January 2021 - Devel - libvirt List Archives

[PATCH] docs/interop/qmp-spec: Document the request queue limit
by Markus Armbruster 02 Feb '21

02 Feb '21

Signed-off-by: Markus Armbruster <armbru(a)redhat.com> --- docs/interop/qmp-spec.txt | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/docs/interop/qmp-spec.txt b/docs/interop/qmp-spec.txt index cdf5842555..b0e8351d5b 100644 --- a/docs/interop/qmp-spec.txt +++ b/docs/interop/qmp-spec.txt @@ -133,9 +133,11 @@ to pass "id" with out-of-band commands. Passing it with all commands is recommended for clients that accept capability "oob". If the client sends in-band commands faster than the server can -execute them, the server will stop reading the requests from the QMP -channel until the request queue length is reduced to an acceptable -range. +execute them, the server will stop reading requests until the request +queue length is reduced to an acceptable range. + +To ensure commands to be executed out-of-band get read and executed, +the client should have at most eight in-band commands in flight. Only a few commands support out-of-band execution. The ones that do have "allow-oob": true in output of query-qmp-schema. -- 2.26.2

2 2

[PATCH] virsh: Simplify @flags handing in cmdSetmem() and cmdSetmaxmem()
by Michal Privoznik 02 Feb '21

02 Feb '21

What code tries to achieve is that if no flags were provided to either 'setmem' or 'setmaxmem' commands then the old (no flags) API is called to be able to communicate with older daemons. Well, the code can be simplified a bit. Note that with this change the old no flag version of APIs is used more often. Previously if --current argument was given it resulted in *Flags() version to be called even though it is not necessary - VIR_DOMAIN_AFFECT_CURRENT is implied. Therefore, this change in fact allows virsh to talk with broader set of daemons. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- tools/virsh-domain.c | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index 2bb136333f..9746117bdb 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -9018,9 +9018,6 @@ cmdSetmem(vshControl *ctl, const vshCmd *cmd) flags |= VIR_DOMAIN_AFFECT_CONFIG; if (live) flags |= VIR_DOMAIN_AFFECT_LIVE; - /* none of the options were specified */ - if (!current && !live && !config) - flags = -1; if (!(dom = virshCommandOptDomain(ctl, cmd, NULL))) return false; @@ -9037,7 +9034,7 @@ cmdSetmem(vshControl *ctl, const vshCmd *cmd) } kibibytes = VIR_DIV_UP(bytes, 1024); - if (flags == -1) { + if (flags == 0) { if (virDomainSetMemory(dom, kibibytes) != 0) ret = false; } else { @@ -9090,7 +9087,7 @@ cmdSetmaxmem(vshControl *ctl, const vshCmd *cmd) bool config = vshCommandOptBool(cmd, "config"); bool live = vshCommandOptBool(cmd, "live"); bool current = vshCommandOptBool(cmd, "current"); - unsigned int flags = VIR_DOMAIN_AFFECT_CURRENT | VIR_DOMAIN_MEM_MAXIMUM; + unsigned int flags = VIR_DOMAIN_AFFECT_CURRENT; VSH_EXCLUSIVE_OPTIONS_VAR(current, live); VSH_EXCLUSIVE_OPTIONS_VAR(current, config); @@ -9099,9 +9096,6 @@ cmdSetmaxmem(vshControl *ctl, const vshCmd *cmd) flags |= VIR_DOMAIN_AFFECT_CONFIG; if (live) flags |= VIR_DOMAIN_AFFECT_LIVE; - /* none of the options were specified */ - if (!current && !live && !config) - flags = -1; if (!(dom = virshCommandOptDomain(ctl, cmd, NULL))) return false; @@ -9118,13 +9112,13 @@ cmdSetmaxmem(vshControl *ctl, const vshCmd *cmd) } kibibytes = VIR_DIV_UP(bytes, 1024); - if (flags == -1) { + if (flags == 0) { if (virDomainSetMaxMemory(dom, kibibytes) != 0) { vshError(ctl, "%s", _("Unable to change MaxMemorySize")); ret = false; } } else { - if (virDomainSetMemoryFlags(dom, kibibytes, flags) < 0) { + if (virDomainSetMemoryFlags(dom, kibibytes, flags | VIR_DOMAIN_MEM_MAXIMUM) < 0) { vshError(ctl, "%s", _("Unable to change MaxMemorySize")); ret = false; } -- 2.26.2

2 4

[PATCH 00/12] Hyper-V serial ports, NICs, and screeshots
by Matt Coleman 02 Feb '21

02 Feb '21

This patchset makes the following changes to the Hyper-V driver: * enable XML parsing and creation of serial ports and NICs * implement several networking APIs * implement screenshots Matt Coleman (12): hyperv: XML parsing of serial ports hyperv: add support for creating serial devices hyperv: XML parsing of Ethernet adapters hyperv: add support for creating network adapters hyperv: implement connectListAllNetworks and connectNumOfNetworks hyperv: implement networkLookupByName and networkLookupByUUID hyperv: implement connectNumOfDefinedNetworks and connectListDefinedNetworks hyperv: implement networkGetAutostart, networkIsActive, and networkIsPersistent hyperv: implement networkGetXMLDesc hyperv: implement domainScreenshot hyperv: provide a more detailed error message for WSMan faults news: implement new Hyper-V APIs NEWS.rst | 12 + po/POTFILES.in | 1 + src/hyperv/hyperv_driver.c | 620 ++++++++++++++++++++++++++ src/hyperv/hyperv_network_driver.c | 242 ++++++++++ src/hyperv/hyperv_network_driver.h | 26 ++ src/hyperv/hyperv_wmi.c | 45 +- src/hyperv/hyperv_wmi.h | 12 + src/hyperv/hyperv_wmi_classes.h | 36 ++ src/hyperv/hyperv_wmi_generator.input | 309 +++++++++++++ src/hyperv/meson.build | 1 + 10 files changed, 1301 insertions(+), 3 deletions(-) create mode 100644 src/hyperv/hyperv_network_driver.c create mode 100644 src/hyperv/hyperv_network_driver.h -- 2.30.0

3 17

[PATCH] virNetDevOpenvswitchGetVhostuserIfname: Remove a single '\n' from ifname
by Yalei Li 02 Feb '21

02 Feb '21

Ovs-vsctl returns a newline result. Signed-off-by: Yalei Li <liyl43(a)chinatelecom.cn> --- src/util/virnetdevopenvswitch.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/util/virnetdevopenvswitch.c b/src/util/virnetdevopenvswitch.c index f9b3369b2a..bd840bd3b7 100644 --- a/src/util/virnetdevopenvswitch.c +++ b/src/util/virnetdevopenvswitch.c @@ -575,6 +575,7 @@ virNetDevOpenvswitchGetVhostuserIfname(const char *path, return 0; } + virStringTrimOptionalNewline(*ifname); if (virNetDevOpenvswitchMaybeUnescapeReply(*ifname) < 0) { VIR_FREE(*ifname); return -1; -- 2.27.0

2 2

[PATCH] Revert "remote: Add libvirtd dependency to virt-guest-shutdown.target"
by Jim Fehlig 02 Feb '21

02 Feb '21

Further testing revealed commit f035f53baa regresses Debian bug 955216 https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=955216 Restarting libvirt-guests on libvirtd restart is worse than the original dependency issue, so revert the commit until a better solution is found. This reverts commit f035f53baa2e5dc00b8e866e594672a90b4cea78. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- src/remote/virt-guest-shutdown.target | 1 - 1 file changed, 1 deletion(-) diff --git a/src/remote/virt-guest-shutdown.target b/src/remote/virt-guest-shutdown.target index e2efa3e63a..25d4aaa267 100644 --- a/src/remote/virt-guest-shutdown.target +++ b/src/remote/virt-guest-shutdown.target @@ -1,4 +1,3 @@ [Unit] Description=Libvirt guests shutdown -Requires=libvirtd.service Documentation=https://libvirt.org -- 2.29.2

3 2

[PATCH v2] docs: Clarify the documentation of the <css> elements
by Thomas Huth 02 Feb '21

02 Feb '21

The channel subsystem elements describe a channel in the I/O subsystem of a s390x machine, and not a normal device (like a disk or network card). Reword the documentation here to make it this a little bit clearer. Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1898074 Reviewed-by: Boris Fiuczynski <fiuczy(a)linux.ibm.com> Signed-off-by: Thomas Huth <thuth(a)redhat.com> --- v2: "channel" -> "subchannel" as suggested by Cornelia docs/formatnode.html.in | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/formatnode.html.in b/docs/formatnode.html.in index 1010a37a3d..5c7286df8a 100644 --- a/docs/formatnode.html.in +++ b/docs/formatnode.html.in @@ -396,15 +396,15 @@ </dl> </dd> <dt><code>css</code></dt> - <dd>Describes a Channel SubSystem (CSS) device commonly found on - the S390 architecture. Sub-elements include: + <dd>Describes a subchannel in the Channel SubSystem (CSS) commonly + found on the S390 architecture. Sub-elements include: <dl> <dt><code>cssid</code></dt> <dd>The channel subsystem identifier.</dd> <dt><code>ssid</code></dt> <dd>The subchannel-set identifier.</dd> <dt><code>devno</code></dt> - <dd>The device number.</dd> + <dd>The subchannel number.</dd> <dt><code>capability</code></dt> <dd> This optional element can occur multiple times. If it -- 2.27.0

3 2

[PATCH] virfile: workaround for when posix_fallocate() is not supported by FS
by Roman Bogorodskiy 01 Feb '21

01 Feb '21

posix_fallocate() might be not supported by a filesystem, for example, it's not supported by ZFS. In that case it fails with return code 22 (EINVAL), and thus safezero_posix_fallocate() returns -1. As safezero_posix_fallocate() is the first function tried by safezero() and it tries other functions only when it returns -2, it fails immediately without falling back to other methods, such as safezero_slow(). Fix that by returning -2 if posix_fallocate() returns EINVAL, to give safezero() a chance to try other functions. Signed-off-by: Roman Bogorodskiy <bogorodskiy(a)gmail.com> --- src/util/virfile.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/src/util/virfile.c b/src/util/virfile.c index 609cafdd34..67a350deb3 100644 --- a/src/util/virfile.c +++ b/src/util/virfile.c @@ -1119,6 +1119,16 @@ safezero_posix_fallocate(int fd, off_t offset, off_t len) int ret = posix_fallocate(fd, offset, len); if (ret == 0) return 0; + else if (ret == EINVAL) + /* EINVAL is returned when either: + - Operation is not supported by the underlying filesystem, + - offset or len argument values are invalid. + Assuming that offset and len are valid, this error means + the operation is not supported, and we need to fall back + to other methods. + */ + return -2; + errno = ret; return -1; } -- 2.30.0

2 1

[PULL 00/13] Misc patches
by Daniel P. Berrangé 30 Jan '21

30 Jan '21

The following changes since commit 5101d00d2f1138a73344dc4833587f76d7a5fa5c: Merge remote-tracking branch 'remotes/vivier2/tags/trivial-branch-for-6.0-p= ull-request' into staging (2021-01-29 10:10:43 +0000) are available in the Git repository at: https://gitlab.com/berrange/qemu tags/misc-fixes-pull-request for you to fetch changes up to ecb98f5c7589ba8ecd15c8b1baa2ec7192e47c75: tests: Replace deprecated ASN1 code (2021-01-29 17:07:53 +0000) ---------------------------------------------------------------- * Replace --enable/disable-git-update with --with-git-submodules to allow improved control over use of git submodules * Deprecate the -enable-fips option * Ensure docs use prefer format for bool options * Clarify platform support rules * Misc fixes to keymap conversions * Fix misc problems on macOS ---------------------------------------------------------------- Dan Streetman (1): configure: replace --enable/disable-git-update with --with-git-submodules Daniel P. Berrang=C3=A9 (5): os: deprecate the -enable-fips option and QEMU's FIPS enforcement Prefer 'on' | 'off' over 'yes' | 'no' for bool options docs: simplify and clarify the platform support rules docs: fix missing backslash in certtool shell example ui: update keycodemapdb submodule commit Kevin Wolf (3): crypto: Move USER_CREATABLE to secret_common base class crypto: Forbid broken unloading of secrets crypto: Fix memory leaks in set_loaded for tls-* Stefan Weil (2): tests: Fix runtime error in test-authz-pam tests: Replace deprecated ASN1 code shiliyang (2): crypto: Fix some code style problems, add spaces around operator crypto: Add spaces around operator Makefile | 24 ++----------- configure | 51 +++++++++++++++++++------- crypto/aes.c | 6 ++-- crypto/desrfb.c | 4 +-- crypto/secret.c | 14 -------- crypto/secret_common.c | 21 +++++++++-- crypto/secret_keyring.c | 14 -------- crypto/tlscredsanon.c | 3 +- crypto/tlscredspsk.c | 3 +- crypto/tlscredsx509.c | 5 ++- docs/system/build-platforms.rst | 63 ++++++++++++--------------------- docs/system/deprecated.rst | 12 +++++++ docs/system/tls.rst | 2 +- docs/system/vnc-security.rst | 10 +++--- include/authz/listfile.h | 2 +- os-posix.c | 3 ++ qemu-options.hx | 4 +-- scripts/git-submodule.sh | 34 +++++++++++++----- tests/crypto-tls-x509-helpers.c | 10 +++--- tests/crypto-tls-x509-helpers.h | 2 +- tests/pkix_asn1_tab.c | 2 +- tests/qemu-iotests/233 | 4 +-- tests/test-authz-pam.c | 10 +++++- ui/keycodemapdb | 2 +- ui/meson.build | 3 +- 25 files changed, 161 insertions(+), 147 deletions(-) --=20 2.29.2

3 15

[libvirt PATCH 00/19] Overhaul test/commandhelper.c
by Tim Wiederhake 30 Jan '21

30 Jan '21

I stumbled upon a buffer overflow / stack smash present in "test/commandhelper.c" that could be triggered by e.g. $ ./tests/commandhelper --readfd 0 --readfd 0 --readfd 0 --readfd x Could not parse fd x *** stack smashing detected ***: terminated Aborted (core dumped) This series cleans up the file, fixes the buffer overflow and converts (most) memory handling to g_auto*. Note that it does not touch the "prevent malloc with zero size" issue discussed in https://www.redhat.com/archives/libvir-list/2021-January/msg01160.html, this will be done in the other series. Please feel free to comment on whether the copyright year in the file's header should be updated and whether a prefix for the function names and the new type is required. Cheers, Tim Tim Wiederhake (19): commandhelper: Remove origenv variable commandhelper: Remove numpollfds variable commandhelper: Simplify envsort commandhelper: Consolidate error paths commandhelper: Consolidate argument parsing commandhelper: Split argument parsing and printing commandhelper: Factor out parseArguments commandhelper: Factor out printArguments commandhelper: Factor out printEnvironment commandhelper: Factor out printFds commandhelper: Factor out printDaemonization commandhelper: Factor out printCwd commandhelper: Factor out printInput commandhelper: Make number of fds variable in printInput commandhelper: Make number of fds variable in parseArguments commandhelper: Convert parseArguments to g_auto* commandhelper: Convert printEnvironment to g_auto* commandhelper: Convert printCwd to g_auto* commandhelper: Convert main to g_auto* tests/commandhelper.c | 295 +++++++++++++++++++++++++++--------------- 1 file changed, 188 insertions(+), 107 deletions(-) -- 2.26.2

4 28

[libvirt PATCH v2 0/2] stop using netcf for the interface driver backend
by Laine Stump 30 Jan '21

30 Jan '21

V1 here: https://www.redhat.com/archives/libvir-list/2021-January/msg00922.html A short version of the cover letter from V1: this is a followup to my proposal to stop using netcf for the interface driver backend in Fedora/RHEL/CentOS builds and use the udev backend instead (it turns out Ubuntu already disabled netcf in 2018). Changes in V2: * removed the patch that made the default netcf=disabled even when netcf-devel was found on the host. If someone has netcf-devel installed and still wants to build without netcf, then can add "-Dnetcf=disabled" on the meson commandline. * Made the specfile changes more intelligent: * instead of hardcoding -Dnetcf=disabled, we now have a variable %{with_netcf} that is set to 1 for current Fedora (< 34) and current RHEL (< 9) but will be set to 0 for future Fedora/RHEL. This way the behavior on current OS releases will remain the same even for future libvirt. * it is possible to for netcf support off even in current/older OS releases by adding "--without netcf" to the rpmbuild commandline. I think at this point I would be comfortable pushing these patches, unless someone has misgivings about it... Laine Stump (2): build: support explicitly disabling netcf rpm: disable netcf for the interface driver in rpm build on new targets libvirt.spec.in | 22 +++++++++++++++++----- meson.build | 10 ++++++---- 2 files changed, 23 insertions(+), 9 deletions(-) -- 2.29.2

4 8