- 09 8月, 2019 26 次提交
-
-
由 Jiri Denemark 提交于
Since qemuDomainDefPostParse callback requires qemuCaps, we need to make sure it gets the capabilities stored in the domain's private data if the domain is running. Passing NULL may cause QEMU capabilities probing to be triggered in case QEMU binary changed in the meantime. When this happens while a running domain object is locked, QMP event delivered to the domain before QEMU capabilities probing finishes will deadlock the event loop. This patch fixes all paths leading to qemuMigrationCookieXMLParse. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
由 Jiri Denemark 提交于
Since qemuDomainDefPostParse callback requires qemuCaps, we need to make sure it gets the capabilities stored in the domain's private data if the domain is running. Passing NULL may cause QEMU capabilities probing to be triggered in case QEMU binary changed in the meantime. When this happens while a running domain object is locked, QMP event delivered to the domain before QEMU capabilities probing finishes will deadlock the event loop. This patch fixes all paths leading to virDomainDefParseString. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
由 Jiri Denemark 提交于
Since qemuDomainDefPostParse callback requires qemuCaps, we need to make sure it gets the capabilities stored in the domain's private data if the domain is running. Passing NULL may cause QEMU capabilities probing to be triggered in case QEMU binary changed in the meantime. When this happens while a running domain object is locked, QMP event delivered to the domain before QEMU capabilities probing finishes will deadlock the event loop. This patch fixes all paths leading to qemuMigrationAnyPrepareDef. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
由 Jiri Denemark 提交于
Since qemuDomainDefPostParse callback requires qemuCaps, we need to make sure it gets the capabilities stored in the domain's private data if the domain is running. Passing NULL may cause QEMU capabilities probing to be triggered in case QEMU binary changed in the meantime. When this happens while a running domain object is locked, QMP event delivered to the domain before QEMU capabilities probing finishes will deadlock the event loop. This patch fixes all paths leading to qemuDomainSaveImageOpen. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
由 Jiri Denemark 提交于
Since qemuDomainDefPostParse callback requires qemuCaps, we need to make sure it gets the capabilities stored in the domain's private data if the domain is running. Passing NULL may cause QEMU capabilities probing to be triggered in case QEMU binary changed in the meantime. When this happens while a running domain object is locked, QMP event delivered to the domain before QEMU capabilities probing finishes will deadlock the event loop. This patch fixes all paths leading to qemuDomainDefFormatBufInternal. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
由 Jiri Denemark 提交于
Since qemuDomainDefPostParse callback requires qemuCaps, we need to make sure it gets the capabilities stored in the domain's private data if the domain is running. Passing NULL may cause QEMU capabilities probing to be triggered in case QEMU binary changed in the meantime. When this happens while a running domain object is locked, QMP event delivered to the domain before QEMU capabilities probing finishes will deadlock the event loop. This patch fixes all paths leading to qemuDomainDefCopy. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
由 Michal Privoznik 提交于
If yajl_parse() fails, we try to print an error message. For that, yajl_get_error() is used. However, its documentation say that caller is also responsible for freeing the memory it allocates by using yajl_free_error(). Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Michal Privoznik 提交于
The NSS module has a compile time option which when enabled makes ERROR() and DEBUG() print messages onto stderr. But now that the module no longer links with libvirt, we need to include stdio.h and define NULLSTR(). Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Michal Privoznik 提交于
Due to latest rewrite of NSS module, we are doing yajl parsing ourselves. This means, we had to introduce couple of callback that yajl calls. According to its documentation, a callback can cancel parsing if it returns a zero value. Well, we do just that in the string callback (findLeasesParserString()). If the JSON file we are parsing contains a key that we are not interested in, zero is returned meaning stop all parsing. This is not correct, because the JSON file can contain some other keys which are not harmful for our address translation (e.g. 'client-id'). Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Paolo Bonzini 提交于
host-passthrough documentation menions that the source and destination hosts are not identical in both hardware and configuration. Configuration actually includes microcode version and QEMU version, but this is not clear so make it explicit Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com> Message-Id: <20190802125415.15227-1-pbonzini@redhat.com> Reviewed-by: NJiri Denemark <jdenemar@redhat.com>
-
由 Ilias Stamatis 提交于
Signed-off-by: NIlias Stamatis <stamatis.iliass@gmail.com> Reviewed-by: NErik Skultety <eskultet@redhat.com>
-
由 Michal Privoznik 提交于
This completer can be used to complete pool types. Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Now that there is no code in virsh-completer.c it doesn't make much sense to keep those #include-s around. Delete them. Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to host (e.g. they complete various checkpoint aspects) into virsh-completer-checkpoint.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to host (e.g. they complete various host aspects) into virsh-completer-host.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to snapshot (e.g. they complete various snapshot aspects) into virsh-completer-snapshot.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to secret (e.g. they complete various secret aspects) into virsh-completer-secret.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to nwfilter (e.g. they complete various nwfilter aspects) into virsh-completer-nwfilter.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to nodedev (e.g. they complete various nodedev aspects) into virsh-completer-nodedev.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to networks (e.g. they complete various network aspects) into virsh-completer-network.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to interfaces (e.g. they complete various interface aspects) into virsh-completer-interface.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to storage volumes (e.g. they complete various storage volume aspects) into virsh-completer-volume.c Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to storage pools (e.g. they complete various storage pool aspects) into virsh-completer-pool.c. Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
Mixing all completers in one file does not support maintainability. Separate those completers which relate to domains (e.g. they complete various domain aspects) into virsh-completer-domain.c. Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
In next commits the virsh-completer.c is going to be split into smaller files. Expose virshCommaStringListComplete() so that it can still be used from those new files. Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
由 Michal Privoznik 提交于
The proper name is [vir|virsh]NodeDevice* and not Nodedev. Fortunately, there are only handful of offenders. Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NJonathon Jongsma <jjongsma@redhat.com>
-
- 08 8月, 2019 7 次提交
-
-
由 Roman Bolshakov 提交于
Confusing message is printed when a parse/format sockettest fails. E.g. there's a test that parses/formats ::ffff and the format fails like that: 38) Test format ::ffff family AF_UNSPEC ... Offset 2 Expect [0.0.255.255] Actual [ffff] It should be instead: 38) Test format ::ffff family AF_UNSPEC ... Offset 2 Expect [ffff] Actual [0.0.255.255] Signed-off-by: NRoman Bolshakov <r.bolshakov@yadro.com> Reviewed-by: NAndrea Bolognani <abologna@redhat.com>
-
由 Daniel P. Berrangé 提交于
The conversion to drop gnulib in the previous patch: commit 8242ce4f Author: Daniel P. Berrangé <berrange@redhat.com> Date: Thu Aug 8 10:23:26 2019 +0100 tools: avoid accidentally using files from gnulib Missed a few conversions needed for FreeBSD. In particular netdb.h doesn't pull in sys/socket.h or netinet/in.h Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
Use the correct enum constant when validating vlan usage. This fixes a merge error in commit 6cb0ec48 Author: Daniel P. Berrangé <berrange@redhat.com> Date: Mon Sep 3 17:34:22 2018 +0100 network: convert networkAllocateActualDevice to virNetworkPortDef Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
The AM_CPPFLAGS setting includes the gnulib headers, which means we can get some replacement functions defined. Since virt-login-shell and the NSS module intentionally don't link to gnulib, these replacement functions causes link failures. This was seen cross-compiling on Debian for example: virt-login-shell.o: In function `main': /builds/libvirt/libvirt/build/tools/../../tools/virt-login-shell.c:81: undefined reference to `rpl_strerror' /builds/libvirt/libvirt/build/tools/../../tools/virt-login-shell.c:66: undefined reference to `rpl_strerror' /builds/libvirt/libvirt/build/tools/../../tools/virt-login-shell.c:75: undefined reference to `rpl_strerror' The only way to avoid these replacement gnulib headers is to drop the -Ignulib/lib flags. We do still want to use gnulib for configmake.h and intprops.h, but those can be included via their full path. We must also stop using internal.h, since that expects -Ignulib/lib to be on the include path in order to resolve the verify.h header. Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Eric Blake 提交于
Requires adjustments to use verify_expr() which replaces verify_true(), and to disable the new syntax check 'sc_prohibit_gnu_make_extensions' since we require GNU make. Signed-off-by: NEric Blake <eblake@redhat.com>
-
由 Eric Blake 提交于
This reverts commit 8a8e8656. The patch was premature, as it still fails to build on mingw at least in the environment used by https://travis-ci.org/libvirt/libvirt/jobs/569132417Signed-off-by: NEric Blake <eblake@redhat.com>
-
由 Eric Blake 提交于
Commit fed58d83 was a hack to fix a mingw build failure due to header inclusion order resulting in a clash over the use of DATADIR, repeating a trick made several other times in the past. Better is to revert that, and instead use pragmas to avoid the clash in the first place, regardless of header ordering, solving it for everyone. Signed-off-by: NEric Blake <eblake@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
- 07 8月, 2019 7 次提交
-
-
由 Daniel P. Berrangé 提交于
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
Now that the code does not refer to any libvirt headers, except internal.h macros, it does not need to link to any libvirt code, nor gnulib either. The only thing it needs is yajl. Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
Use the plain libc APIs to avoid a dependancy on the main libvirt code from the nss module. Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
Use the plain libc socket APIs to avoid a dependancy on the main libvirt code from the nss module. Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
The .leases file is currently loaded using the virLease class, which in turn uses the virJSON parsing code. This pulls in a heap of libvirt code (logging, hash tables, etc) which we do not wish to depend on. This uses the yajl parser code directly, so the only dep is yajl and plain libc functions. Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
The .macs file is currently loaded using the virMacMap class, which in turn uses the virJSON parsing code. This pulls in a heap of libvirt code (logging, hash tables, objects, etc) which we do not wish to depend on. This uses the yajl parser code directly, so the only dep is yajl and plain libc functions. Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Daniel P. Berrangé 提交于
Build a list of mac addresses immediately, so that later code searching for leases can be simplified and avoid needing to use the virMacMap object. Reviewed-by: NMichal Privoznik <mprivozn@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-