- Devel - libvirt List Archives

[libvirt] [PATCH] qemu: fix msg could be a wild pointer in qemuMonitorIOProcess()
by Shannon Zhao 06 Jun '18

06 Jun '18

From: Weilun Zhu <zhuweilun(a)huawei.com> As qemuMonitorJSONIOProcess() will unlock the qemu monitor, there is some extreme situation, eg qemu send message to monitor twice in a short time, where the local viriable 'msg' of qemuMonitorIOProcess() could be a wild point: 1. qemuMonitorSend() assign mon->msg to parameter 'msg', which is alse a local variable of its caller qemuMonitorJSONCommandWithFd(), cause eventloop to send message to monitor, then wait condition. 2. qemu send message to monitor for the first time immediately. 3. qemuMonitorIOProcess() is called, then wake up the qemuMonitorSend() thread, but the qemuMonitorSend() thread stuck for a while, which means the qemu monitor is still unlocked. 4. qemu send message to monitor for the second time, such as RTC_CHANGE event 5. qemuMonitorIOProcess() is called, the local viriable 'msg' is assigned to mon->msg. 6. qemuMonitorIOProcess() call qemuMonitorJSONIOProcess() to deal with the message 7. qemuMonitorJSONIOProcess() unlock the qemu monitor, qemuMonitorSend() thread get the lock and free the mon->msg, assign mon->msg to NULL. so the local viriable 'msg' of qemuMonitorIOProcess() is a wild pointer now. AFAIK, it is not harmful to call again virCondBroadcast() while msg is a wild pointer, but just in case, we fix it in this patch. --- src/qemu/qemu_monitor.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index 43f1d2f..464f200 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -454,7 +454,7 @@ qemuMonitorIOProcess(qemuMonitorPtr mon) #if DEBUG_IO VIR_DEBUG("Process done %d used %d", (int)mon->bufferOffset, len); #endif - if (msg && msg->finished) + if (msg && msg == mon->msg && msg->finished) virCondBroadcast(&mon->notify); return len; } -- 1.8.3.1

3 3

[libvirt] [PATCH] docs: document mandatory signoffs in governance.html
by Ján Tomko 06 Jun '18

06 Jun '18

Amend the paragraphs about no CLAs and implicit license agreements to mention mandatory Signed-off-by tags. Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- This is duplicating information, but I did not find a reasonable way to link the numbered list in hacking.html without making it look ugly. Very probably due to my lack of trying. docs/governance.html.in | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/docs/governance.html.in b/docs/governance.html.in index df5999c0f2..e9608f673c 100644 --- a/docs/governance.html.in +++ b/docs/governance.html.in @@ -141,13 +141,17 @@ than having the interest and ability to provide a contribution. The libvirt project does not require any "Contributor License Agreement" - to be signed prior to engagement with the community. + to be signed prior to engagement with the community. However for + contributing patches, providing a 'Signed-off-by' line with the + author's legal name and e-mail address to demonstrate agreement + and compliance with the <a href="https://developercertificate.org/"> + Developer Certificate of Origin</a> is required. - In making a contribution to the project, the community member is - implicitly stating that they accept the terms of the license under - which the work they are contributing to is distributed. They are + In making a non-patch contribution to the project, the community + member is implicitly stating that they accept the terms of the license + under which the work they are contributing to is distributed. They are also implicitly stating that they have the legal right to make the contribution, if doing so on behalf of a broader organization / company. Most of the project's code is distributed under the GNU -- 2.13.6

2 1

[libvirt] [PATCH] qemu: Fix double free in qemuDomainSecretAESClear
by John Ferlan 06 Jun '18

06 Jun '18

Commit id 02b031a4 added a secondary path from which the incoming @secinfo would not be free'd until the private data was freed in qemuDomainStorageSourcePrivateDispose. However, by doing this the original intention to free @*secinfo afterwards is lost and thus the pass by value of the secinfo->s.aes (or secinfo->s.plain for its method) results in not keeping the NULL setting in the various secret.{username|iv|ciphertext} fields upon return to qemuDomainSecretInfoClear and eventually will result in a double free at domain destroy: raise () abort () __libc_message () malloc_printerr () _int_free () virFree qemuDomainSecretAESClear qemuDomainSecretInfoClear qemuDomainSecretInfoFree qemuDomainStorageSourcePrivateDispose virObjectUnref virStorageSourceClear virStorageSourceFree virDomainDiskDefFree virDomainDefFree virDomainObjRemoveTransientDef qemuProcessStop qemuDomainDestroyFlags virDomainDestroy Signed-off-by: John Ferlan <jferlan(a)redhat.com> --- src/qemu/qemu_domain.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) Domains w/ secrets weren't very happy when I went to destroy them today during testing... Fortunately issue is not in 4.4.0... I modified both Plain and AES just because it's probably best to avoid something like this in the future. diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index f135117a95..1fb1ef1deb 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -945,23 +945,23 @@ qemuDomainMasterKeyCreate(virDomainObjPtr vm) static void -qemuDomainSecretPlainClear(qemuDomainSecretPlain secret) +qemuDomainSecretPlainClear(qemuDomainSecretPlainPtr secret) { - VIR_FREE(secret.username); - VIR_DISPOSE_N(secret.secret, secret.secretlen); + VIR_FREE(secret->username); + VIR_DISPOSE_N(secret->secret, secret->secretlen); } static void -qemuDomainSecretAESClear(qemuDomainSecretAES secret, +qemuDomainSecretAESClear(qemuDomainSecretAESPtr secret, bool keepAlias) { if (!keepAlias) - VIR_FREE(secret.alias); + VIR_FREE(secret->alias); - VIR_FREE(secret.username); - VIR_FREE(secret.iv); - VIR_FREE(secret.ciphertext); + VIR_FREE(secret->username); + VIR_FREE(secret->iv); + VIR_FREE(secret->ciphertext); } @@ -974,11 +974,11 @@ qemuDomainSecretInfoClear(qemuDomainSecretInfoPtr secinfo, switch ((qemuDomainSecretInfoType) secinfo->type) { case VIR_DOMAIN_SECRET_INFO_TYPE_PLAIN: - qemuDomainSecretPlainClear(secinfo->s.plain); + qemuDomainSecretPlainClear(&secinfo->s.plain); break; case VIR_DOMAIN_SECRET_INFO_TYPE_AES: - qemuDomainSecretAESClear(secinfo->s.aes, keepAlias); + qemuDomainSecretAESClear(&secinfo->s.aes, keepAlias); break; case VIR_DOMAIN_SECRET_INFO_TYPE_LAST: -- 2.14.4

2 1

[libvirt] [PATCH v2 0/2] nwfilter: fix IP address learning
by Daniel P. Berrangé 05 Jun '18

05 Jun '18

Changed in v2: - Use if() instead of switch() Daniel P. Berrangé (2): nwfilter: fix IP address learning nwfilter: directly use poll to wait for packets instead of pcap_next src/nwfilter/nwfilter_learnipaddr.c | 55 ++++++++++++++++++++++------- src/nwfilter/nwfilter_learnipaddr.h | 2 +- 2 files changed, 43 insertions(+), 14 deletions(-) -- 2.17.0

2 3

[libvirt] [perl PATCH 0/6] Update RPM spec to match latest Fedora
by Daniel P. Berrangé 05 Jun '18

05 Jun '18

Syncing changes back from Fedora to follow best practice Daniel P. Berrangé (6): Remove obsolete RPM group tag Replace search.cpan.org with metacpan.org Remove commented out cruft leftover from Module::Build conversion Remove obsolete %defattr statement Use %license for LICENSE file instead of %doc Modernize list of build deps to match actual required packages perl-Sys-Virt.spec.PL | 45 ++++++++++++++++++++++++++++--------------- 1 file changed, 30 insertions(+), 15 deletions(-) -- 2.17.0

2 12

[libvirt] [PATCH v2 0/3] adding virGetLastErrorCode/Domain to paritally replace virGetLastError
by ramyelkest 05 Jun '18

05 Jun '18

Changes from v1[1]: * removed virHasLastError() and s/virHasLastError/virGetLastErrorCode/g * replaced in missed files: virmodule.c and virnetlibsshsession.c * better split of patches [1] https://www.redhat.com/archives/libvir-list/2018-May/msg00259.html ramyelkest (3): util: cleanup: using virGetLastErrorMessage instead of err->message util: added virGetLastErrorCode/Domain all: replacing virGetLastError with virGetLastErrorCode where we can include/libvirt/virterror.h | 2 ++ src/libvirt_public.syms | 6 ++++++ src/locking/lock_driver_lockd.c | 3 +-- src/lxc/lxc_controller.c | 4 +--- src/qemu/qemu_agent.c | 3 +-- src/qemu/qemu_conf.c | 3 +-- src/qemu/qemu_domain.c | 2 +- src/qemu/qemu_driver.c | 12 ++++++------ src/qemu/qemu_hotplug.c | 2 +- src/qemu/qemu_migration.c | 4 ++-- src/qemu/qemu_monitor.c | 5 ++--- src/qemu/qemu_monitor_json.c | 2 +- src/qemu/qemu_process.c | 4 ++-- src/remote/remote_driver.c | 3 +-- src/rpc/virnetclient.c | 2 +- src/rpc/virnetlibsshsession.c | 4 +--- src/util/virerror.c | 42 +++++++++++++++++++++++++++++++++++++++++ src/util/virfilecache.c | 3 +-- src/util/virmodule.c | 3 +-- src/util/virxml.c | 4 ++-- tests/commandtest.c | 2 +- tests/testutils.c | 6 ++---- tests/virhostcputest.c | 2 +- tests/virstoragetest.c | 8 ++++---- tools/virsh-domain-monitor.c | 7 +++---- tools/virsh-domain.c | 4 +--- tools/virsh-util.c | 3 +-- tools/vsh.c | 2 +- 28 files changed, 90 insertions(+), 57 deletions(-) -- 2.7.4

3 8

[libvirt] [PATCH] tests: use a fixed NBD TLS certificate path
by Daniel P. Berrangé 05 Jun '18

05 Jun '18

The default NBD TLS certificate path varies based on prefix given to configure, causing tests to fail depending on build options. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- Pushed as a build fix for CI tests/qemuxml2argvdata/disk-drive-network-tlsx509.args | 4 ++-- tests/qemuxml2argvtest.c | 4 ++++ 2 files changed, 6 insertions(+), 2 deletions(-) diff --git a/tests/qemuxml2argvdata/disk-drive-network-tlsx509.args b/tests/qemuxml2argvdata/disk-drive-network-tlsx509.args index 970b8a32a6..e25f45742c 100644 --- a/tests/qemuxml2argvdata/disk-drive-network-tlsx509.args +++ b/tests/qemuxml2argvdata/disk-drive-network-tlsx509.args @@ -44,8 +44,8 @@ file.server.host=192.168.0.3,file.server.port=9999,format=raw,if=none,\ id=drive-virtio-disk2,serial=eb90327c-8302-4725-9e1b-4e85ed4dc252,cache=none \ -device virtio-blk-pci,bus=pci.0,addr=0x6,drive=drive-virtio-disk2,\ id=virtio-disk2 \ --object tls-creds-x509,id=objvirtio-disk3_tls0,dir=/etc/pki/qemu,\ -endpoint=client,verify-peer=yes \ +-object tls-creds-x509,id=objvirtio-disk3_tls0,dir=/etc/pki/libvirt-nbd/dummy,,\ +path,endpoint=client,verify-peer=yes \ -drive file.driver=nbd,file.server.type=inet,file.server.host=example.com,\ file.server.port=1234,file.tls-creds=objvirtio-disk3_tls0,format=raw,if=none,\ id=drive-virtio-disk3,cache=none \ diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 36bff26d33..8e5504f845 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -655,6 +655,9 @@ mymain(void) VIR_FREE(driver.config->vxhsTLSx509certdir); if (VIR_STRDUP_QUIET(driver.config->vxhsTLSx509certdir, "/etc/pki/libvirt-vxhs/dummy,path") < 0) return EXIT_FAILURE; + VIR_FREE(driver.config->nbdTLSx509certdir); + if (VIR_STRDUP_QUIET(driver.config->nbdTLSx509certdir, "/etc/pki/libvirt-nbd/dummy,path") < 0) + return EXIT_FAILURE; VIR_FREE(driver.config->hugetlbfs); if (VIR_ALLOC_N(driver.config->hugetlbfs, 2) < 0) @@ -2860,6 +2863,7 @@ mymain(void) if (getenv("LIBVIRT_SKIP_CLEANUP") == NULL) virFileDeleteTree(fakerootdir); + VIR_FREE(driver.config->nbdTLSx509certdir); qemuTestDriverFree(&driver); VIR_FREE(fakerootdir); VIR_FREE(capslatest_x86_64); -- 2.17.0

1 0

[libvirt] libvirt default machine-type guarantees? (was Re: [Qemu-devel] [PATCH RFC] hw/pc: set q35 as the default x86 machine)
by Eduardo Habkost 05 Jun '18

05 Jun '18

(CCing libvir-list) On Tue, Jun 05, 2018 at 09:43:00AM +0100, Daniel P. Berrangé wrote: > On Tue, Jun 05, 2018 at 09:27:46AM +0200, Gerd Hoffmann wrote: > > Hi, > > > > > > Add to that shortcuts like -cdrom > > > > stop working, > > > > > > Maybe is fixable. > > > > Already fixed for ages. > > > > > I see marking Q35 as the default machine a first step. > > > > Maybe the better option is to go the arm route: Just don't define a > > default, so users have to specify pc or q35. That will make them notice > > there is a world beside 'pc', and we also avoid breaking things > > silently. > > If QEMU removes the default, then libvirt will have to hardcode > 'pc' as the default to maintain back compatibility, so I don't > think that ends up as a net win Is there an actual promise to never change the default machine-type documented in the libvirt API, or is this just fear of breaking existing code? -- Eduardo

4 10

[libvirt] [PATCH v3 0/4] qemu: use FD passing for chardev UNIX sockets
by Daniel P. Berrangé 05 Jun '18

05 Jun '18

This series makes use of the chardev fd passing arriving in QEMU 2.12 to get rid of the startup race wrt opening the QEMU monitor. It is actually enabled in all chardev UNIX sockets for sake of having the same codepath everywhere, but is only important for the monitor socket. Changed in v3: - Refactor UNIX socket opening code to allow it to be mocked in the unit tests to avoid creating real UNIX sockets Daniel P. Berrangé (4): qemu: probe for -chardev 'fd' parameter for FD passing qemu: support passing pre-opened UNIX socket listen FD qemu: don't retry connect() if doing FD passing qemu: remove pointless connect retry logic in agent src/qemu/qemu_agent.c | 84 ++----------------- src/qemu/qemu_capabilities.c | 4 +- src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 64 +++++++++++++- src/qemu/qemu_command.h | 4 + src/qemu/qemu_monitor.c | 54 +++++++----- src/qemu/qemu_monitor.h | 1 + src/qemu/qemu_process.c | 27 ++++-- .../caps_2.12.0.aarch64.xml | 1 + .../caps_2.12.0.ppc64.xml | 1 + .../caps_2.12.0.s390x.xml | 1 + .../caps_2.12.0.x86_64.xml | 1 + tests/qemumonitortestutils.c | 1 + .../disk-drive-write-cache.x86_64-latest.args | 3 +- ...irtio-scsi-reservations.x86_64-latest.args | 3 +- tests/qemuxml2argvmock.c | 16 ++++ 16 files changed, 153 insertions(+), 113 deletions(-) -- 2.17.0

3 7

[libvirt] [PATCH 0/3] travis: Stop performing Ubuntu builds
by Andrea Bolognani 05 Jun '18

05 Jun '18

They're all failing at the moment anyway. We'll get the Ubuntu coverage back as soon as we add Ubuntu builders to the CentOS CI environment. Andrea Bolognani (3): travis: Stop performing Ubuntu builds travis: Remove matrix travis: Minor style tweaks .travis.yml | 118 ++++++++-------------------------------------------- 1 file changed, 17 insertions(+), 101 deletions(-) -- 2.17.1

2 4