February 2021 - Devel - libvirt List Archives

[libvirt PATCH] Revert "tests: Avoid gnulib replacements in mocks"
by Andrea Bolognani 05 Feb '21

05 Feb '21

Now that we're no longer using gnulib, we can treat macOS the same as all other targets. This reverts commit 0ae6f5cea54d95c0d1dedf04a0a2accfe2529fb2 Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- Roman, this seems to work fine both locally and on Cirrus CI[1], but I'd like to have your thumbs up before pushing. Thanks! [1] https://gitlab.com/abologna/libvirt/-/jobs/1007362310 tests/virfilewrapper.c | 5 ----- tests/virmockstathelpers.c | 10 ---------- 2 files changed, 15 deletions(-) diff --git a/tests/virfilewrapper.c b/tests/virfilewrapper.c index ca2356b5c9..7034e22420 100644 --- a/tests/virfilewrapper.c +++ b/tests/virfilewrapper.c @@ -163,12 +163,7 @@ int access(const char *path, int mode) return real_access(newpath ? newpath : path, mode); } -# ifdef __APPLE__ -int _open(const char *path, int flags, ...) __asm("_open"); -int _open(const char *path, int flags, ...) -# else int open(const char *path, int flags, ...) -# endif { g_autofree char *newpath = NULL; va_list ap; diff --git a/tests/virmockstathelpers.c b/tests/virmockstathelpers.c index 3bd2437ffe..49485cd360 100644 --- a/tests/virmockstathelpers.c +++ b/tests/virmockstathelpers.c @@ -204,12 +204,7 @@ static int virMockStatRedirect(const char *path, char **newpath); #endif #ifdef MOCK_STAT -# ifdef __APPLE__ -int _stat(const char *path, struct stat *sb) __asm("_stat$INODE64"); -int _stat(const char *path, struct stat *sb) -# else int stat(const char *path, struct stat *sb) -# endif { g_autofree char *newpath = NULL; @@ -279,13 +274,8 @@ __xstat64(int ver, const char *path, struct stat64 *sb) #endif #ifdef MOCK_LSTAT -# ifdef __APPLE__ -int _lstat(const char *path, struct stat *sb) __asm("_lstat$INODE64"); -int _lstat(const char *path, struct stat *sb) -# else int lstat(const char *path, struct stat *sb) -# endif { g_autofree char *newpath = NULL; -- 2.26.2

2 1

[PATCH] util: Remove '\n' from vhostuser ifname
by Yalei Li 05 Feb '21

05 Feb '21

When deleting the vhostuserclient interface, OVS prompts that the interface does not exist, Through the XML file, I found that the "target dev" has a '\n', results in an XML parsing error. XML file: <target dev='vm-20ac9c030a47 '/> That is because 'ovs-vsctl' returns a newline result, always come with a '\n', and the vircommandrun function puts it in ifname. So virNetDevOpenvswitchGetVhostuserIfname should remove '\n' from ifname. 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 1

[PATCH] docs: Remove broken link to Xen channel doc
by Jim Fehlig 05 Feb '21

05 Feb '21

Many of Xen's text documents have been converted to man pages over the years, the channel doc being one of them. Replace the broken channel.txt link with the name of the man page providing the same information. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- docs/formatdomain.rst | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 0f78aa3a8c..aa7ea61258 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -6353,11 +6353,10 @@ types have different ``target`` attributes. ``xen`` Paravirtualized Xen channel. Channel is exposed in the guest as a Xen console but identified with a name. Setup and consumption of a Xen channel depends on - software and configuration in the guest (for more info, please see - http://xenbits.xen.org/docs/unstable/misc/channel.txt) Channel source path - semantics are the same as the virtio target type. The ``state`` attribute is - not supported since Xen channels lack the necessary probing mechanism. - :since:`Since 2.3.0` + software and configuration in the guest. See the xen-pv-channel(7) man page + for more information. Channel source path semantics are the same as the virtio + target type. The ``state`` attribute is not supported since Xen channels lack + the necessary probing mechanism. :since:`Since 2.3.0` ``spicevmc`` Paravirtualized SPICE channel. The domain must also have a SPICE server as a `graphics device <#elementsGraphics>`__, at which point the host piggy-backs -- 2.29.2

2 1

[libvirt PATCH 00/15] eliminate VIR_FREE in all *Dispose() functions
by Laine Stump 04 Feb '21

04 Feb '21

A *Dispose() function is similar to a *Free() function, except that 1) the object is always sent as a void* so it has to be typecast into some other object-specific pointer at the top of the function, and 2) it frees all the resources inside the object, but never frees the object itself (this is done by the caller, somewhere deep in the bowels of virObjectDispose or something I guess; frankly I've always ignored the details simply "because I could"). The important point is that the contents of the object are never referenced in any way after return from the Dispose function, so it is unnecessary to clear any pointers, ergo (always wanted to use that word!) it's completely safe to replace all VIR_FREEs in a *Dispose() function with g_free (as long as there's nothing within the Dispose function itself that depends on the pointers being cleared). After this series is applied, there will be exactly 0 instances of VIR_FREE in any *Dispose() (or *Free()) function. As with the *Free() series, almost all were accomplished by directly replacing VIR_FREE with g_free, but there were a couple oddities that needed separate patches just to call them out: * Patch 1 & 2 - in both cases a Dispose() function was the only caller to a *Free() function that didn't fit the normal pattern of a *Free() function. Since each of the Free()s had only one caller (their respective *Dispose() parents), their contents were just moved into the caller, clearing the way for their VIR_FREEs to be g_free-ified (in later patches, along with the other *Dispose() functions in the same directories). 220 VIR_FREE uses eliminated in this series, so a total of 762 for the two series combined (nearly 20% of all remaining VIR_FREEs). Laine Stump (15): conf: simplify virDomainCapsDispose() rpc: eliminate static function virNetLibsshSessionAuthMethodsFree() bhyve: replace VIR_FREE with g_free in all *Dispose() functions libxl: replace VIR_FREE with g_free in all *Dispose() functions qemu: replace VIR_FREE with g_free in all *Dispose() functions interface: replace VIR_FREE with g_free in all *Dispose() functions access: replace VIR_FREE with g_free in all *Dispose() functions hypervisor: replace VIR_FREE with g_free in all *Dispose() functions logging: replace VIR_FREE with g_free in all *Dispose() functions rpc: replace VIR_FREE with g_free in all *Dispose() functions security: replace VIR_FREE with g_free in all *Dispose() functions util: replace VIR_FREE with g_free in all *Dispose() functions conf: replace VIR_FREE with g_free in all *Dispose() functions tests: replace VIR_FREE with g_free in all *Dispose() functions datatypes: replace VIR_FREE with g_free in all *Dispose() functions src/access/viraccessmanager.c | 2 +- src/bhyve/bhyve_conf.c | 2 +- src/conf/capabilities.c | 22 ++--- src/conf/checkpoint_conf.c | 2 +- src/conf/domain_capabilities.c | 29 ++---- src/conf/domain_conf.c | 2 +- src/conf/domain_event.c | 52 +++++----- src/conf/moment_conf.c | 6 +- src/conf/object_event.c | 4 +- src/conf/snapshot_conf.c | 4 +- src/conf/virsecretobj.c | 6 +- src/conf/virstorageobj.c | 4 +- src/datatypes.c | 34 +++---- src/hypervisor/virhostdev.c | 2 +- src/interface/interface_backend_netcf.c | 2 +- src/libxl/libxl_conf.c | 20 ++-- src/libxl/libxl_migration.c | 2 +- src/logging/log_handler.c | 2 +- src/qemu/qemu_agent.c | 2 +- src/qemu/qemu_capabilities.c | 10 +- src/qemu/qemu_conf.c | 122 ++++++++++++------------ src/qemu/qemu_domain.c | 10 +- src/qemu/qemu_monitor.c | 4 +- src/rpc/virnetclient.c | 4 +- src/rpc/virnetdaemon.c | 6 +- src/rpc/virnetlibsshsession.c | 37 +++---- src/rpc/virnetsaslcontext.c | 2 +- src/rpc/virnetserver.c | 8 +- src/rpc/virnetserverservice.c | 2 +- src/rpc/virnetsocket.c | 6 +- src/rpc/virnetsshsession.c | 8 +- src/rpc/virnettlscontext.c | 6 +- src/security/security_manager.c | 2 +- src/util/virdnsmasq.c | 2 +- src/util/virfilecache.c | 4 +- src/util/virmdev.c | 2 +- src/util/virnvme.c | 2 +- src/util/virpci.c | 2 +- src/util/virresctrl.c | 40 ++++---- src/util/virscsi.c | 2 +- src/util/virscsivhost.c | 2 +- src/util/virusb.c | 2 +- tests/virfilecachetest.c | 2 +- 43 files changed, 235 insertions(+), 251 deletions(-) -- 2.29.2

2 17

Re: [PATCH 3/3] utils: Deprecate inexact fractional suffix sizes
by Eric Blake 04 Feb '21

04 Feb '21

On 2/4/21 1:07 PM, Eric Blake wrote: > The value '1.1k' is inexact; 1126.4 bytes is not possible, so we > happen to truncate it to 1126. Our use of fractional sizes is > intended for convenience, but when a user specifies a fraction that is > not a clean translation to binary, truncating/rounding behind their > backs can cause confusion. Better is to deprecate inexact values, > which still leaves '1.5k' as valid, but alerts the user to spell out > their values as a precise byte number in cases where they are > currently being rounded. And I promptly forgot to save my changes in my editor. Consider this squashed in: diff --git i/docs/system/deprecated.rst w/docs/system/deprecated.rst index c404c3926e6f..8e5e05a10642 100644 --- i/docs/system/deprecated.rst +++ w/docs/system/deprecated.rst @@ -154,6 +154,15 @@ Input parameters that take a size value should only use a size suffix the value is hexadecimal. That is, '0x20M' is deprecated, and should be written either as '32M' or as '0x2000000'. +inexact sizes via scaled fractions (since 6.0) +'''''''''''''''''''''''''''''''''''''''''''''' + +Input parameters that take a size value should only use a fractional +size (such as '1.5M') that will result in an exact byte value. The +use of inexact values (such as '1.1M') that require truncation or +rounding is deprecated, and you should instead consider writing your +unusual size in bytes (here, '1153433' or '1153434' as desired). + QEMU Machine Protocol (QMP) commands ------------------------------------ -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org

1 0

[libvirt PATCH v2 00/24] eliminate VIR_FREE in all *Free() functions
by Laine Stump 04 Feb '21

04 Feb '21

When I sent a patch to convert (what I thought was) all of the VIR_FREEs in any *Free() function in the conf directory to g_free (with the justification that it was trivial to verify that the pointers weren't being referenced after they had been freed), there were two responses that stuck in my mind: 1) Daniel said that this was a reasonable class of uses to change and a reasonable way to arrive at the desired end result. 2) Peter justifiably expressing concern over the amount of churn in the code, and also said that he'd rather not have things "halfway" for an indeterminate time. The combination of those two comments (ignoring the part about "concern for churn" :-) let me to sit down last night and make this patch series that eliminates all uses of VIR_FREE from any function whose name ends in "Free" (or eliminate/rename the function, just for completeness' sake) In the vast majority of cases, this was done by replacing the VIR_FREEs in the functions with g_free (while verifying that everything being g_freed was actually something pointed to by the parent object that was sent to the *Free() function, and that the parent object itself was subsequently freed). But there were a a couple of cases where this wasn't quite the right thing to do: * in patch 20, two functions ending with Free() don't actually free the pointer they're sent. They instead behave like a *Clear() function, VIR_FREEing their components. Since we can see that these functions aren't actually be used as *Clear() functions (and never will be used that way), but don't want to promote confusion by leaving them misnamed, the patch renames them to *FreeContent() (more accurate name) and changes their VIR_FREEs to g_frees (it takes more than just looking at one function, but it's very easy to verify that the pointers are never referenced after they're freed). * In Patch 21-23, several *Free() functions were actually being passed a full copy of an object, and then VIR_FREEing all the pointers in the copied object. So the name was misleading, the function was inefficiently copying entire objects, and was unnecessarily NULLing the pointers that were then tossed out when the function returned. So I renamed the functions (21), changed the calling sequences to use pointers instead of copied objects (22), and (of course) changed the VIR_FREEs into g_free (23). * in Patch 24, a *Free() function actually appears to be used as a Clear() function. So I left the VIR_FREEs in place, but renamed the function to *Clear(). However, its one use as a *Clear() is when being called from a "Reset" function that (as far as I can see) is never used. If we could get rid of that Reset function, I could just fold the *Clear() function into its one remaining caller, and eliminate the VIR_FREEs. If someone knows more about the function virNetSSHSessionAuthReset() (which is in a section of the file labeled "Public API") please enlighten me. ** The last time I ran wc over the diffs of these patches, it showed they are eliminating 542 uses of VIR_FREE, which is > 13% of the 4k remaining. There is a companion series that eliminates VIR_FREE from all *Dispose() functions, but I'm sending it separately since the two series are completely independent, and I didn't want to scare anyone off. (Really, most of these can be very mechanically reviewed - just verify that all the items being g_freed are subordinates of the main argument to the function, and that that object is freed before return) Laine Stump (24): conf: replace remaining straggler VIR_FREE with g_free in vir*Free() util: replace VIR_FREE with g_free in all vir*Free() functions bhyve: replace VIR_FREE with g_free in all vir*Free() functions libxl: replace VIR_FREE with g_free in all vir*Free() functions qemu: replace VIR_FREE with g_free in all vir*Free() functions test_driver: replace VIR_FREE with g_free in all vir*Free() functions vbox: replace VIR_FREE with g_free in all vir*Free() functions vmx: replace VIR_FREE with g_free in all vir*Free() functions vz: replace VIR_FREE with g_free in all vir*Free() functions admin: replace VIR_FREE with g_free in all vir*Free() functions locking: replace VIR_FREE with g_free in all vir*Free() functions logging: replace VIR_FREE with g_free in all vir*Free() functions remote: replace VIR_FREE with g_free in all vir*Free() functions rpc: replace VIR_FREE with g_free in all vir*Free() functions security: replace VIR_FREE with g_free in all vir*Free() functions tools: replace VIR_FREE with g_free in all vir*Free() functions tests: replace VIR_FREE with g_free in all vir*Free() functions storage: replace VIR_FREE with g_free in all vir*Free() functions libvirtd: replace straggler VIR_FREE with g_free in all vir*Free() functions util: rename two *Free() functions while changing VIR_FREE to g_free qemu: rename virFirmware*Free() functions to have more accurate names qemu: pass pointers instead of copying objects for qemuFirmware*FreeContent() qemu: replace VIR_FREE with g_free in qemuFirmware*FreeContent() rpc: rename virNetSessionAuthMethodsFree to virNetSessionAuthMethodsClear src/admin/admin_server_dispatch.c | 2 +- src/bhyve/bhyve_conf.c | 6 +- src/bhyve/bhyve_domain.c | 2 +- src/conf/domain_conf.c | 6 +- src/conf/numa_conf.c | 10 +-- src/conf/storage_encryption_conf.c | 2 +- src/conf/virchrdev.c | 6 +- src/libvirt-domain.c | 18 +++--- src/libvirt-network.c | 14 ++--- src/libxl/libxl_domain.c | 2 +- src/libxl/libxl_driver.c | 2 +- src/libxl/libxl_migration.c | 6 +- src/locking/lock_daemon.c | 6 +- src/locking/lock_daemon_config.c | 6 +- src/locking/lock_driver_lockd.c | 10 +-- src/locking/lock_driver_sanlock.c | 6 +- src/locking/lock_manager.c | 2 +- src/logging/log_daemon.c | 4 +- src/logging/log_daemon_config.c | 6 +- src/logging/log_handler.c | 6 +- src/logging/log_manager.c | 2 +- src/qemu/qemu_block.c | 36 +++++------ src/qemu/qemu_capabilities.c | 8 +-- src/qemu/qemu_cgroup.c | 4 +- src/qemu/qemu_conf.c | 6 +- src/qemu/qemu_domain.c | 14 ++--- src/qemu/qemu_firmware.c | 42 ++++++------- src/qemu/qemu_migration_params.c | 4 +- src/qemu/qemu_monitor.c | 48 +++++++-------- src/qemu/qemu_monitor_json.c | 22 +++---- src/qemu/qemu_process.c | 24 ++++---- src/qemu/qemu_saveimage.c | 6 +- src/qemu/qemu_slirp.c | 2 +- src/qemu/qemu_vhost_user.c | 8 +-- src/remote/remote_daemon_config.c | 48 +++++++-------- src/remote/remote_daemon_dispatch.c | 4 +- src/remote/remote_driver.c | 2 +- src/rpc/virnetmessage.c | 2 +- src/rpc/virnetsshsession.c | 6 +- src/security/security_dac.c | 8 +-- src/security/security_selinux.c | 10 +-- src/storage/storage_backend_fs.c | 6 +- src/storage/storage_backend_rbd.c | 10 +-- src/storage/storage_backend_scsi.c | 4 +- src/storage/storage_driver.c | 4 +- src/test/test_driver.c | 6 +- src/util/virarptable.c | 8 +-- src/util/virauthconfig.c | 4 +- src/util/virbitmap.c | 4 +- src/util/vircgroup.c | 12 ++-- src/util/vircommand.c | 24 ++++---- src/util/virconf.c | 10 +-- src/util/virdnsmasq.c | 32 +++++----- src/util/virebtables.c | 4 +- src/util/virfdstream.c | 10 +-- src/util/virfile.c | 4 +- src/util/virfirewall.c | 16 ++--- src/util/virfirmware.c | 8 +-- src/util/virjson.c | 12 ++-- src/util/virlockspace.c | 12 ++-- src/util/virlog.c | 12 ++-- src/util/virmacaddr.c | 2 +- src/util/virmdev.c | 16 ++--- src/util/virnetdev.c | 12 ++-- src/util/virnetdevbandwidth.c | 6 +- src/util/virnetdevip.c | 6 +- src/util/virnetdevmacvlan.c | 8 +-- src/util/virnetdevvlan.c | 2 +- src/util/virnvme.c | 2 +- src/util/virobject.c | 2 +- src/util/virpci.c | 18 +++--- src/util/virperf.c | 2 +- src/util/virportallocator.c | 4 +- src/util/virresctrl.c | 6 +- src/util/virrotatingfile.c | 14 ++--- src/util/virscsi.c | 16 ++--- src/util/virscsivhost.c | 10 +-- src/util/virseclabel.c | 16 ++--- src/util/virsocketaddr.c | 2 +- src/util/virsysinfo.c | 96 ++++++++++++++--------------- src/util/virsystemd.c | 6 +- src/util/virthreadpool.c | 6 +- src/util/virtypedparam-public.c | 2 +- src/util/virtypedparam.c | 8 +-- src/util/viruri.c | 20 +++--- src/util/virusb.c | 8 +-- src/util/virxml.c | 4 +- src/vbox/vbox_snapshot_conf.c | 54 ++++++++-------- src/vmx/vmx.c | 6 +- src/vz/vz_driver.c | 8 +-- src/vz/vz_utils.c | 2 +- tests/nodedevmdevctltest.c | 4 +- tests/nwfilterxml2firewalltest.c | 2 +- tests/qemuhotplugtest.c | 12 ++-- tests/qemumonitortestutils.c | 28 ++++----- tests/virnetdaemontest.c | 2 +- tests/virnetserverclienttest.c | 2 +- tools/virsh-checkpoint.c | 6 +- tools/virsh-domain-monitor.c | 4 +- tools/virsh-domain.c | 2 +- tools/virsh-interface.c | 4 +- tools/virsh-network.c | 8 +-- tools/virsh-nodedev.c | 4 +- tools/virsh-nwfilter.c | 8 +-- tools/virsh-pool.c | 4 +- tools/virsh-secret.c | 4 +- tools/virsh-snapshot.c | 6 +- tools/virsh-volume.c | 4 +- tools/vsh-table.c | 10 +-- tools/vsh.c | 6 +- 110 files changed, 557 insertions(+), 557 deletions(-) -- 2.29.2

2 26

[libvirt PATCH] qemu_monitor_json: fix JSON generator for VC chardev
by Pavel Hrdina 04 Feb '21

04 Feb '21

The correct backend type is 'vc', same as in qemuBuildChrChardevStr() where we generate qemu command line. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/qemu/qemu_monitor_json.c | 5 ++++- tests/qemumonitorjsontest.c | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 97c5e5b36c..231f0ded32 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -7438,10 +7438,13 @@ qemuMonitorJSONAttachCharDevCommand(const char *chrID, switch ((virDomainChrType)chr->type) { case VIR_DOMAIN_CHR_TYPE_NULL: - case VIR_DOMAIN_CHR_TYPE_VC: backend_type = "null"; break; + case VIR_DOMAIN_CHR_TYPE_VC: + backend_type = "vc"; + break; + case VIR_DOMAIN_CHR_TYPE_PTY: backend_type = "pty"; break; diff --git a/tests/qemumonitorjsontest.c b/tests/qemumonitorjsontest.c index 29c396891b..956423f10f 100644 --- a/tests/qemumonitorjsontest.c +++ b/tests/qemumonitorjsontest.c @@ -718,7 +718,7 @@ qemuMonitorJSONTestAttachChardev(virDomainXMLOptionPtr xmlopt, chr = (virDomainChrSourceDef) { .type = VIR_DOMAIN_CHR_TYPE_VC }; CHECK("vc", false, - "{'id':'alias','backend':{'type':'null','data':{}}}"); + "{'id':'alias','backend':{'type':'vc','data':{}}}"); chr = (virDomainChrSourceDef) { .type = VIR_DOMAIN_CHR_TYPE_PTY }; if (qemuMonitorJSONTestAttachOneChardev(xmlopt, schema, "pty", &chr, -- 2.29.2

2 3

[libvirt PATCH] qemu_driver: increase recorded counter for disk block stats
by Pavel Hrdina 04 Feb '21

04 Feb '21

Commit <318d807a0bd3372b634d1952b559c5c627ccfa5b> added a fix to skip most of the block stat code to not log error message for missing storage sources but forgot to increase the recordnr counter. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/qemu/qemu_driver.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 69fcd28666..c34af6b7d1 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -18414,8 +18414,14 @@ qemuDomainGetStatsBlockExportDisk(virDomainDiskDefPtr disk, VIR_INFO("optional disk '%s' source file is missing, " "skip getting stats", disk->dst); - return qemuDomainGetStatsBlockExportHeader(disk, disk->src, *recordnr, - params); + if (qemuDomainGetStatsBlockExportHeader(disk, disk->src, *recordnr, + params) < 0) { + return -1; + } + + (*recordnr)++; + + return 0; } /* vhost-user disk doesn't support getting block stats */ -- 2.29.2

2 1

[libvirt PATCH 0/2] tests: Improve macOS stat() mocking logic
by Andrea Bolognani 04 Feb '21

04 Feb '21

See 2/2 for more information. Andrea Bolognani (2): tests: Split macOS stat() mocking logic tests: Improve macOS stat() mocking logic tests/virmockstathelpers.c | 75 ++++++++++++++++++++++---------------- 1 file changed, 43 insertions(+), 32 deletions(-) -- 2.26.2

2 3

[libvirt PATCH 00/21] eliminate VIR_FREE in all *Free() functions
by Laine Stump 04 Feb '21

04 Feb '21

When I sent a patch to convert (what I thought was) all of the VIR_FREEs in any *Free() function in the conf directory to g_free (with the justification that it was trivial to verify that the pointers weren't being referenced after they had been freed), there were two responses that stuck in my mind: 1) Daniel said that this was a reasonable class of uses to change and a reasonable way to arrive at the desired end result. 2) Peter justifiably expressing concern over the amount of churn in the code, and also said that he'd rather not have things "halfway" for an indeterminate time. The combination of those two comments (ignoring the part about "concern for churn" :-) let me to sit down last night and make this patch series that eliminates all uses of VIR_FREE from any function whose name ends in "Free" (or eliminate/rename the function, just for completeness' sake) In the vast majority of cases, this was done by replacing the VIR_FREEs in the functions with g_free (while verifying that everything being g_freed was actually something pointed to by the parent object that was sent to the *Free() function, and that the parent object itself was subsequently freed). But there were a a couple of cases where this wasn't quite the right thing to do: * in patch 20, two functions ending with Free() don't actually free the pointer they're sent. They instead behave like a *Clear() function, VIR_FREEing their components. Since we can see that these functions aren't actually be used as *Clear() functions (and never will be used that way), but don't want to promote confusion by leaving them misnamed, the patch renames them to *FreeContent() (more accurate name) and changes their VIR_FREEs to g_frees (it takes more than just looking at one function, but it's very easy to verify that the pointers are never referenced after they're freed). * In Patch 21, a *Free() function actually appears to be used as a *Clear() function, I left the VIR_FREEs in place, but renamed the *function to *Clear(). However, its one use as a *Clear() is when *being called from a "Reset" function that (as far as I can see) *is never used. If we could get rid of that Reset function, I *could just fold the *Clear() function into its one remaining *caller, and eliminate the VIR_FREEs. There is a companion series that eliminates VIR_FREE from all *Dispose() functions, but I'm sending it separately since the two series are completely independent. The last time I ran wc over the diffs of these patches, it showed they are eliminating 436 uses of VIR_FREE, which is > 10% of the 4k remaining. Laine Stump (21): conf: replace remaining straggler VIR_FREE with g_free in vir*Free() util: replace VIR_FREE with g_free in all vir*Free() functions bhyve: replace VIR_FREE with g_free in all vir*Free() functions libxl: replace VIR_FREE with g_free in all vir*Free() functions qemu: replace VIR_FREE with g_free in all vir*Free() functions test_driver: replace VIR_FREE with g_free in all vir*Free() functions vbox: replace VIR_FREE with g_free in all vir*Free() functions vmx: replace VIR_FREE with g_free in all vir*Free() functions vz: replace VIR_FREE with g_free in all vir*Free() functions admin: replace VIR_FREE with g_free in all vir*Free() functions locking: replace VIR_FREE with g_free in all vir*Free() functions logging: replace VIR_FREE with g_free in all vir*Free() functions remote: replace VIR_FREE with g_free in all vir*Free() functions rpc: replace VIR_FREE with g_free in all vir*Free() functions security: replace VIR_FREE with g_free in all vir*Free() functions tools: replace VIR_FREE with g_free in all vir*Free() functions tests: replace VIR_FREE with g_free in all vir*Free() functions storage: replace VIR_FREE with g_free in all vir*Free() functions libvirtd: replace straggler VIR_FREE with g_free in all vir*Free() functions util: rename two *Free() functions while changing VIR_FREE to g_free rpc: rename virNetSessionAuthMethodsFree to virNetSessionAuthMethodsClear src/admin/admin_server_dispatch.c | 2 +- src/bhyve/bhyve_conf.c | 6 +- src/bhyve/bhyve_domain.c | 2 +- src/conf/domain_conf.c | 6 +- src/conf/numa_conf.c | 10 +-- src/conf/storage_encryption_conf.c | 2 +- src/conf/virchrdev.c | 6 +- src/libvirt-domain.c | 18 +++--- src/libvirt-network.c | 14 ++--- src/libxl/libxl_domain.c | 2 +- src/libxl/libxl_driver.c | 2 +- src/libxl/libxl_migration.c | 6 +- src/locking/lock_daemon.c | 6 +- src/locking/lock_daemon_config.c | 6 +- src/locking/lock_driver_lockd.c | 10 +-- src/locking/lock_driver_sanlock.c | 6 +- src/locking/lock_manager.c | 2 +- src/logging/log_daemon.c | 4 +- src/logging/log_daemon_config.c | 6 +- src/logging/log_handler.c | 6 +- src/logging/log_manager.c | 2 +- src/qemu/qemu_block.c | 36 +++++------ src/qemu/qemu_capabilities.c | 8 +-- src/qemu/qemu_cgroup.c | 4 +- src/qemu/qemu_conf.c | 6 +- src/qemu/qemu_domain.c | 14 ++--- src/qemu/qemu_firmware.c | 10 +-- src/qemu/qemu_migration_params.c | 4 +- src/qemu/qemu_monitor.c | 48 +++++++-------- src/qemu/qemu_monitor_json.c | 22 +++---- src/qemu/qemu_process.c | 24 ++++---- src/qemu/qemu_saveimage.c | 6 +- src/qemu/qemu_slirp.c | 2 +- src/qemu/qemu_vhost_user.c | 8 +-- src/remote/remote_daemon_config.c | 48 +++++++-------- src/remote/remote_daemon_dispatch.c | 4 +- src/remote/remote_driver.c | 2 +- src/rpc/virnetmessage.c | 2 +- src/rpc/virnetsshsession.c | 6 +- src/security/security_dac.c | 8 +-- src/security/security_selinux.c | 10 +-- src/storage/storage_backend_fs.c | 6 +- src/storage/storage_backend_rbd.c | 10 +-- src/storage/storage_backend_scsi.c | 4 +- src/storage/storage_driver.c | 4 +- src/test/test_driver.c | 6 +- src/util/virarptable.c | 8 +-- src/util/virauthconfig.c | 4 +- src/util/virbitmap.c | 4 +- src/util/vircgroup.c | 12 ++-- src/util/vircommand.c | 24 ++++---- src/util/virconf.c | 10 +-- src/util/virdnsmasq.c | 32 +++++----- src/util/virebtables.c | 4 +- src/util/virfdstream.c | 10 +-- src/util/virfile.c | 4 +- src/util/virfirewall.c | 16 ++--- src/util/virfirmware.c | 8 +-- src/util/virjson.c | 12 ++-- src/util/virlockspace.c | 12 ++-- src/util/virlog.c | 12 ++-- src/util/virmacaddr.c | 2 +- src/util/virmdev.c | 16 ++--- src/util/virnetdev.c | 12 ++-- src/util/virnetdevbandwidth.c | 6 +- src/util/virnetdevip.c | 6 +- src/util/virnetdevmacvlan.c | 8 +-- src/util/virnetdevvlan.c | 2 +- src/util/virnvme.c | 2 +- src/util/virobject.c | 2 +- src/util/virpci.c | 18 +++--- src/util/virperf.c | 2 +- src/util/virportallocator.c | 4 +- src/util/virresctrl.c | 6 +- src/util/virrotatingfile.c | 14 ++--- src/util/virscsi.c | 16 ++--- src/util/virscsivhost.c | 10 +-- src/util/virseclabel.c | 16 ++--- src/util/virsocketaddr.c | 2 +- src/util/virsysinfo.c | 96 ++++++++++++++--------------- src/util/virsystemd.c | 6 +- src/util/virthreadpool.c | 6 +- src/util/virtypedparam-public.c | 2 +- src/util/virtypedparam.c | 8 +-- src/util/viruri.c | 20 +++--- src/util/virusb.c | 8 +-- src/util/virxml.c | 4 +- src/vbox/vbox_snapshot_conf.c | 54 ++++++++-------- src/vmx/vmx.c | 6 +- src/vz/vz_driver.c | 8 +-- src/vz/vz_utils.c | 2 +- tests/nodedevmdevctltest.c | 4 +- tests/nwfilterxml2firewalltest.c | 2 +- tests/qemuhotplugtest.c | 12 ++-- tests/qemumonitortestutils.c | 28 ++++----- tests/virnetdaemontest.c | 2 +- tests/virnetserverclienttest.c | 2 +- tools/virsh-checkpoint.c | 6 +- tools/virsh-domain-monitor.c | 4 +- tools/virsh-domain.c | 2 +- tools/virsh-interface.c | 4 +- tools/virsh-network.c | 8 +-- tools/virsh-nodedev.c | 4 +- tools/virsh-nwfilter.c | 8 +-- tools/virsh-pool.c | 4 +- tools/virsh-secret.c | 4 +- tools/virsh-snapshot.c | 6 +- tools/virsh-volume.c | 4 +- tools/vsh-table.c | 10 +-- tools/vsh.c | 6 +- 110 files changed, 541 insertions(+), 541 deletions(-) -- 2.29.2

1 22