[libvirt] [PATCH] Remove some C99 variable decls in parallels driver
by Daniel P. Berrange
From: "Daniel P. Berrange" <berrange(a)redhat.com>
The parallels storage driver declared some loop variables
inside the for(;;). This is not allowed by libvirt coding
standards
Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com>
Pushed under build breaker rule - RHEL-5 gcc chokes on them
---
src/parallels/parallels_storage.c | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/src/parallels/parallels_storage.c b/src/parallels/parallels_storage.c
index f696fcb..ac65a1d 100644
--- a/src/parallels/parallels_storage.c
+++ b/src/parallels/parallels_storage.c
@@ -129,9 +129,11 @@ static char *parallelsMakePoolName(virConnectPtr conn, const char *path)
{
parallelsConnPtr privconn = conn->privateData;
char *name;
+ unsigned int i;
- for (unsigned int i = 0; i < UINT_MAX; i++) {
+ for (i = 0; i < UINT_MAX; i++) {
bool found = false;
+ int j;
if (!(name = strdup(path))) {
virReportOOMError();
@@ -148,11 +150,11 @@ static char *parallelsMakePoolName(virConnectPtr conn, const char *path)
return 0;
}
- for (int j = 0; j < strlen(name); j++)
+ for (j = 0; j < strlen(name); j++)
if (name[j] == '/')
name[j] = '-';
- for (int j = 0; j < privconn->pools.count; j++) {
+ for (j = 0; j < privconn->pools.count; j++) {
if (STREQ(name, privconn->pools.objs[j]->def->name)) {
found = true;
break;
@@ -226,6 +228,7 @@ parallelsPoolAddByDomain(virConnectPtr conn, virDomainObjPtr dom)
virStoragePoolObjListPtr pools = &privconn->pools;
char *poolPath;
virStoragePoolObjPtr pool = NULL;
+ int j;
if (!(poolPath = strdup(pdom->home))) {
virReportOOMError();
@@ -234,7 +237,7 @@ parallelsPoolAddByDomain(virConnectPtr conn, virDomainObjPtr dom)
poolPath = dirname(poolPath);
- for (int j = 0; j < pools->count; j++) {
+ for (j = 0; j < pools->count; j++) {
if (STREQ(poolPath, pools->objs[j]->def->target.path)) {
pool = pools->objs[j];
break;
--
1.8.1.2
12 years, 2 months
[libvirt] [PATCH 0/2] fix some RHEL 5 test failures
by Eric Blake
A couple of patches that I noticed while testing the release
candidate on a RHEL 5 VM.
Eric Blake (2):
tests: consistent skip messages
tests: old automake lacks abs_builddir
tests/Makefile.am | 25 ++-----------------------
tests/qemumonitorjsontest.c | 2 +-
tests/virstoragetest.c | 4 ++--
3 files changed, 5 insertions(+), 26 deletions(-)
--
1.8.1.2
12 years, 2 months
[libvirt] [PATCH 0/4] Virtio-RNG followup
by Peter Krempa
This series adds rate limiting and multiple device support and fixes
docs and adds test for XML entities and multiple RNG devices.
Peter Krempa (4):
virtio-rng: allow multiple RNG devices
docs: Fix attribute name for virtio-rng backend
virtio-rng: Add rate limiting options for virtio-RNG
tests: Test multiple RNG devices and support for XML entities in paths
docs/formatdomain.html.in | 14 +++-
docs/schemas/domaincommon.rng | 18 ++++-
src/conf/domain_conf.c | 78 +++++++++++++---------
src/conf/domain_conf.h | 6 +-
src/qemu/qemu_command.c | 28 +++++---
.../qemuxml2argv-virtio-rng-random.args | 2 +-
.../qemuxml2argv-virtio-rng-random.xml | 4 ++
7 files changed, 106 insertions(+), 44 deletions(-)
--
1.8.1.1
12 years, 2 months
[libvirt] [PATCH] vl.c: Support multiple CPU ranges on -numa option
by Eduardo Habkost
This allows "," to be used a separator between each CPU range. Note
that commas inside key=value command-line options have to be escaped
using ",,", so the command-line will look like:
-numa node,cpus=A,,B,,C,,D
Note that the following format, currently used by libvirt:
-numa nodes,cpus=A,B,C,D
will _not_ work yet, as "," is the option separator for the command-line
option parser, and it will require changing the -numa option parsing
code to handle "cpus" as a special case.
Signed-off-by: Eduardo Habkost <ehabkost(a)redhat.com>
---
vl.c | 14 +++++++++++++-
1 file changed, 13 insertions(+), 1 deletion(-)
diff --git a/vl.c b/vl.c
index 955d2ff..cd247be 100644
--- a/vl.c
+++ b/vl.c
@@ -1244,7 +1244,7 @@ char *get_boot_devices_list(size_t *size)
return list;
}
-static void numa_node_parse_cpus(int nodenr, const char *cpus)
+static void numa_node_parse_cpu_range(int nodenr, const char *cpus)
{
char *endptr;
unsigned long long value, endvalue;
@@ -1288,6 +1288,18 @@ error:
exit(1);
}
+static void numa_node_parse_cpus(int nodenr, const char *option)
+{
+ char **parts;
+ int i;
+
+ parts = g_strsplit(option, ",", 0);
+ for (i = 0; parts[i]; i++) {
+ numa_node_parse_cpu_range(nodenr, parts[i]);
+ }
+ g_strfreev(parts);
+}
+
static void numa_add(const char *optarg)
{
char option[128];
--
1.8.1
12 years, 2 months
[libvirt] [PATCH] qemu: fix graphics port allocation
by Ján Tomko
Right now, we allocate a port or a TLS port for SPICE if it's set to -1,
even if autoport is off. But we only free them if autoport is on.
With autoport on, we only allocate TLS port if cfg->spiceTLS is set, but
we free it even if it's not, leading to an error message.
This patch separates the autoport=yes|no XML option into autoport and
tlsAutoport in virDomainGraphicsDef:
if autoport is yes, we set them both to 1 (and vice versa)
if either of the port numbers is -1, the corresponding autoport is set
to 1 and it's used as a condition for acquiring/releasing the port.
For TLS ports, cfg->spiceTLS is checked before releasing the port as
well.
Also check the port number before trying to release it - the vnc port
will be 0, the spice port will be -1 or 0 if it hasn't been allocated
yet (if qemuProcessStart fails before port allocation).
---
src/conf/domain_conf.c | 17 +++++++++++------
src/conf/domain_conf.h | 1 +
src/qemu/qemu_hotplug.c | 3 ++-
src/qemu/qemu_process.c | 26 ++++++++++++++------------
4 files changed, 28 insertions(+), 19 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index b65e52a..cb27f82 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7083,8 +7083,10 @@ virDomainGraphicsDefParseXML(xmlNodePtr node,
}
if ((autoport = virXMLPropString(node, "autoport")) != NULL) {
- if (STREQ(autoport, "yes"))
+ if (STREQ(autoport, "yes")) {
def->data.spice.autoport = 1;
+ def->data.spice.tlsAutoport = 1;
+ }
VIR_FREE(autoport);
}
@@ -7102,12 +7104,14 @@ virDomainGraphicsDefParseXML(xmlNodePtr node,
VIR_FREE(defaultMode);
}
- if (def->data.spice.port == -1 && def->data.spice.tlsPort == -1) {
- /* Legacy compat syntax, used -1 for auto-port */
+ /* Legacy compat syntax, used -1 for auto-port */
+ if (def->data.spice.port == -1)
def->data.spice.autoport = 1;
- }
+ if (def->data.spice.tlsPort == -1)
+ def->data.spice.tlsAutoport = 1;
- if (def->data.spice.autoport && (flags & VIR_DOMAIN_XML_INACTIVE)) {
+ if (def->data.spice.autoport && def->data.spice.tlsAutoport &&
+ (flags & VIR_DOMAIN_XML_INACTIVE)) {
def->data.spice.port = 0;
def->data.spice.tlsPort = 0;
}
@@ -14201,7 +14205,8 @@ virDomainGraphicsDefFormat(virBufferPtr buf,
def->data.spice.tlsPort);
virBufferAsprintf(buf, " autoport='%s'",
- def->data.spice.autoport ? "yes" : "no");
+ (def->data.spice.autoport &&
+ def->data.spice.tlsAutoport) ? "yes" : "no");
if (listenAddr)
virBufferAsprintf(buf, " listen='%s'", listenAddr);
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index 0828954..cc67716 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -1356,6 +1356,7 @@ struct _virDomainGraphicsDef {
char *keymap;
virDomainGraphicsAuthDef auth;
unsigned int autoport :1;
+ unsigned int tlsAutoport :1;
int channels[VIR_DOMAIN_GRAPHICS_SPICE_CHANNEL_LAST];
int defaultMode; /* enum virDomainGraphicsSpiceChannelMode */
int image;
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index 78961a7..d8f9007 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -1879,9 +1879,10 @@ qemuDomainChangeGraphics(virQEMUDriverPtr driver,
case VIR_DOMAIN_GRAPHICS_TYPE_SPICE:
if ((olddev->data.spice.autoport != dev->data.spice.autoport) ||
+ (olddev->data.spice.tlsAutoport != dev->data.spice.tlsAutoport) ||
(!dev->data.spice.autoport &&
(olddev->data.spice.port != dev->data.spice.port)) ||
- (!dev->data.spice.autoport &&
+ (!dev->data.spice.tlsAutoport &&
(olddev->data.spice.tlsPort != dev->data.spice.tlsPort))) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("cannot change port settings on spice graphics"));
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index 732964f..7c2a54e 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -3622,8 +3622,7 @@ int qemuProcessStart(virConnectPtr conn,
graphics->data.vnc.port = port;
} else if (graphics->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
unsigned short port = 0;
- if (graphics->data.spice.autoport ||
- graphics->data.spice.port == -1) {
+ if (graphics->data.spice.autoport) {
if (virPortAllocatorAcquire(driver->remotePorts, &port) < 0)
goto cleanup;
@@ -3635,9 +3634,7 @@ int qemuProcessStart(virConnectPtr conn,
graphics->data.spice.port = port;
}
- if (cfg->spiceTLS &&
- (graphics->data.spice.autoport ||
- graphics->data.spice.tlsPort == -1)) {
+ if (cfg->spiceTLS && graphics->data.spice.tlsAutoport) {
unsigned short tlsPort;
if (virPortAllocatorAcquire(driver->remotePorts, &tlsPort) < 0)
goto cleanup;
@@ -4318,16 +4315,21 @@ retry:
for (i = 0 ; i < vm->def->ngraphics; ++i) {
virDomainGraphicsDefPtr graphics = vm->def->graphics[i];
if (graphics->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
- graphics->data.vnc.autoport) {
+ graphics->data.vnc.autoport && graphics->data.vnc.port) {
ignore_value(virPortAllocatorRelease(driver->remotePorts,
graphics->data.vnc.port));
}
- if (graphics->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE &&
- graphics->data.spice.autoport) {
- ignore_value(virPortAllocatorRelease(driver->remotePorts,
- graphics->data.spice.port));
- ignore_value(virPortAllocatorRelease(driver->remotePorts,
- graphics->data.spice.tlsPort));
+ if (graphics->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
+ if (graphics->data.spice.autoport &&
+ graphics->data.spice.port > 0) {
+ ignore_value(virPortAllocatorRelease(driver->remotePorts,
+ graphics->data.spice.port));
+ }
+ if (cfg->spiceTLS && graphics->data.spice.tlsAutoport &&
+ graphics->data.spice.tlsPort > 0) {
+ ignore_value(virPortAllocatorRelease(driver->remotePorts,
+ graphics->data.spice.tlsPort));
+ }
}
}
--
1.7.12.4
12 years, 2 months
[libvirt] [PATCH] libvirt: fix error message when connection can't be opened
by Ján Tomko
VIR_ERR_NO_CONNECT already contains "no connection driver available".
This patch changes:
no connection driver available for No connection for URI hello
to:
no connection driver available for hello
Bug: https://bugzilla.redhat.com/show_bug.cgi?id=851413
---
src/libvirt.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/libvirt.c b/src/libvirt.c
index 934997a..8a28e4a 100644
--- a/src/libvirt.c
+++ b/src/libvirt.c
@@ -1216,7 +1216,7 @@ do_open(const char *name,
if (!ret->driver) {
/* If we reach here, then all drivers declined the connection. */
virLibConnError(VIR_ERR_NO_CONNECT,
- _("No connection for URI %s"),
+ "%s",
NULLSTR(name));
goto failed;
}
--
1.7.12.4
12 years, 2 months
[libvirt] [PATCH 0/4] uid_t|gid_t fixes for 32 bit Linux
by Philipp Hahn
Hello,
on 2012-11-23 I already reported a problem with running libvirt on 32 Bit
Linuxes (libvirt 0.9.12 on Debian Squeeze in my case) on this ML:
[BUG] storage.xml: owner|group=-1 → 2^32-1 on 32 Bit
Due to multiple castings the initial -1 gets casted to (unsigned
int)4294967295, which failes to be cased to an (signed int) on read-back.
Thus the default storage pool is not launched:
2013-02-22 16:50:10.182+0000: 28550: info : libvirt version: 0.9.12
2013-02-22 16:50:10.182+0000: 28550: error : virStorageDefParsePerms:635 : XML error: malformed owner element
This is caused by uid_t and gid_t being opaque types: their exact type ranges
from s32 to u32 to u64. (Solaris until 2007, Linux, Windows64)
The following patch series fixes several issues, which I found while
investigating the storage pool startup problem:
- mismatch between printf format and types
- implicit casts of -1 to uid_t and gid_t
- missing casts of uid_t and gid_t for printing
- virXPathLong() failing on +(2^32-1)
@Guido: for Debian only the 4th patch is relevant, which applies with some
re-indention.
Additional issues:
- UID and GID 4294967295 happens to be the same as -1, so hopefully nobody
uses it. Should this be documented somewhere?
- Some regression test?
Philipp Hahn (4):
util: Fix printf format for uid_t|gid_t
storage: Cast uid_t|gid_t to unsigned int
storage: cast -1 for uid_t|gid_t
storage: fix uid_t|gid_t handling on 32 bit Linux
src/conf/storage_conf.c | 82 ++++++++++++++++++++++++++++++++---------
src/storage/storage_backend.c | 22 ++++++-----
src/util/virutil.c | 10 ++---
3 files changed, 81 insertions(+), 33 deletions(-)
--
1.7.10.4
12 years, 2 months
[libvirt] [PATCH] qemu: do not set unpriv_sgio if neither supported nor requested
by Paolo Bonzini
Currently we call virSetDeviceUnprivSGIO with val == 0 if a block device
has an sgio attribute. But for sgio='filtered', we know that a
kernel with no unpriv_sgio support will always behave as the user
wanted. In this case, there is no need to call the function and
report a (bogus) error.
Signed-off-by: Paolo Bonzini <pbonzini(a)redhat.com>
---
src/qemu/qemu_process.c | 35 +++++++++++++++++------------------
1 file changed, 17 insertions(+), 18 deletions(-)
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index b560d2e..4c152b2 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -3455,36 +3455,35 @@ qemuProcessReconnectAll(virConnectPtr conn, virQEMUDriverPtr driver)
int
qemuSetUnprivSGIO(virDomainDiskDefPtr disk)
{
+ char *sysfs_path = NULL;
int val = -1;
+ int ret = 0;
/* "sgio" is only valid for block disk; cdrom
* and floopy disk can have empty source.
*/
if (disk->type != VIR_DOMAIN_DISK_TYPE_BLOCK ||
+ disk->device != VIR_DOMAIN_DISK_DEVICE_LUN ||
!disk->src)
return 0;
- if (disk->sgio)
- val = (disk->sgio == VIR_DOMAIN_DISK_SGIO_UNFILTERED);
-
- /* Ignore the setting if unpriv_sgio is not supported by the
- * kernel, otherwise defaults to filter the SG_IO commands,
- * I.E. Set unpriv_sgio to 0.
- */
- if (disk->sgio == VIR_DOMAIN_DISK_SGIO_DEFAULT &&
- disk->device == VIR_DOMAIN_DISK_DEVICE_LUN) {
- char *sysfs_path = NULL;
+ sysfs_path = virGetUnprivSGIOSysfsPath(disk->src, NULL);
+ if (sysfs_path == NULL)
+ return -1;
- if ((sysfs_path = virGetUnprivSGIOSysfsPath(disk->src, NULL)) &&
- virFileExists(sysfs_path))
- val = 0;
- VIR_FREE(sysfs_path);
- }
+ /* By default, filter the SG_IO commands, i.e. set unpriv_sgio to 0. */
+ val = (disk->sgio == VIR_DOMAIN_DISK_SGIO_UNFILTERED);
- if (val >= 0 && virSetDeviceUnprivSGIO(disk->src, NULL, val) < 0)
- return -1;
+ /* Do not do anything if unpriv_sgio is not supported by the kernel and the
+ * whitelist is enabled. But if requesting unfiltered access, always call
+ * virSetDeviceUnprivSGIO, to report an error for unsupported unpriv_sgio.
+ */
+ if ((virFileExists(sysfs_path) || val == 1) &&
+ virSetDeviceUnprivSGIO(disk->src, NULL, val) < 0)
+ ret = -1;
- return 0;
+ VIR_FREE(sysfs_path);
+ return ret;
}
int qemuProcessStart(virConnectPtr conn,
--
1.8.1.2
12 years, 2 months
[libvirt] RFC: APIs for managing resource groups
by Daniel P. Berrange
Historically for the QEMU/LXC drivers we've simply put each virtual
instance in a dedicated cgroup, under the path
$LIBVIRT_CGROUP_LOCATION
|
+- libvirt
|
+- qemu
| |
| +- vm1
| +- vm2
| +- vm3
|
+- lxc
|
+- cont1
+- cont2
+- cont3
for a variety of reasons this nesting sucks. It is too deep causing
kernel performance problems, its structure does not easily allow for
calculating fixed % shares, it does not allow for grouping of VMs.
We need to simplify our layout and also introduce some APIs for the
grouping of VMs. I won't go into specifics of a new cgroups layout
here, just focus on the question of defining a set of APIs that are
generic to any hypervisor, for the purpose of setting up VM resource
groups.
I'm calling the resource cgroup a "partition", since this is all about
partitioning workloads.
I anticipate a new top level object and APIs for creating/defining it
in the usual manner:
typedef struct _virPartition virPartition;
typedef virPartition *virPartitionPtr;
int virConnectListAllPartitions(virConnectPtr conn,
virPartitionPtr **partitions,
unsigned int flags);
virPartitionPtr virPartitionDefineXML(virConnectPtr conn,
const char *xml,
unsigned int flags);
int virPartitionCreate(virPartitionPtr partition,
unsigned int flags);
int virPartitionCreateXML(virPartitionPtr partition,
const char *xml,
unsigned int flags);
int virPartitionDestroy(virPartitionPtr partition,
unsigned int flags);
int virPartitionUndefine(virPartitionPtr partition,
unsigned int flags);
Then I think we'll duplicate all the APIs for setting resource tunables
from virDomainPtr against the new object, so we get
int virPartitionGetSchedulerParameters(virPartitionPtr partition,
virTypedParameterPtr params,
int *nparams,
unsigned int flags);
int virPartitionSetSchedulerParameters(virPartitionPtr partition,
virTypedParameterPtr params,
int nparams,
unsigned int flags)
int virDomainSetBlkioParameters(virDomainPtr domain,
virTypedParameterPtr params,
int nparams,
unsigned int flags);
int virDomainGetBlkioParameters(virDomainPtr domain,
virTypedParameterPtr params,
int *nparams,
unsigned int flags);
int virDomainSetMemoryParameters(virDomainPtr domain,
virTypedParameterPtr params,
int nparams,
unsigned int flags);
int virDomainGetMemoryParameters(virDomainPtr domain,
virTypedParameterPtr params,
int *nparams,
unsigned int flags);
int virDomainSetNumaParameters(virDomainPtr domain,
virTypedParameterPtr params,
int nparams,
unsigned int flags);
int virDomainGetNumaParameters(virDomainPtr domain,
virTypedParameterPtr params,
int *nparams,
unsigned int flags);
Finally we need a way to associate a domain with a partition
virPartitionPtr virDomainGetPartition(virDomainPtr dom,
unsigned int flags);
void virDomainSetPartition(virDomainPtr dom,
unsigned int flags);
There'd also likely be a new VM XML element
<partition name="..partition name..."/>
which is what the Get/SetPartition methods would be touching.
Regards,
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
12 years, 2 months
