- Devel - libvirt List Archives

[libvirt] [GSoC] Project of libvirt/qemu fuzzing
by Dan 30 May '17

30 May '17

Dear all, The project of qemu command line fuzzing has been accepted as a GSoC project [1] [2]. As a student participating Google Summer of Code activity, I am extremely exitited to get started today on May 30th, 2017. During the past months, I have received tremendous guidance from my mentors as well as many other contributors on the mailinglist. I look forward to contributing to the community and learning a lot over the summer. Any advice, comment, feedback, suggestion to my emails/commit, would always be highly appreciated and more than welcome. Thank you all for your time, Daniel Liu [1]. https://summerofcode.withgoogle.com/projects/#5088017038442496 [2]. https://wiki.libvirt.org/page/Google_Summer_of_Code_Ideas#QEMU_command_line…

2 2

[libvirt] [PATCH] maint: add sanitizers to the build process
by claudioandre.br＠gmail.com 30 May '17

30 May '17

From: Claudio André <claudioandre.br(a)gmail.com> Sanitizers are based on compile-time instrumentation. They are available in gcc and clang for a range of supported operation systems and platforms. More info at: https://github.com/google/sanitizers The address sanitizer finds bugs related to addressing memory: use after free, heap buffer overflow, stack buffer overflow, memory leaks, ... The undefined behavior sanitizer detects situations not prescribed by the language specification: bound violations, data overflows, ... The llvm.org states that Sanitizers have found thousands of bugs everywhere. Sanitizers running during CI can prevent bugs from taking up residence. A helper tool to keep bugs out. --- - I mean CI (in general) not only Travis; - The functionality is not tied to CI; it is useful for local testing; - A way to think about this (including the ongoing GSOC): - Phase 1: test with Sanitizers to achieve basic code sanity; - Phase 2: use fuzzing for stronger security & reliability; - MISSING: should I add the flag to which Makefile.am? Or, what do you guys think about this? configure.ac | 2 ++ m4/virt-compile-sanitizer.m4 | 51 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 53 insertions(+) create mode 100644 m4/virt-compile-sanitizer.m4 diff --git a/configure.ac b/configure.ac index 246f4e0..4334614 100644 --- a/configure.ac +++ b/configure.ac @@ -237,6 +237,7 @@ LIBVIRT_COMPILE_WARNINGS LIBVIRT_COMPILE_PIE LIBVIRT_LINKER_RELRO LIBVIRT_LINKER_NO_INDIRECT +LIBVIRT_COMPILE_SANITIZER LIBVIRT_ARG_APPARMOR LIBVIRT_ARG_ATTR @@ -1011,6 +1012,7 @@ AC_MSG_NOTICE([]) AC_MSG_NOTICE([Miscellaneous]) AC_MSG_NOTICE([]) LIBVIRT_RESULT_DEBUG +LIBVIRT_RESULT_SANITIZER AC_MSG_NOTICE([ Use -Werror: $enable_werror]) AC_MSG_NOTICE([ Warning Flags: $WARN_CFLAGS]) LIBVIRT_RESULT_DTRACE diff --git a/m4/virt-compile-sanitizer.m4 b/m4/virt-compile-sanitizer.m4 new file mode 100644 index 0000000..a7cac31 --- /dev/null +++ b/m4/virt-compile-sanitizer.m4 @@ -0,0 +1,51 @@ +dnl +dnl Check for support for Sanitizers +dnl Check for -fsanitize=address and -fsanitize=undefined support +dnl +dnl This library is free software; you can redistribute it and/or +dnl modify it under the terms of the GNU Lesser General Public +dnl License as published by the Free Software Foundation; either +dnl version 2.1 of the License, or (at your option) any later version. +dnl +dnl This library is distributed in the hope that it will be useful, +dnl but WITHOUT ANY WARRANTY; without even the implied warranty of +dnl MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +dnl Lesser General Public License for more details. +dnl +dnl You should have received a copy of the GNU Lesser General Public +dnl License along with this library. If not, see +dnl <http://www.gnu.org/licenses/>. +dnl + +AC_DEFUN([LIBVIRT_COMPILE_SANITIZER],[ + LIBVIRT_ARG_ENABLE([ASAN], [Build with address sanitizer support], [no]) + LIBVIRT_ARG_ENABLE([UBSAN], [Build with undefined behavior sanitizer support], [no]) + + SAN_CFLAGS= + SAN_LDFLAGS= + + AS_IF([test "x$enable_asan" = "xyes"], [ + gl_COMPILER_OPTION_IF([-fsanitize=address -fno-omit-frame-pointer], [ + SAN_CFLAGS="-fsanitize=address -fno-omit-frame-pointer" + SAN_LDFLAGS="-fsanitize=address" + ]) + + AC_SUBST([SAN_CFLAGS]) + AC_SUBST([SAN_LDFLAGS]) + ]) + + AS_IF([test "x$enable_ubsan" = "xyes"], [ + gl_COMPILER_OPTION_IF([-fsanitize=undefined -fno-omit-frame-pointer], [ + SAN_CFLAGS="$SAN_CFLAGS -fsanitize=undefined -fno-omit-frame-pointer" + SAN_LDFLAGS="$SAN_LDFLAGS -fsanitize=undefined" + ]) + + AC_SUBST([SAN_CFLAGS]) + AC_SUBST([SAN_LDFLAGS]) + ]) +]) + +AC_DEFUN([LIBVIRT_RESULT_SANITIZER], [ + AC_MSG_NOTICE([ ASan: $enable_asan]) + AC_MSG_NOTICE([ UBSan: $enable_ubsan]) +]) -- 2.11.0

3 2

[libvirt] [PATCH 0/9] virObject adjustments for common object
by John Ferlan 30 May '17

30 May '17

Originally posted in part as an RFC: https://www.redhat.com/archives/libvir-list/2017-April/msg00321.html Obviously not a 3.4.0 discussion, but I'm trying to get ahead of the curve for 3.5.0 and posting this now as I'm "close enough" to at least convert secrets, nwfilters, nodedevs, and interfaces to use a more common object methodolgy to add/store, fetch/find, search, list, etc. the objects in something other than a linear forward linked list. Converting the storage pool, storage volume, and network is taking longer than hoped since there are so many patches to get from pointA to pointB. So I'll focus on the 4 and work through getting the complete model published/adopted as those have been in my branches for a while now. Still these patches get as far as the creation of the common object which is a start. The first 4 patches are essentially setup. Patch 5 is new and covers the need for nwfilters (at least temporarily until self locking list mgmt is possible). Patches 6-7 were more or less what patch 7-8 were in the RFC. Patches 8-9 are essentially the driver/common object extension. Patches 5-6 from the RFC still exist in my branches and would be the next logical step. It was a conscious decision to just use the "default" string comparison for hash table key. That key could be a UUID, but since the UUID can easily be converted from unsigned to signed - I believe it would be far simpler to keep that mechanism rather than complexity introduced by having "different" key search algorithms when the key isn't a char *. John Ferlan (9): util: Formatting cleanups to virobject API util: Introduce virObjectGetLockableObj util: Generate a common internal only error print util: Add magic_marker for all virObjects util: Introduce virObjectLockableRecursiveNew util: Introduce virObjectPoolableHashElement util: Add virObjectPoolableHashElementGet*Key util: Introduce virObjectPoolableDef util: Add virObjectPoolableDef* accessor API's src/libvirt_private.syms | 12 ++ src/util/virobject.c | 492 +++++++++++++++++++++++++++++++++++++++++++---- src/util/virobject.h | 130 +++++++++++-- 3 files changed, 580 insertions(+), 54 deletions(-) -- 2.9.4

3 17

[libvirt] [PATCH v3 0/3] Loadparm support
by Farhan Ali 30 May '17

30 May '17

This patch series introduces the support for new s390x 'loadparm' feature. The 'loadparm' can be used to select the boot entry to boot from, for a boot device. Here is a link to the QEMU patches: https://lists.nongnu.org/archive/html/qemu-devel/2017-05/msg00192.html ChangeLog --------- v2 -> v3: - Updated news.xml and formatdomain.html.in with a more architectural description of loadparm (patch 1) v1 -> v2: - Rebased the patch series on the latest master, commit 2f69dd3 virfiletest: include linux/falloc.h Thanks Farhan Ali Farhan Ali (3): conf : Add loadparm boot option for a boot device qemu : Add loadparm to qemu command line string tests : Testcases for loadparm docs/formatdomain.html.in | 9 ++- docs/news.xml | 11 ++++ docs/schemas/domaincommon.rng | 7 +++ src/conf/device_conf.h | 1 + src/conf/domain_conf.c | 69 +++++++++++++++++++++- src/qemu/qemu_capabilities.c | 2 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 37 ++++++++++++ ...-machine-loadparm-multiple-disks-nets-s390.args | 28 +++++++++ ...v-machine-loadparm-multiple-disks-nets-s390.xml | 43 ++++++++++++++ .../qemuxml2argv-machine-loadparm-net-s390.args | 20 +++++++ .../qemuxml2argv-machine-loadparm-net-s390.xml | 26 ++++++++ ...xml2argv-machine-loadparm-s390-char-invalid.xml | 26 ++++++++ ...uxml2argv-machine-loadparm-s390-len-invalid.xml | 26 ++++++++ .../qemuxml2argv-machine-loadparm-s390.args | 20 +++++++ .../qemuxml2argv-machine-loadparm-s390.xml | 26 ++++++++ tests/qemuxml2argvtest.c | 19 ++++++ 17 files changed, 367 insertions(+), 4 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-multiple-disks-nets-s390.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-multiple-disks-nets-s390.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-net-s390.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-net-s390.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-s390-char-invalid.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-s390-len-invalid.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-s390.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-machine-loadparm-s390.xml -- 1.9.1

2 10

[libvirt] [PATCH] qemu: Don't error out if allocation info can't be queried
by Peter Krempa 30 May '17

30 May '17

qemuDomainGetBlockInfo would error out if qemu did not report 'wr_highest_offset'. This usually does not happen, but can happen briefly during active layer block commit. There's no need to report the error, we can simply report that the disk is fully alocated at that point. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1452045 --- src/qemu/qemu_driver.c | 8 -------- 1 file changed, 8 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 67f54282a..f6b352b56 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -11541,14 +11541,6 @@ qemuDomainGetBlockInfo(virDomainPtr dom, } if (!entry->wr_highest_offset_valid) { - if (virStorageSourceGetActualType(disk->src) == VIR_STORAGE_TYPE_BLOCK && - disk->src->format != VIR_STORAGE_FILE_RAW) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("failed to query the maximum written offset of " - "block device '%s'"), disk->dst); - goto endjob; - } - info->allocation = entry->physical; } else { if (virStorageSourceGetActualType(disk->src) == VIR_STORAGE_TYPE_FILE && -- 2.12.2

2 1

[libvirt] [libvirt-sandbox PATCH] Drop library/ from template name and image path
by Guido Günther 30 May '17

30 May '17

If one pastes from the output of virt-sansbox-image $ virt-sandbox-image list docker:/library/ubuntu?tag=17.04 docker:/library/debian?tag=latest verbatim $ virt-sandbox-image run -c qemu:///session docker:/library/debian?tag=latest This fails like /home/<usr>/.local/share/libvirt/images/library/debian:qbeilwxard.qcow2: Could not create file: No such file or directory Unable to start sandbox: Failed to create domain: XML error: name library/debian:qbeilwxard cannot contain '/' /usr/bin/virt-sandbox-image: [Errno 2] No such file or directory: '/home/<user>/.local/share/libvirt/images/library/debian:qbeilwxard.qcow2' so strip of any leading components. --- This might not be the correct fix but I hope you get the idea. libvirt-sandbox/image/cli.py | 2 +- libvirt-sandbox/image/sources/docker.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/libvirt-sandbox/image/cli.py b/libvirt-sandbox/image/cli.py index 3a5ccfa..0f27a16 100644 --- a/libvirt-sandbox/image/cli.py +++ b/libvirt-sandbox/image/cli.py @@ -96,7 +96,7 @@ def run(args): name = args.name if name is None: randomid = ''.join(random.choice(string.lowercase) for i in range(10)) - name = tmpl.path[1:] + ":" + randomid + name = tmpl.path[1:].split('/')[-1] + ":" + randomid diskfile = source.get_disk(template=tmpl, templatedir=template_dir, diff --git a/libvirt-sandbox/image/sources/docker.py b/libvirt-sandbox/image/sources/docker.py index aa5675e..6ca086c 100755 --- a/libvirt-sandbox/image/sources/docker.py +++ b/libvirt-sandbox/image/sources/docker.py @@ -655,7 +655,7 @@ class DockerSource(base.Source): def get_disk(self, template, templatedir, imagedir, sandboxname): image = DockerImage.from_template(template) configfile, diskfile = self._get_template_data(image, templatedir) - tempfile = imagedir + "/" + sandboxname + ".qcow2" + tempfile = imagedir + "/" + sandboxname.split('/')[-1] + ".qcow2" if not os.path.exists(imagedir): os.makedirs(imagedir) cmd = ["qemu-img","create","-q","-f","qcow2"] -- 2.11.0

2 1

[libvirt] [PATCH v2] nodedev: Increase the netlink socket buffer size to the one used by udev
by Erik Skultety 30 May '17

30 May '17

From: "ning.bo" <ning.bo9(a)zte.com.cn> When a number of SRIOV VFs (up to 128 on Intel XL710) is created: for i in `seq 0 1`; do echo 63 > /sys/class/net/<interface>/device/sriov_numvfs done libvirtd will then report "udev_monitor_receive_device returned NULL" error because the netlink socket buffer is not big enough (using GDB on libudev confirmed this with ENOBUFFS) and thus some udev events were dropped. This results in some devices being missing in the nodedev-list output. This patch overrides the system's rmem_max limit but for that, we need to make sure we've got root privileges. https://bugzilla.redhat.com/show_bug.cgi?id=1450960 Signed-off-by: ning.bo <ning.bo9(a)zte.com.cn> Signed-off-by: Erik Skultety <eskultet(a)redhat.com> --- Additionally, we might want to check for the errno, providing a specific message if such issue occurs again in a further non-specified point in time and return the generic, yet cryptic one for all other cases. src/node_device/node_device_udev.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/src/node_device/node_device_udev.c b/src/node_device/node_device_udev.c index 4ecb0b18f..0b3717397 100644 --- a/src/node_device/node_device_udev.c +++ b/src/node_device/node_device_udev.c @@ -1728,6 +1728,13 @@ static int nodeStateInitialize(bool privileged, udev_monitor_enable_receiving(priv->udev_monitor); + /* mimic udevd's behaviour and override the systems rmem_max limit in case + * there's a significant number of device 'add' events + */ + if (geteuid() == 0) + udev_monitor_set_receive_buffer_size(priv->udev_monitor, + 128 * 1024 * 1024); + /* We register the monitor with the event callback so we are * notified by udev of device changes before we enumerate existing * devices because libvirt will simply recreate the device if we -- 2.13.0

3 4

[libvirt] [PATCH] qemu:json: Fix daemon crash on handling domain shutdown event
by Erik Skultety 30 May '17

30 May '17

commit a8eba5036 added further checking of the guest shutdown cause, but this enhancement is available since qemu 2.10, causing a crash because of a NULL pointer dereference on older qemus. Thread 1 "libvirtd" received signal SIGSEGV, Segmentation fault. 0x00007ffff72441af in virJSONValueObjectGet (object=0x0, key=0x7fffd5ef11bf "guest") at util/virjson.c:769 769 if (object->type != VIR_JSON_TYPE_OBJECT) (gdb) bt 0 in virJSONValueObjectGet 1 in virJSONValueObjectGetBoolean 2 in qemuMonitorJSONHandleShutdown 3 in qemuMonitorJSONIOProcessEvent 4 in qemuMonitorJSONIOProcessLine 5 in qemuMonitorJSONIOProcess 6 in qemuMonitorIOProcess 7 in qemuMonitorIO 8 in virEventPollDispatchHandles 9 in virEventPollRunOnce 10 in virEventRunDefaultImpl 11 in virNetDaemonRun 12 in main Signed-off-by: Erik Skultety <eskultet(a)redhat.com> --- src/qemu/qemu_monitor_json.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 757595dd7..f208dd05a 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -528,7 +528,7 @@ static void qemuMonitorJSONHandleShutdown(qemuMonitorPtr mon, virJSONValuePtr da bool guest = false; virTristateBool guest_initiated = VIR_TRISTATE_BOOL_ABSENT; - if (virJSONValueObjectGetBoolean(data, "guest", &guest) == 0) + if (data && virJSONValueObjectGetBoolean(data, "guest", &guest) == 0) guest_initiated = guest ? VIR_TRISTATE_BOOL_YES : VIR_TRISTATE_BOOL_NO; qemuMonitorEmitShutdown(mon, guest_initiated); -- 2.13.0

2 2

[libvirt] [PATCH v3] Regression: Report correct CPUs present on executing virsh nodecpumap
by Nitesh Konkar 30 May '17

30 May '17

On executing the virsh nodecpumap command, the number of CPUs present was shown as (last cpu online id + 1). This patch fixes the issue by reporting the current number of CPUs present. Signed-off-by: Nitesh Konkar <nitkon12(a)linux.vnet.ibm.com> --- src/util/virhostcpu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/util/virhostcpu.c b/src/util/virhostcpu.c index aa9cfeac2..c485a9721 100644 --- a/src/util/virhostcpu.c +++ b/src/util/virhostcpu.c @@ -1093,7 +1093,7 @@ virHostCPUGetMap(unsigned char **cpumap, if (online) *online = virBitmapCountBits(cpus); - ret = virBitmapSize(cpus); + ret = virHostCPUGetCount(); cleanup: if (ret < 0 && cpumap) -- 2.11.0

2 1

[libvirt] [libvirt-sandbox PATCH] mkinitrd: Add missing fscrypto module
by Guido Günther 30 May '17

30 May '17

--- libvirt-sandbox/libvirt-sandbox-builder-machine.c | 1 + 1 file changed, 1 insertion(+) diff --git a/libvirt-sandbox/libvirt-sandbox-builder-machine.c b/libvirt-sandbox/libvirt-sandbox-builder-machine.c index bdec490..7204f71 100644 --- a/libvirt-sandbox/libvirt-sandbox-builder-machine.c +++ b/libvirt-sandbox/libvirt-sandbox-builder-machine.c @@ -186,6 +186,7 @@ static gchar *gvir_sandbox_builder_machine_mkinitrd(GVirSandboxConfig *config, /* In case ext4 is built as a module, include it and its deps * for the root mount */ + gvir_sandbox_config_initrd_add_module(initrd, "fscrypto.ko"); gvir_sandbox_config_initrd_add_module(initrd, "mbcache.ko"); gvir_sandbox_config_initrd_add_module(initrd, "jbd2.ko"); gvir_sandbox_config_initrd_add_module(initrd, "crc16.ko"); -- 2.11.0

2 2