8:14 a.m.
I/O threads and vCPU threads already support setting schedulers, but until now
it was impossible to do so for the main QEMU thread (emulator thread in the
libvirt naming). This is, however, requested for some very specific scenarios,
for example when vCPU threads are running at such priority that could starve the
main thread.
v2:
- news.xml addition in a separate patch
Martin Kletzander (6):
conf: Parse common scheduler attributes in separate function
conf: Format thread IDs optionally
docs: Mention iothreadsched element in the docs and reword
conf: Add support for emulatorsched
qemu: Add support for emulatorsched
docs: Add emulatorsched support to news.xml
docs/formatdomain.html.in | 25 +++--
docs/news.xml | 12 +++
docs/schemas/domaincommon.rng | 5 +
src/conf/domain_conf.c | 136 +++++++++++++++++-------
src/conf/domain_conf.h | 1 +
src/qemu/qemu_process.c | 2 +-
tests/genericxml2xmlindata/cputune.xml | 37 +++++++
tests/genericxml2xmloutdata/cputune.xml | 40 +++++++
tests/genericxml2xmltest.c | 2 +
9 files changed, 212 insertions(+), 48 deletions(-)
create mode 100644 tests/genericxml2xmlindata/cputune.xml
create mode 100644 tests/genericxml2xmloutdata/cputune.xml
--
2.21.0
8:14 a.m.
New subject: [libvirt] [PATCH v2 1/6] conf: Parse common scheduler attributes in separate function
This will become useful later when parsing emulatorsched parameters which don't
need the rest of the current function.
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
src/conf/domain_conf.c | 70 +++++++++++++++++++++++++-----------------
1 file changed, 41 insertions(+), 29 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index b969a9f6e566..15838c2a23f5 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -18441,45 +18441,24 @@ virDomainLoaderDefParseXML(xmlNodePtr node,
}
-static virBitmapPtr
-virDomainSchedulerParse(xmlNodePtr node,
- const char *name,
- virProcessSchedPolicy *policy,
- int *priority)
+static int
+virDomainSchedulerParseCommonAttrs(xmlNodePtr node,
+ virProcessSchedPolicy *policy,
+ int *priority)
{
- virBitmapPtr ret = NULL;
int pol = 0;
VIR_AUTOFREE(char *) tmp = NULL;
- if (!(tmp = virXMLPropString(node, name))) {
- virReportError(VIR_ERR_XML_ERROR,
- _("Missing attribute '%s' in element
'%sched'"),
- name, name);
- goto error;
- }
-
- if (virBitmapParse(tmp, &ret, VIR_DOMAIN_CPUMASK_LEN) < 0)
- goto error;
-
- if (virBitmapIsAllClear(ret)) {
- virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
- _("'%s' scheduler bitmap '%s' is
empty"),
- name, tmp);
- goto error;
- }
-
- VIR_FREE(tmp);
-
if (!(tmp = virXMLPropString(node, "scheduler"))) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Missing scheduler attribute"));
- goto error;
+ return -1;
}
if ((pol = virProcessSchedPolicyTypeFromString(tmp)) <= 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Invalid scheduler attribute: '%s'"), tmp);
- goto error;
+ return -1;
}
*policy = pol;
@@ -18490,16 +18469,49 @@ virDomainSchedulerParse(xmlNodePtr node,
if (!(tmp = virXMLPropString(node, "priority"))) {
virReportError(VIR_ERR_XML_ERROR, "%s",
_("Missing scheduler priority"));
- goto error;
+ return -1;
}
if (virStrToLong_i(tmp, NULL, 10, priority) < 0) {
virReportError(VIR_ERR_XML_ERROR, "%s",
_("Invalid value for element priority"));
- goto error;
+ return -1;
}
}
+ return 0;
+}
+
+
+static virBitmapPtr
+virDomainSchedulerParse(xmlNodePtr node,
+ const char *name,
+ virProcessSchedPolicy *policy,
+ int *priority)
+{
+ virBitmapPtr ret = NULL;
+ VIR_AUTOFREE(char *) tmp = NULL;
+
+ if (!(tmp = virXMLPropString(node, name))) {
+ virReportError(VIR_ERR_XML_ERROR,
+ _("Missing attribute '%s' in element
'%sched'"),
+ name, name);
+ goto error;
+ }
+
+ if (virBitmapParse(tmp, &ret, VIR_DOMAIN_CPUMASK_LEN) < 0)
+ goto error;
+
+ if (virBitmapIsAllClear(ret)) {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("'%s' scheduler bitmap '%s' is
empty"),
+ name, tmp);
+ goto error;
+ }
+
+ if (virDomainSchedulerParseCommonAttrs(node, policy, priority) < 0)
+ goto error;
+
return ret;
error:
--
2.21.0
8:55 a.m.
New subject: [libvirt] [PATCH v2 1/6] conf: Parse common scheduler attributes in separate function
On Mon, Apr 15, 2019 at 03:14:22PM +0200, Martin Kletzander wrote:
This will become useful later when parsing emulatorsched parameters
which don't
need the rest of the current function.
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
src/conf/domain_conf.c | 70 +++++++++++++++++++++++++-----------------
1 file changed, 41 insertions(+), 29 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index b969a9f6e566..15838c2a23f5 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -18441,45 +18441,24 @@ virDomainLoaderDefParseXML(xmlNodePtr node,
}
-static virBitmapPtr
-virDomainSchedulerParse(xmlNodePtr node,
- const char *name,
Consider making the 'name' XML attribute mandatory only if this 'name'
argument is non-NULL.
- virProcessSchedPolicy *policy,
- int *priority)
Jano
9:19 a.m.
New subject: [libvirt] [PATCH v2 1/6] conf: Parse common scheduler attributes in separate function
On Mon, Apr 15, 2019 at 03:55:26PM +0200, Ján Tomko wrote:
On Mon, Apr 15, 2019 at 03:14:22PM +0200, Martin Kletzander wrote:
>This will become useful later when parsing emulatorsched parameters which don't
>need the rest of the current function.
>
>Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
>---
> src/conf/domain_conf.c | 70 +++++++++++++++++++++++++-----------------
> 1 file changed, 41 insertions(+), 29 deletions(-)
>
>diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
>index b969a9f6e566..15838c2a23f5 100644
>--- a/src/conf/domain_conf.c
>+++ b/src/conf/domain_conf.c
>@@ -18441,45 +18441,24 @@ virDomainLoaderDefParseXML(xmlNodePtr node,
> }
>
>
>-static virBitmapPtr
>-virDomainSchedulerParse(xmlNodePtr node,
>- const char *name,
Consider making the 'name' XML attribute mandatory only if this 'name'
argument is non-NULL.
You mean keeping the function the way it is and calling it for emulatorsched
with NULL? Sure, I can change that.
>- virProcessSchedPolicy *policy,
>- int *priority)
Jano
9:57 a.m.
New subject: [libvirt] [PATCH v2 1/6] conf: Parse common scheduler attributes in separate function
On Mon, Apr 15, 2019 at 04:19:43PM +0200, Martin Kletzander wrote:
On Mon, Apr 15, 2019 at 03:55:26PM +0200, Ján Tomko wrote:
>On Mon, Apr 15, 2019 at 03:14:22PM +0200, Martin Kletzander wrote:
>>This will become useful later when parsing emulatorsched parameters which
don't
>>need the rest of the current function.
>>
>>Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
>>---
>>src/conf/domain_conf.c | 70 +++++++++++++++++++++++++-----------------
>>1 file changed, 41 insertions(+), 29 deletions(-)
>>
>>diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
>>index b969a9f6e566..15838c2a23f5 100644
>>--- a/src/conf/domain_conf.c
>>+++ b/src/conf/domain_conf.c
>>@@ -18441,45 +18441,24 @@ virDomainLoaderDefParseXML(xmlNodePtr node,
>>}
>>
>>
>>-static virBitmapPtr
>>-virDomainSchedulerParse(xmlNodePtr node,
>>- const char *name,
>
>Consider making the 'name' XML attribute mandatory only if this
'name'
>argument is non-NULL.
>
You mean keeping the function the way it is and calling it for emulatorsched
with NULL? Sure, I can change that.
Yes.
Jano
8:14 a.m.
New subject: [libvirt] [PATCH v2 2/6] conf: Format thread IDs optionally
This will be used later when we want to format emulator scheduler parameters
which don't apply for multiple threads.
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
squash! conf: Do not format negative thread IDs
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
src/conf/domain_conf.c | 22 +++++++++++++---------
1 file changed, 13 insertions(+), 9 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 15838c2a23f5..e7b8b51aad0e 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -27205,22 +27205,25 @@ static void
virDomainSchedulerFormat(virBufferPtr buf,
const char *name,
virDomainThreadSchedParamPtr sched,
- size_t id)
+ size_t id,
+ bool multiple_threads)
{
switch (sched->policy) {
case VIR_PROC_POLICY_BATCH:
case VIR_PROC_POLICY_IDLE:
- virBufferAsprintf(buf, "<%ssched "
- "%ss='%zu'
scheduler='%s'/>\n",
- name, name, id,
+ virBufferAsprintf(buf, "<%ssched", name);
+ if (multiple_threads)
+ virBufferAsprintf(buf, " %ss='%zu'", name, id);
+ virBufferAsprintf(buf, " scheduler='%s'/>\n",
virProcessSchedPolicyTypeToString(sched->policy));
break;
case VIR_PROC_POLICY_RR:
case VIR_PROC_POLICY_FIFO:
- virBufferAsprintf(buf, "<%ssched "
- "%ss='%zu' scheduler='%s'
priority='%d'/>\n",
- name, name, id,
+ virBufferAsprintf(buf, "<%ssched", name);
+ if (multiple_threads)
+ virBufferAsprintf(buf, " %ss='%zu'", name, id);
+ virBufferAsprintf(buf, " scheduler='%s'
priority='%d'/>\n",
virProcessSchedPolicyTypeToString(sched->policy),
sched->priority);
break;
@@ -27489,13 +27492,14 @@ virDomainCputuneDefFormat(virBufferPtr buf,
for (i = 0; i < def->maxvcpus; i++) {
virDomainSchedulerFormat(&childrenBuf, "vcpu",
- &def->vcpus[i]->sched, i);
+ &def->vcpus[i]->sched, i, true);
}
for (i = 0; i < def->niothreadids; i++) {
virDomainSchedulerFormat(&childrenBuf, "iothread",
&def->iothreadids[i]->sched,
- def->iothreadids[i]->iothread_id);
+ def->iothreadids[i]->iothread_id,
+ true);
}
for (i = 0; i < def->nresctrls; i++)
--
2.21.0
8:58 a.m.
New subject: [libvirt] [PATCH v2 2/6] conf: Format thread IDs optionally
On Mon, Apr 15, 2019 at 03:14:23PM +0200, Martin Kletzander wrote:
This will be used later when we want to format emulator scheduler
parameters
which don't apply for multiple threads.
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
squash! conf: Do not format negative thread IDs
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
Rebase artifacts.
---
src/conf/domain_conf.c | 22 +++++++++++++---------
1 file changed, 13 insertions(+), 9 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 15838c2a23f5..e7b8b51aad0e 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -27205,22 +27205,25 @@ static void
virDomainSchedulerFormat(virBufferPtr buf,
const char *name,
virDomainThreadSchedParamPtr sched,
- size_t id)
+ size_t id,
+ bool multiple_threads)
{
switch (sched->policy) {
case VIR_PROC_POLICY_BATCH:
case VIR_PROC_POLICY_IDLE:
- virBufferAsprintf(buf, "<%ssched "
- "%ss='%zu'
scheduler='%s'/>\n",
- name, name, id,
+ virBufferAsprintf(buf, "<%ssched", name);
+ if (multiple_threads)
+ virBufferAsprintf(buf, " %ss='%zu'", name, id);
+ virBufferAsprintf(buf, " scheduler='%s'/>\n",
virProcessSchedPolicyTypeToString(sched->policy));
break;
case VIR_PROC_POLICY_RR:
case VIR_PROC_POLICY_FIFO:
- virBufferAsprintf(buf, "<%ssched "
- "%ss='%zu' scheduler='%s'
priority='%d'/>\n",
- name, name, id,
+ virBufferAsprintf(buf, "<%ssched", name);
+ if (multiple_threads)
+ virBufferAsprintf(buf, " %ss='%zu'", name, id);
+ virBufferAsprintf(buf, " scheduler='%s'
priority='%d'/>\n",
virProcessSchedPolicyTypeToString(sched->policy),
sched->priority);
This function would look nicer with a temporary attrBuf and a virXMLFormatElement
at the end.
Jano
break;
9:16 a.m.
New subject: [libvirt] [PATCH v2 2/6] conf: Format thread IDs optionally
On Mon, Apr 15, 2019 at 03:58:05PM +0200, Ján Tomko wrote:
On Mon, Apr 15, 2019 at 03:14:23PM +0200, Martin Kletzander wrote:
>This will be used later when we want to format emulator scheduler parameters
>which don't apply for multiple threads.
>
>Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
>
>squash! conf: Do not format negative thread IDs
>
>Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
Rebase artifacts.
Yep, thanks, fixed locally.
>---
> src/conf/domain_conf.c | 22 +++++++++++++---------
> 1 file changed, 13 insertions(+), 9 deletions(-)
>
>diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
>index 15838c2a23f5..e7b8b51aad0e 100644
>--- a/src/conf/domain_conf.c
>+++ b/src/conf/domain_conf.c
>@@ -27205,22 +27205,25 @@ static void
> virDomainSchedulerFormat(virBufferPtr buf,
> const char *name,
> virDomainThreadSchedParamPtr sched,
>- size_t id)
>+ size_t id,
>+ bool multiple_threads)
> {
> switch (sched->policy) {
> case VIR_PROC_POLICY_BATCH:
> case VIR_PROC_POLICY_IDLE:
>- virBufferAsprintf(buf, "<%ssched "
>- "%ss='%zu'
scheduler='%s'/>\n",
>- name, name, id,
>+ virBufferAsprintf(buf, "<%ssched", name);
>+ if (multiple_threads)
>+ virBufferAsprintf(buf, " %ss='%zu'", name, id);
>+ virBufferAsprintf(buf, " scheduler='%s'/>\n",
> virProcessSchedPolicyTypeToString(sched->policy));
> break;
>
> case VIR_PROC_POLICY_RR:
> case VIR_PROC_POLICY_FIFO:
>- virBufferAsprintf(buf, "<%ssched "
>- "%ss='%zu' scheduler='%s'
priority='%d'/>\n",
>- name, name, id,
>+ virBufferAsprintf(buf, "<%ssched", name);
>+ if (multiple_threads)
>+ virBufferAsprintf(buf, " %ss='%zu'", name, id);
>+ virBufferAsprintf(buf, " scheduler='%s'
priority='%d'/>\n",
> virProcessSchedPolicyTypeToString(sched->policy),
> sched->priority);
This function would look nicer with a temporary attrBuf and a virXMLFormatElement
at the end.
Would it? I don't really see how. If you do, please show me.
Jano
> break;
8:14 a.m.
New subject: [libvirt] [PATCH v2 3/6] docs: Mention iothreadsched element in the docs and reword
Just one missing occurrence of iothreadsched fixed plus some rewording for this
to make more sense for the readers.
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
docs/formatdomain.html.in | 18 ++++++++++--------
1 file changed, 10 insertions(+), 8 deletions(-)
diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in
index 0e3799061d93..af4b88c60984 100644
--- a/docs/formatdomain.html.in
+++ b/docs/formatdomain.html.in
@@ -937,14 +937,16 @@
<dt><code>vcpusched</code> and
<code>iothreadsched</code></dt>
<dd>
- The optional <code>vcpusched</code> elements specifies the scheduler
- type (values <code>batch</code>, <code>idle</code>,
<code>fifo</code>,
- <code>rr</code>) for particular vCPU/IOThread threads (based on
- <code>vcpus</code> and <code>iothreads</code>, leaving
out
- <code>vcpus</code>/<code>iothreads</code> sets the
default). Valid
- <code>vcpus</code> values start at 0 through one less than the
- number of vCPU's defined for the domain. Valid
<code>iothreads</code>
- values are described in the <code>iothreadids</code>
+ The optional <code>vcpusched</code> and
<code>iothreadsched</code>
+ elements specify the scheduler type
+ (values <code>batch</code>, <code>idle</code>,
<code>fifo</code>,
+ <code>rr</code>) for particular vCPU and IOThread threads
respecively.
+ The attributes <code>vcpus</code> and
<code>iothreads</code> select
+ which vCPUs/IOThreads this setting applies to, leaving them out sets the
+ default. Valid <code>vcpus</code> values start at 0 through one
less
+ than the number of vCPU's defined for the
+ domain. Valid <code>iothreads</code> values are described in
+ the <code>iothreadids</code>
<a
href="#elementsIOThreadsAllocation"><code>description</code></a>.
If no <code>iothreadids</code> are defined, then libvirt numbers
IOThreads from 1 to the number of <code>iothreads</code> available
--
2.21.0
8:14 a.m.
New subject: [libvirt] [PATCH v2 4/6] conf: Add support for emulatorsched
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
docs/formatdomain.html.in | 17 ++++++----
docs/schemas/domaincommon.rng | 5 +++
src/conf/domain_conf.c | 44 +++++++++++++++++++++++++
src/conf/domain_conf.h | 1 +
tests/genericxml2xmlindata/cputune.xml | 37 +++++++++++++++++++++
tests/genericxml2xmloutdata/cputune.xml | 40 ++++++++++++++++++++++
tests/genericxml2xmltest.c | 2 ++
7 files changed, 140 insertions(+), 6 deletions(-)
create mode 100644 tests/genericxml2xmlindata/cputune.xml
create mode 100644 tests/genericxml2xmloutdata/cputune.xml
diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in
index af4b88c60984..74d32e2a9b8c 100644
--- a/docs/formatdomain.html.in
+++ b/docs/formatdomain.html.in
@@ -935,15 +935,19 @@
<span class="since">Only QEMU driver support since
2.1.0</span>
</dd>
- <dt><code>vcpusched</code> and
<code>iothreadsched</code></dt>
+ <dt><code>vcpusched</code>,
<code>iothreadsched</code>
+ and <code>emulatorsched</code></dt>
<dd>
- The optional <code>vcpusched</code> and
<code>iothreadsched</code>
- elements specify the scheduler type
+ The optional
+ <code>vcpusched</code>, <code>iothreadsched</code>
+ and <code>emulatorsched</code> elements specify the scheduler type
(values <code>batch</code>, <code>idle</code>,
<code>fifo</code>,
- <code>rr</code>) for particular vCPU and IOThread threads
respecively.
- The attributes <code>vcpus</code> and
<code>iothreads</code> select
+ <code>rr</code>) for particular vCPU, IOThread and emulator threads
+ respecively. For <code>vcpusched</code> and
<code>iothreadsched</code>
+ the attributes <code>vcpus</code> and
<code>iothreads</code> select
which vCPUs/IOThreads this setting applies to, leaving them out sets the
- default. Valid <code>vcpus</code> values start at 0 through one
less
+ default. The element <code>emulatorsched</code> does not have that
+ attribute. Valid <code>vcpus</code> values start at 0 through one
less
than the number of vCPU's defined for the
domain. Valid <code>iothreads</code> values are described in
the <code>iothreadids</code>
@@ -955,6 +959,7 @@
well (and is ignored for non-real-time ones). The value range
for the priority depends on the host kernel (usually 1-99).
<span class="since">Since 1.2.13</span>
+ <code>emulatorsched</code> <span class="since">since
5.3.0</span>
</dd>
<dt><code>cachetune</code><span
class="since">Since 4.1.0</span></dt>
diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
index 623ef28719df..111b85c36fd3 100644
--- a/docs/schemas/domaincommon.rng
+++ b/docs/schemas/domaincommon.rng
@@ -967,6 +967,11 @@
<ref name="schedparam"/>
</element>
</zeroOrMore>
+ <optional>
+ <element name="emulatorsched">
+ <ref name="schedparam"/>
+ </element>
+ </optional>
<zeroOrMore>
<element name="cachetune">
<attribute name="vcpus">
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index e7b8b51aad0e..17e8975be399 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -3396,6 +3396,7 @@ void virDomainDefFree(virDomainDefPtr def)
virDomainIOThreadIDDefArrayFree(def->iothreadids, def->niothreadids);
virBitmapFree(def->cputune.emulatorpin);
+ VIR_FREE(def->cputune.emulatorsched);
virDomainNumaFree(def->numa);
@@ -18483,6 +18484,25 @@ virDomainSchedulerParseCommonAttrs(xmlNodePtr node,
}
+static int
+virDomainEmulatorSchedParse(xmlNodePtr node,
+ virDomainDefPtr def)
+{
+ VIR_AUTOFREE(virDomainThreadSchedParamPtr) sched = NULL;
+
+ if (VIR_ALLOC(sched) < 0)
+ return -1;
+
+ if (virDomainSchedulerParseCommonAttrs(node,
+ &sched->policy,
+ &sched->priority) < 0)
+ return -1;
+
+ VIR_STEAL_PTR(def->cputune.emulatorsched, sched);
+ return 0;
+}
+
+
static virBitmapPtr
virDomainSchedulerParse(xmlNodePtr node,
const char *name,
@@ -19913,6 +19933,25 @@ virDomainDefParseXML(xmlDocPtr xml,
}
VIR_FREE(nodes);
+ if ((n = virXPathNodeSet("./cputune/emulatorsched", ctxt, &nodes)) <
0) {
+ virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("cannot extract emulatorsched nodes"));
+ goto error;
+ }
+
+ if (n) {
+ if (n > 1) {
+ virReportError(VIR_ERR_XML_ERROR, "%s",
+ _("only one emulatorsched is supported"));
+ VIR_FREE(nodes);
+ goto error;
+ }
+
+ if (virDomainEmulatorSchedParse(nodes[0], def) < 0)
+ goto error;
+ }
+ VIR_FREE(nodes);
+
if ((n = virXPathNodeSet("./cputune/cachetune", ctxt, &nodes)) < 0)
{
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("cannot extract cachetune nodes"));
@@ -27490,6 +27529,11 @@ virDomainCputuneDefFormat(virBufferPtr buf,
VIR_FREE(cpumask);
}
+ if (def->cputune.emulatorsched) {
+ virDomainSchedulerFormat(&childrenBuf, "emulator",
+ def->cputune.emulatorsched, 0, false);
+ }
+
for (i = 0; i < def->maxvcpus; i++) {
virDomainSchedulerFormat(&childrenBuf, "vcpu",
&def->vcpus[i]->sched, i, true);
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index 12eb71c1977e..988ef90f1189 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -2159,6 +2159,7 @@ struct _virDomainCputune {
unsigned long long iothread_period;
long long iothread_quota;
virBitmapPtr emulatorpin;
+ virDomainThreadSchedParamPtr emulatorsched;
};
diff --git a/tests/genericxml2xmlindata/cputune.xml
b/tests/genericxml2xmlindata/cputune.xml
new file mode 100644
index 000000000000..999c8cef73cd
--- /dev/null
+++ b/tests/genericxml2xmlindata/cputune.xml
@@ -0,0 +1,37 @@
+<domain type='qemu'>
+ <name>QEMUGuest1</name>
+ <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid>
+ <memory unit='KiB'>219100</memory>
+ <currentMemory unit='KiB'>219100</currentMemory>
+ <vcpu placement='static'>4</vcpu>
+ <iothreads>4</iothreads>
+ <cputune>
+ <shares>2048</shares>
+ <period>1000000</period>
+ <quota>-1</quota>
+ <vcpupin vcpu='0' cpuset='0'/>
+ <vcpupin vcpu='1' cpuset='1'/>
+ <emulatorpin cpuset='1'/>
+ <emulatorsched scheduler='fifo' priority='2'/>
+ <vcpusched vcpus='0-1' scheduler='fifo' priority='1'/>
+ <vcpusched vcpus='1-3,^1' scheduler='idle'/>
+ <iothreadsched iothreads='1,3' scheduler='batch'/>
+ <iothreadsched iothreads='2' scheduler='rr'
priority='99'/>
+ </cputune>
+ <os>
+ <type arch='x86_64' machine='q35'>hvm</type>
+ <boot dev='hd'/>
+ </os>
+ <features>
+ <smm state='on'>
+ <tseg unit='MiB'>48</tseg>
+ </smm>
+ </features>
+ <clock offset='utc'/>
+ <on_poweroff>destroy</on_poweroff>
+ <on_reboot>restart</on_reboot>
+ <on_crash>destroy</on_crash>
+ <devices>
+ <emulator>/usr/bin/qemu-system-x86_64</emulator>
+ </devices>
+</domain>
diff --git a/tests/genericxml2xmloutdata/cputune.xml
b/tests/genericxml2xmloutdata/cputune.xml
new file mode 100644
index 000000000000..f9f99a4846e6
--- /dev/null
+++ b/tests/genericxml2xmloutdata/cputune.xml
@@ -0,0 +1,40 @@
+<domain type='qemu'>
+ <name>QEMUGuest1</name>
+ <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid>
+ <memory unit='KiB'>219100</memory>
+ <currentMemory unit='KiB'>219100</currentMemory>
+ <vcpu placement='static'>4</vcpu>
+ <iothreads>4</iothreads>
+ <cputune>
+ <shares>2048</shares>
+ <period>1000000</period>
+ <quota>-1</quota>
+ <vcpupin vcpu='0' cpuset='0'/>
+ <vcpupin vcpu='1' cpuset='1'/>
+ <emulatorpin cpuset='1'/>
+ <emulatorsched scheduler='fifo' priority='2'/>
+ <vcpusched vcpus='0' scheduler='fifo' priority='1'/>
+ <vcpusched vcpus='1' scheduler='fifo' priority='1'/>
+ <vcpusched vcpus='2' scheduler='idle'/>
+ <vcpusched vcpus='3' scheduler='idle'/>
+ <iothreadsched iothreads='1' scheduler='batch'/>
+ <iothreadsched iothreads='2' scheduler='rr'
priority='99'/>
+ <iothreadsched iothreads='3' scheduler='batch'/>
+ </cputune>
+ <os>
+ <type arch='x86_64' machine='q35'>hvm</type>
+ <boot dev='hd'/>
+ </os>
+ <features>
+ <smm state='on'>
+ <tseg unit='MiB'>48</tseg>
+ </smm>
+ </features>
+ <clock offset='utc'/>
+ <on_poweroff>destroy</on_poweroff>
+ <on_reboot>restart</on_reboot>
+ <on_crash>destroy</on_crash>
+ <devices>
+ <emulator>/usr/bin/qemu-system-x86_64</emulator>
+ </devices>
+</domain>
diff --git a/tests/genericxml2xmltest.c b/tests/genericxml2xmltest.c
index 4393d444641d..1840d6e6a8f9 100644
--- a/tests/genericxml2xmltest.c
+++ b/tests/genericxml2xmltest.c
@@ -149,6 +149,8 @@ mymain(void)
DO_TEST("launch-security-sev");
+ DO_TEST_DIFFERENT("cputune");
+
virObjectUnref(caps);
virObjectUnref(xmlopt);
--
2.21.0
8:14 a.m.
New subject: [libvirt] [PATCH v2 5/6] qemu: Add support for emulatorsched
This helps in a scenarios where vCPUs run with a priority that is so high they
might starve the emulator thread. And it also fits with the rest of the
settings.
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
src/qemu/qemu_process.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index f773aa89b78d..55f4074ea146 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -2683,7 +2683,7 @@ qemuProcessSetupEmulator(virDomainObjPtr vm)
0, vm->def->cputune.emulatorpin,
vm->def->cputune.emulator_period,
vm->def->cputune.emulator_quota,
- NULL);
+ vm->def->cputune.emulatorsched);
}
--
2.21.0
8:14 a.m.
New subject: [libvirt] [PATCH v2 6/6] docs: Add emulatorsched support to news.xml
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
docs/news.xml | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/docs/news.xml b/docs/news.xml
index 86c77346940a..84de507b0e9a 100644
--- a/docs/news.xml
+++ b/docs/news.xml
@@ -35,6 +35,18 @@
<libvirt>
<release version="v5.3.0" date="unreleased">
<section title="New features">
+ <change>
+ <summary>
+ qemu: Add support for setting the emulator scheduler parameters
+ </summary>
+ <description>
+ I/O threads and vCPU threads already support setting schedulers, but
+ until now it was impossible to do so for the main QEMU thread
+ (emulator thread in the libvirt naming). This is, however, requested
+ for some very specific scenarios, for example when vCPU threads are
+ running at such priority that could starve the main thread.
+ </description>
+ </change>
</section>
<section title="Improvements">
<change>
--
2.21.0
8:59 a.m.
On Mon, Apr 15, 2019 at 03:14:21PM +0200, Martin Kletzander wrote:
I/O threads and vCPU threads already support setting schedulers, but
until now
it was impossible to do so for the main QEMU thread (emulator thread in the
libvirt naming). This is, however, requested for some very specific scenarios,
for example when vCPU threads are running at such priority that could starve the
main thread.
I don't like the XML format, but it's consistent with the existing
scheds.
v2:
- news.xml addition in a separate patch
Martin Kletzander (6):
conf: Parse common scheduler attributes in separate function
conf: Format thread IDs optionally
docs: Mention iothreadsched element in the docs and reword
conf: Add support for emulatorsched
qemu: Add support for emulatorsched
docs: Add emulatorsched support to news.xml
docs/formatdomain.html.in | 25 +++--
docs/news.xml | 12 +++
docs/schemas/domaincommon.rng | 5 +
src/conf/domain_conf.c | 136 +++++++++++++++++-------
src/conf/domain_conf.h | 1 +
src/qemu/qemu_process.c | 2 +-
tests/genericxml2xmlindata/cputune.xml | 37 +++++++
tests/genericxml2xmloutdata/cputune.xml | 40 +++++++
tests/genericxml2xmltest.c | 2 +
9 files changed, 212 insertions(+), 48 deletions(-)
create mode 100644 tests/genericxml2xmlindata/cputune.xml
create mode 100644 tests/genericxml2xmloutdata/cputune.xml
With at least the commit message in 2/6 fixed:
Reviewed-by: Ján Tomko <jtomko(a)redhat.com>
Jano
9:23 a.m.
On Mon, Apr 15, 2019 at 03:59:27PM +0200, Ján Tomko wrote:
On Mon, Apr 15, 2019 at 03:14:21PM +0200, Martin Kletzander wrote:
>I/O threads and vCPU threads already support setting schedulers, but until now
>it was impossible to do so for the main QEMU thread (emulator thread in the
>libvirt naming). This is, however, requested for some very specific scenarios,
>for example when vCPU threads are running at such priority that could starve the
>main thread.
>
I don't like the XML format,
me neither
but it's consistent with the existing scheds.
yes, unfortunately
>v2:
> - news.xml addition in a separate patch
>
>Martin Kletzander (6):
> conf: Parse common scheduler attributes in separate function
> conf: Format thread IDs optionally
> docs: Mention iothreadsched element in the docs and reword
> conf: Add support for emulatorsched
> qemu: Add support for emulatorsched
> docs: Add emulatorsched support to news.xml
>
> docs/formatdomain.html.in | 25 +++--
> docs/news.xml | 12 +++
> docs/schemas/domaincommon.rng | 5 +
> src/conf/domain_conf.c | 136 +++++++++++++++++-------
> src/conf/domain_conf.h | 1 +
> src/qemu/qemu_process.c | 2 +-
> tests/genericxml2xmlindata/cputune.xml | 37 +++++++
> tests/genericxml2xmloutdata/cputune.xml | 40 +++++++
> tests/genericxml2xmltest.c | 2 +
> 9 files changed, 212 insertions(+), 48 deletions(-)
> create mode 100644 tests/genericxml2xmlindata/cputune.xml
> create mode 100644 tests/genericxml2xmloutdata/cputune.xml
>
With at least the commit message in 2/6 fixed:
Reviewed-by: Ján Tomko <jtomko(a)redhat.com>
I'll fix other things you mentioned and send a v3 if you want, just let me know.
Jano
9:59 a.m.
On Mon, Apr 15, 2019 at 04:23:18PM +0200, Martin Kletzander wrote:
On Mon, Apr 15, 2019 at 03:59:27PM +0200, Ján Tomko wrote:
>On Mon, Apr 15, 2019 at 03:14:21PM +0200, Martin Kletzander wrote:
>>I/O threads and vCPU threads already support setting schedulers, but until now
>>it was impossible to do so for the main QEMU thread (emulator thread in the
>>libvirt naming). This is, however, requested for some very specific scenarios,
>>for example when vCPU threads are running at such priority that could starve the
>>main thread.
>>
>
>I don't like the XML format,
me neither
>but it's consistent with the existing scheds.
>
yes, unfortunately
>>v2:
>>- news.xml addition in a separate patch
>>
>>Martin Kletzander (6):
>> conf: Parse common scheduler attributes in separate function
>> conf: Format thread IDs optionally
>> docs: Mention iothreadsched element in the docs and reword
>> conf: Add support for emulatorsched
>> qemu: Add support for emulatorsched
>> docs: Add emulatorsched support to news.xml
>>
>>docs/formatdomain.html.in | 25 +++--
>>docs/news.xml | 12 +++
>>docs/schemas/domaincommon.rng | 5 +
>>src/conf/domain_conf.c | 136 +++++++++++++++++-------
>>src/conf/domain_conf.h | 1 +
>>src/qemu/qemu_process.c | 2 +-
>>tests/genericxml2xmlindata/cputune.xml | 37 +++++++
>>tests/genericxml2xmloutdata/cputune.xml | 40 +++++++
>>tests/genericxml2xmltest.c | 2 +
>>9 files changed, 212 insertions(+), 48 deletions(-)
>>create mode 100644 tests/genericxml2xmlindata/cputune.xml
>>create mode 100644 tests/genericxml2xmloutdata/cputune.xml
>>
>
>With at least the commit message in 2/6 fixed:
>Reviewed-by: Ján Tomko <jtomko(a)redhat.com>
>
I'll fix other things you mentioned and send a v3 if you want, just let me know.
Depending on your opinion on the refactor:
https://www.redhat.com/archives/libvir-list/2019-April/msg00996.html
Jano
5:59 a.m.
On Mon, Apr 15, 2019 at 04:59:30PM +0200, Ján Tomko wrote:
On Mon, Apr 15, 2019 at 04:23:18PM +0200, Martin Kletzander wrote:
>
>I'll fix other things you mentioned and send a v3 if you want, just let me know.
>
Depending on your opinion on the refactor:
https://www.redhat.com/archives/libvir-list/2019-April/msg00996.html
Given Andrea's comments on that series I withdraw my virXMLFormatElement
suggestion.
Jano
6:24 a.m.
On Tue, Apr 16, 2019 at 12:59:34PM +0200, Ján Tomko wrote:
On Mon, Apr 15, 2019 at 04:59:30PM +0200, Ján Tomko wrote:
>On Mon, Apr 15, 2019 at 04:23:18PM +0200, Martin Kletzander wrote:
>>
>>I'll fix other things you mentioned and send a v3 if you want, just let me
know.
>>
>
>Depending on your opinion on the refactor:
>https://www.redhat.com/archives/libvir-list/2019-April/msg00996.html
>
Given Andrea's comments on that series I withdraw my virXMLFormatElement
suggestion.
Based on the fact virDomainSchedulerParse() callers utilize the return value
(which would not be used in the new one) nicely, I'll leave that function as is
in this series so that the refactor discussion takes place in a separate thread
(if anyone cares enough to do that). I think this way it actually is better as
one function serves one purpose and not two slightly different ones based on
variable parameter.
Thanks.
Jano
2050
days inactive
2051
days old
16 comments
2 participants
participants (2)
-
Ján Tomko -
Martin Kletzander