/share/qemu/firmware/61-ovmf.json"); DO_PARSE_TEST("usr/share/qemu/firmware/70-aavmf.json"); + +#if defined(__x86_64__) || defined(__amd64__) || defined(__aarch64__) + /* XXX Dirty hack, but mocking stat on 32bits is above my skills */ if (virTestRun("QEMU FW precedence test", testFWPrecedence, NULL) < 0) ret = -1; +#endif virFileWrapperClearPrefixes(); diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 0c0dcae197..364792d24d 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -2038,6 +2038,8 @@ mymain(void) DO_TEST("pseries-many-buses-2", QEMU_CAPS_DEVICE_SPAPR_PCI_HOST_BRIDGE, QEMU_CAPS_VIRTIO_SCSI); +# if defined(__x86_64__) || defined(__amd64__) || defined(__aarch64__) + /* XXX Dirty hack, but mocking stat on 32bits is above my skills */ DO_TEST("pseries-hostdevs-1", QEMU_CAPS_DEVICE_SPAPR_PCI_HOST_BRIDGE, QEMU_CAPS_VIRTIO_SCSI, @@ -2050,6 +2052,7 @@ mymain(void) QEMU_CAPS_DEVICE_SPAPR_PCI_HOST_BRIDGE, QEMU_CAPS_VIRTIO_SCSI, QEMU_CAPS_DEVICE_VFIO_PCI); +# endif DO_TEST("pseries-features", QEMU_CAPS_DEVICE_SPAPR_PCI_HOST_BRIDGE, @@ -3157,10 +3160,13 @@ mymain(void) DO_TEST_CAPS_ARCH_LATEST("x86_64-pc-graphics", "x86_64"); DO_TEST_CAPS_ARCH_LATEST("x86_64-q35-graphics", "x86_64"); +# if defined(__x86_64__) || defined(__amd64__) || defined(__aarch64__) + /* XXX Dirty hack, but mocking stat on 32bits is above my skills */ DO_TEST_CAPS_LATEST("os-firmware-bios"); DO_TEST_CAPS_LATEST("os-firmware-efi"); DO_TEST_CAPS_LATEST("os-firmware-efi-secboot"); DO_TEST_CAPS_ARCH_LATEST("aarch64-os-firmware-efi", "aarch64"); +# endif if (getenv("LIBVIRT_SKIP_CLEANUP") == NULL) virFileDeleteTree(fakerootdir); -- 2.19.2

3 2

[libvirt] Planning for next release
by Daniel Veillard 28 Mar '19

28 Mar '19

It's time to get a new release out. Simplest would be to enter freeze tomorrow, then we can plan an RC2 by Friday, and if everything looks good have the release on time next Monday. Hope this works for everybody, thanks, Daniel -- Daniel Veillard | Red Hat Developers Tools http://developer.redhat.com/ veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

3 3

[libvirt] [PATCH 00/16] snapshot refactoring (incremental backup saga)
by Eric Blake 28 Mar '19

28 Mar '19

A couple of these have seen the list before, but most of them are new. The bulk of this series is about refactoring snapshot_conf.c into smaller pieces that I can then reuse for implementing checkpoints, without having to open-code the hierarchy algorithms a second time (one of John's complaints against my v4 incremental backup series). I suspect that src/vz/ builds may break on one or more of these patches; I could not get a working vz build environment. Eric Blake (16): test: Avoid use-after-free on virDomainSnapshotDelete snapshot: Use accessors for virDomainSnapshot members snapshot: Create virDomainMoment base class vbox: Clean up some snapshot usage snapshot: Drop virDomainSnapshotDef.current snapshot: Track current snapshot in virDomainSnapshotObjList snapshot: Add accessors for updating snapshot list relations snapshot: Access snapshot def directly when needed snapshot: Refactor list filtering snapshot: Factor out virDomainMomentDef class snapshot: Switch type of virDomainSnapshotObj.def snapshot: Rename virDomainSnapshotObjPtr snapshot: Rename file for virDomainMomentObj snapshot: Move snapshot list code into generic file snapshot: Tweaks to support new bulk dumpxml/import API backup: Introduce virDomainCheckpointPtr include/libvirt/virterror.h | 6 +- src/util/virerror.c | 12 +- include/libvirt/libvirt.h | 6 +- src/conf/domain_conf.h | 1 - src/conf/moment_conf.h | 41 +++ src/conf/snapshot_conf.h | 23 +- src/conf/virconftypes.h | 12 +- src/conf/virdomainmomentobjlist.h | 95 +++++ src/conf/virdomainsnapshotobj.h | 50 --- src/conf/virdomainsnapshotobjlist.h | 33 +- src/datatypes.h | 83 ++++- src/qemu/qemu_command.h | 2 +- src/qemu/qemu_domain.h | 6 +- src/qemu/qemu_process.h | 4 +- src/conf/Makefile.inc.am | 6 +- src/conf/moment_conf.c | 40 +++ src/conf/snapshot_conf.c | 179 +++++----- src/conf/virdomainmomentobjlist.c | 521 ++++++++++++++++++++++++++++ src/conf/virdomainsnapshotobj.c | 123 ------- src/conf/virdomainsnapshotobjlist.c | 367 +++++++------------- src/datatypes.c | 132 ++++--- src/esx/esx_driver.c | 82 ++--- src/libvirt-domain-snapshot.c | 26 +- src/libvirt_private.syms | 21 +- src/qemu/qemu_command.c | 2 +- src/qemu/qemu_domain.c | 36 +- src/qemu/qemu_driver.c | 290 ++++++++-------- src/qemu/qemu_process.c | 4 +- src/remote/remote_daemon_dispatch.c | 4 +- src/remote/remote_driver.c | 4 +- src/rpc/gendispatch.pl | 2 +- src/test/test_driver.c | 174 ++++------ src/vbox/vbox_common.c | 207 +++++------ src/vz/vz_driver.c | 89 ++--- src/vz/vz_sdk.c | 4 +- tests/domainsnapshotxml2xmltest.c | 5 +- 36 files changed, 1613 insertions(+), 1079 deletions(-) create mode 100644 src/conf/moment_conf.h create mode 100644 src/conf/virdomainmomentobjlist.h delete mode 100644 src/conf/virdomainsnapshotobj.h create mode 100644 src/conf/moment_conf.c create mode 100644 src/conf/virdomainmomentobjlist.c delete mode 100644 src/conf/virdomainsnapshotobj.c -- 2.20.1

7 70

[libvirt] [PATCH] tests: Avoid writing into $HOME during virsh-snapshot
by Eric Blake 28 Mar '19

28 Mar '19

In a constrained CI environment, where it is intentional that attempts to write outside the current directory will fail, virsh-snapshot was failing: @@ -1,2 +1,3 @@ error: invalid argument: parent s3 for snapshot s2 not found error: marker +error: Failed to create '/home/travis/.cache/libvirt/virsh': Permission denied FAIL virsh-snapshot (exit status: 1) But we've already solved the problem in virsh-uriprecedence: tell virsh to use XDG locations pointing to somewhere we can write rather than its default of falling back to $HOME with the test being at risk of breaking due to the user's environment and/or unacceptably altering the user's normal cache. Hoist that solution into test-lib.sh, so that all scripts can use it as needed. Fixes: 280a2b41e Reported-by: Daniel P. Berrangé <berrange(a)redhat.com> Signed-off-by: Eric Blake <eblake(a)redhat.com> --- tests/test-lib.sh | 13 +++++++++++++ tests/virsh-snapshot | 2 ++ tests/virsh-uriprecedence | 12 +----------- 3 files changed, 16 insertions(+), 11 deletions(-) diff --git a/tests/test-lib.sh b/tests/test-lib.sh index 49e8d22095..64f0b0d401 100644 --- a/tests/test-lib.sh +++ b/tests/test-lib.sh @@ -222,6 +222,19 @@ mkfifo_or_skip_() fi } +# Create mock XDG files/directories to avoid permission problems. +# As it points inside $test_dir_, it is automatically cleaned. +mock_xdg_() +{ + export XDG_CONFIG_HOME="$test_dir_/.config" + export XDG_CACHE_HOME="$test_dir_/.cache" + export XDG_RUNTIME_HOME="XDG_CACHE_HOME" + + mkdir -p "$XDG_CONFIG_HOME/libvirt" "$XDG_CONFIG_HOME/virsh" + mkdir -p "$XDG_CACHE_HOME/libvirt" "$XDG_CACHE_HOME/virsh" + mkdir -p "$XDG_RUNTIME_HOME/libvirt" "$XDG_RUNTIME_HOME/virsh" +} + test_dir_=$(pwd) this_test_() { echo "./$0" | sed 's,.*/,,'; } diff --git a/tests/virsh-snapshot b/tests/virsh-snapshot index fb8a99dd43..cb498cf54e 100755 --- a/tests/virsh-snapshot +++ b/tests/virsh-snapshot @@ -26,6 +26,8 @@ fi fail=0 +mock_xdg_ || framework_failure + # The test driver loses states between restarts, so we perform a script # with some convenient markers for later post-processing of output. $abs_top_builddir/tools/virsh --connect test:///default >out 2>err ' diff --git a/tests/virsh-uriprecedence b/tests/virsh-uriprecedence index 564e3dc42c..fd6ce108c0 100755 --- a/tests/virsh-uriprecedence +++ b/tests/virsh-uriprecedence @@ -11,17 +11,7 @@ virsh_cmd="$virsh_bin" counter=0 ret=0 -cleanup_() { rm -rf "$tmphome"; } - -# Create all mock files/directories to avoid permission problems -tmphome="$PWD/tmp_home" -export XDG_CONFIG_HOME="$tmphome/.config" -export XDG_CACHE_HOME="$tmphome/.cache" -export XDG_RUNTIME_HOME="XDG_CACHE_HOME" - -mkdir -p "$XDG_CONFIG_HOME/libvirt" "$XDG_CONFIG_HOME/virsh" -mkdir -p "$XDG_CACHE_HOME/libvirt" "$XDG_CACHE_HOME/virsh" -mkdir -p "$XDG_RUNTIME_HOME/libvirt" "$XDG_RUNTIME_HOME/virsh" +mock_xdg_ || framework_failure is_uri_good() { -- 2.20.1

3 4

[libvirt] [PATCH 0/3] qemu: clean up USB address releasing
by Peter Krempa 28 Mar '19

28 Mar '19

Peter Krempa (3): qemu: hotplug: Don't release USB address twice when removing disk qemu: Always use 'alias' in warning message when removing USB address qemu: address: Stop reporting warning when USB address can't be released src/qemu/qemu_domain_address.c | 9 ++---- src/qemu/qemu_domain_address.h | 3 +- src/qemu/qemu_hotplug.c | 53 +++++++++++++++++----------------- 3 files changed, 29 insertions(+), 36 deletions(-) -- 2.20.1

2 4

Re: [libvirt] Can jobs suck like qemu-pr-helper does be transfered to libvirtd?
by Zhangbo (Oscar) 28 Mar '19

28 Mar '19

>> Hi all: >> qemu-pr-helper exits to help qemu do the high-privileged scsi related jobs. >LIBVIRTD is responsible to launch qemu-pr-helper and qemu, and set >selinux/DAC labels for them and their socket. >> >> # >> # >> # >> # >> # ___________ >> # ___________|libvirtd |__________ >> # | |___________| | >> # | | | >> # | | | >> # _________|________ _____|______ ___|____ >> # | qemu-pr-helper |__| vm1.sock |_____| qemu1 | >> # |__________________| |____________| |________| >> >> There may be other jobs quite like qemu-pr-helper, shall we make them as >"complex" as the qemu-pr-helper scheme? >> Will it be OK to just let qemu send an EVENT to libvirtd, and let libvirtd do the >SCSI/other jobs? Will it be OK if we remove qemu-pr-helper and similar >processes? >> What's the disadvantage if we let libvirtd do there high privileged jobs instead >qemu-pr-helper-like processes? > >No what you describe is not desired. Any process which is vital for the >runtime of the VM needs to stay as simple and minimal as possible and >separate from libvirtd. > >Libvirtd is a complex big application dealing with multiple things at >once and also the architecture allows for libvirtd to be stopped and >upgraded. > >This does not play well with the fact that processes as the PR helper >are always required. > >Merging them into libvirtd would make the VM stop until libvirtd is >running again. Additionally if any of the operations require persistent >kernel state as e.g. file descriptors, this would be impossible as >stopping libvirtd process would close the FDs which may be then >impossible to reopen properly e.g. due to state. Thanks! Besides these reasons above, will it weaken security if we let libvirtd do these jobs? For example, Such sayings, like, "libvirtd would become the focus suffered from attacking forces", make sense?

1 0

[libvirt] [PATCH] qemu: don't fail on destroy if domain is inactive
by Nikolay Shirokovskiy 28 Mar '19

28 Mar '19

Mgmt can not track if domain is already inactive before calling destroy because domain can become inactive because of crash/shutdown from guest. Thus it is make sense to report success in this case. Another option is to return special error code but this is a bit more complicated. Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy(a)virtuozzo.com> --- src/qemu/qemu_driver.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 62d8d97..0789efc 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2172,8 +2172,10 @@ qemuDomainDestroyFlags(virDomainPtr dom, if (virDomainDestroyFlagsEnsureACL(dom->conn, vm->def) < 0) goto cleanup; - if (virDomainObjCheckActive(vm) < 0) + if (!virDomainObjIsActive(vm)) { + ret = 0; goto cleanup; + } state = virDomainObjGetState(vm, &reason); starting = (state == VIR_DOMAIN_PAUSED && -- 1.8.3.1

3 7

[libvirt] [PATCH] util: suppress unimportant ovs-vsctl errors when getting interface stats
by Laine Stump 28 Mar '19

28 Mar '19

commit edaf13565 modified the stats retrieval for OVS interfaces to not fail when one of the fields was unrecognized by the ovs-vsctl command, but ovs-vsctl was still returning an error, and libvirt was cluttering the logs with these inconsequential error messages. This patch modifies the GET_STAT macro to add "--if-exists" to the ovs-vsctl command, which causes it to return an empty string (and exit with success) if the requested statistic isn't in its database, thus eliminating the ugly error messages from the log. Resolves: https://bugzilla.redhat.com/1683175 Signed-off-by: Laine Stump <laine(a)laine.org> --- src/util/virnetdevopenvswitch.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/util/virnetdevopenvswitch.c b/src/util/virnetdevopenvswitch.c index 4fa3a5742a..e5a45f59ec 100644 --- a/src/util/virnetdevopenvswitch.c +++ b/src/util/virnetdevopenvswitch.c @@ -335,10 +335,10 @@ virNetDevOpenvswitchInterfaceStats(const char *ifname, virCommandFree(cmd); \ cmd = virCommandNew(OVSVSCTL); \ virNetDevOpenvswitchAddTimeout(cmd); \ - virCommandAddArgList(cmd, "get", "Interface", ifname, \ - "statistics:" name, NULL); \ + virCommandAddArgList(cmd, "--if-exists", "get", "Interface", \ + ifname, "statistics:" name, NULL); \ virCommandSetOutputBuffer(cmd, &output); \ - if (virCommandRun(cmd, NULL) < 0) { \ + if (virCommandRun(cmd, NULL) < 0 || !*output || *output == '\n') { \ stats->member = -1; \ } else { \ if (virStrToLong_ll(output, &tmp, 10, &stats->member) < 0 || \ -- 2.20.1

2 1

[libvirt] [PATCH] tools: Tweak workding for iothreadset
by John Ferlan 27 Mar '19

27 Mar '19

Update the wording to note the values for polling are purely dynamic and won't be saved across domain stop/(re)start or save/restore. Signed-off-by: John Ferlan <jferlan(a)redhat.com> --- tools/virsh.pod | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/tools/virsh.pod b/tools/virsh.pod index a0fe949c55..f00564fa5f 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -1772,7 +1772,10 @@ reach the maximum polling time. If a 0 (zero) is provided, then the default factor will be used. The I<--poll-shrink> is the quotient by which the current polling time will be reduced in order to get below the maximum polling interval. If a 0 (zero) is provided, then -the default quotient will be used. +the default quotient will be used. The polling values are purely dynamic +for a running guest. Saving, destroying, stopping, etc. the guest will +result in the polling values returning to hypervisor defaults at the +next start, restore, etc. If I<--live> is specified, affect a running guest. If the guest is not running an error is returned. -- 2.20.1

2 2

[libvirt] [PATCH v2 0/2] qemu_capabilities; Drop virQEMUCapsSetVAList
by Michal Privoznik 27 Mar '19

27 Mar '19

Technically, this is v2 of: https://www.redhat.com/archives/libvir-list/2019-March/msg01778.html but is this one implements different approach. Plus it breaks dependency between two unrelated enums that I've noticed when debugging broken build on a 32bit arch. Michal Prívozník (2): qemu_capabilities; Drop virQEMUCapsSetVAList qemuxml2argvtest: Drop dependency between testInfoArgName and virQEMUCapsFlags enums src/qemu/qemu_capabilities.c | 15 +++---------- src/qemu/qemu_capabilities.h | 2 -- tests/qemuxml2argvtest.c | 41 +++++++++++++++++++++++------------- 3 files changed, 29 insertions(+), 29 deletions(-) -- 2.19.2

2 4

[libvirt] [PATCH] qemu_capabilities: Pass pointer to va_list in virQEMUCapsSetVAList
by Michal Privoznik 27 Mar '19

27 Mar '19

There is one specific caller (testInfoSetArgs() in qemuxml2argvtest.c) which expect the va_list argument to change after returning from the virQEMUCapsSetVAList() function. However, since we are passing plain va_list this is not guaranteed. The man page of stdarg(3) says: If ap is passed to a function that uses va_arg(ap,type), then the value of ap is undefined after the return of that function. (ap is a variable of type va_list) I've seen this in action in fact: on i686 the qemuxml2argvtest fails on the second test case because testInfoSetArgs() sees ARG_QEMU_CAPS and callse virQEMUCapsSetVAList to process the capabilities (in this case there's just one QEMU_CAPS_SECCOMP_BLACKLIST). But since the changes are not reflected in the caller, in the next iteration testInfoSetArgs() sees the qemu capability and not ARG_END. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- This passed successfully on x86_64 and i686 in my testing. src/qemu/qemu_capabilities.c | 6 +++--- src/qemu/qemu_capabilities.h | 2 +- tests/qemuxml2argvtest.c | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index c5954edaf0..e182a2c2e5 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -1665,11 +1665,11 @@ virQEMUCapsSet(virQEMUCapsPtr qemuCaps, void virQEMUCapsSetVAList(virQEMUCapsPtr qemuCaps, - va_list list) + va_list *list) { int flag; - while ((flag = va_arg(list, int)) < QEMU_CAPS_LAST) + while ((flag = va_arg(*list, int)) < QEMU_CAPS_LAST) ignore_value(virBitmapSetBit(qemuCaps->flags, flag)); } @@ -1680,7 +1680,7 @@ virQEMUCapsSetList(virQEMUCapsPtr qemuCaps, ...) va_list list; va_start(list, qemuCaps); - virQEMUCapsSetVAList(qemuCaps, list); + virQEMUCapsSetVAList(qemuCaps, &list); va_end(list); } diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index 7625d754a3..ceab8c9154 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -519,7 +519,7 @@ void virQEMUCapsSet(virQEMUCapsPtr qemuCaps, virQEMUCapsFlags flag) ATTRIBUTE_NONNULL(1); void virQEMUCapsSetVAList(virQEMUCapsPtr qemuCaps, - va_list list) ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2); + va_list *list) ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2); void virQEMUCapsSetList(virQEMUCapsPtr qemuCaps, ...) ATTRIBUTE_NONNULL(1); void virQEMUCapsClear(virQEMUCapsPtr qemuCaps, diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 6e7d1b0b9a..a7a8d77e1d 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -650,7 +650,7 @@ testInfoSetArgs(struct testInfo *info, case ARG_QEMU_CAPS: if (qemuCaps || !(qemuCaps = virQEMUCapsNew())) goto cleanup; - virQEMUCapsSetVAList(qemuCaps, argptr); + virQEMUCapsSetVAList(qemuCaps, &argptr); break; case ARG_GIC: -- 2.19.2

4 6