Re: [libvirt] [PATCH 2/3] docs, conf, schema: add support for shared memory mapping
On Wed, Sep 10, 2014 at 04:46:27PM +0100, Daniel P. Berrange wrote: On Wed, Sep 10, 2014 at 05:36:58PM +0200, Ján Tomko wrote: On 09/08/2014 01:40 PM, Martin Kletzander wrote: Signed-off-by: Martin Kletzander mklet...@redhat.com --- docs/formatdomain.html.in | 7 +++- docs/schemas/domaincommon.rng | 5 +++ src/conf/cpu_conf.c| 25 +++- src/conf/cpu_conf.h| 7 ++-- .../qemuxml2argv-cpu-numa-memshared.xml| 28 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++ tests/qemuxml2xmltest.c| 2 + 7 files changed, 113 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 94236dd..b284d6e 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1105,7 +1105,7 @@ ... lt;numagt; lt;cell id='0' cpus='0-3' memory='512000'/gt; - lt;cell id='1' cpus='4-7' memory='512000'/gt; + lt;cell id='1' cpus='4-7' memory='512000' memShared='on'/gt; I wonder if shared='on' would be enough, avoiding the need for a multi-word attribute. Or how about access=shared|private ? I prepended the mem so that it is visible that it has something to do with the memory, not the whole node. But I'm OK with pushing shared= as well. Using access= seems too ambiguously worded to me, although if most of you agree... Martin signature.asc Description: Digital signature -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Re: [libvirt] [PATCH 2/3] docs, conf, schema: add support for shared memory mapping
On Mon, Sep 15, 2014 at 09:47:45AM +0200, Martin Kletzander wrote: On Wed, Sep 10, 2014 at 04:46:27PM +0100, Daniel P. Berrange wrote: On Wed, Sep 10, 2014 at 05:36:58PM +0200, Ján Tomko wrote: On 09/08/2014 01:40 PM, Martin Kletzander wrote: Signed-off-by: Martin Kletzander mklet...@redhat.com --- docs/formatdomain.html.in | 7 +++- docs/schemas/domaincommon.rng | 5 +++ src/conf/cpu_conf.c| 25 +++- src/conf/cpu_conf.h| 7 ++-- .../qemuxml2argv-cpu-numa-memshared.xml| 28 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++ tests/qemuxml2xmltest.c| 2 + 7 files changed, 113 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 94236dd..b284d6e 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1105,7 +1105,7 @@ ... lt;numagt; lt;cell id='0' cpus='0-3' memory='512000'/gt; - lt;cell id='1' cpus='4-7' memory='512000'/gt; + lt;cell id='1' cpus='4-7' memory='512000' memShared='on'/gt; I wonder if shared='on' would be enough, avoiding the need for a multi-word attribute. Or how about access=shared|private ? I prepended the mem so that it is visible that it has something to do with the memory, not the whole node. But I'm OK with pushing shared= as well. Using access= seems too ambiguously worded to me, although if most of you agree... Sure, memAccess is fine with me. 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 :| -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Re: [libvirt] [PATCH 2/3] docs, conf, schema: add support for shared memory mapping
On Mon, Sep 15, 2014 at 10:20:01AM +0100, Daniel P. Berrange wrote: On Mon, Sep 15, 2014 at 09:47:45AM +0200, Martin Kletzander wrote: On Wed, Sep 10, 2014 at 04:46:27PM +0100, Daniel P. Berrange wrote: On Wed, Sep 10, 2014 at 05:36:58PM +0200, Ján Tomko wrote: On 09/08/2014 01:40 PM, Martin Kletzander wrote: Signed-off-by: Martin Kletzander mklet...@redhat.com --- docs/formatdomain.html.in | 7 +++- docs/schemas/domaincommon.rng | 5 +++ src/conf/cpu_conf.c| 25 +++- src/conf/cpu_conf.h| 7 ++-- .../qemuxml2argv-cpu-numa-memshared.xml| 28 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++ tests/qemuxml2xmltest.c| 2 + 7 files changed, 113 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 94236dd..b284d6e 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1105,7 +1105,7 @@ ... lt;numagt; lt;cell id='0' cpus='0-3' memory='512000'/gt; - lt;cell id='1' cpus='4-7' memory='512000'/gt; + lt;cell id='1' cpus='4-7' memory='512000' memShared='on'/gt; I wonder if shared='on' would be enough, avoiding the need for a multi-word attribute. Or how about access=shared|private ? I prepended the mem so that it is visible that it has something to do with the memory, not the whole node. But I'm OK with pushing shared= as well. Using access= seems too ambiguously worded to me, although if most of you agree... Sure, memAccess is fine with me. Is there any possibility of that option having another value (in the future)? Otherwise shared= seems more appropriate to me. Let's see what others think, so I can finally get rid of this problem :) Martin signature.asc Description: Digital signature -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Re: [libvirt] [PATCH 2/3] docs, conf, schema: add support for shared memory mapping
On Mon, Sep 15, 2014 at 01:19:25PM +0200, Martin Kletzander wrote: On Mon, Sep 15, 2014 at 10:20:01AM +0100, Daniel P. Berrange wrote: On Mon, Sep 15, 2014 at 09:47:45AM +0200, Martin Kletzander wrote: On Wed, Sep 10, 2014 at 04:46:27PM +0100, Daniel P. Berrange wrote: On Wed, Sep 10, 2014 at 05:36:58PM +0200, Ján Tomko wrote: On 09/08/2014 01:40 PM, Martin Kletzander wrote: Signed-off-by: Martin Kletzander mklet...@redhat.com --- docs/formatdomain.html.in | 7 +++- docs/schemas/domaincommon.rng | 5 +++ src/conf/cpu_conf.c| 25 +++- src/conf/cpu_conf.h| 7 ++-- .../qemuxml2argv-cpu-numa-memshared.xml| 28 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++ tests/qemuxml2xmltest.c| 2 + 7 files changed, 113 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 94236dd..b284d6e 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1105,7 +1105,7 @@ ... lt;numagt; lt;cell id='0' cpus='0-3' memory='512000'/gt; - lt;cell id='1' cpus='4-7' memory='512000'/gt; + lt;cell id='1' cpus='4-7' memory='512000' memShared='on'/gt; I wonder if shared='on' would be enough, avoiding the need for a multi-word attribute. Or how about access=shared|private ? I prepended the mem so that it is visible that it has something to do with the memory, not the whole node. But I'm OK with pushing shared= as well. Using access= seems too ambiguously worded to me, although if most of you agree... Sure, memAccess is fine with me. Is there any possibility of that option having another value (in the future)? Otherwise shared= seems more appropriate to me. Let's see what others think, so I can finally get rid of this problem :) I prefer the approach of having values reflect the usage, as 'shared' vs 'private' for the value is clearer than 'on' vs 'off' IMHO. 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 :| -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Re: [libvirt] [PATCH 2/3] docs, conf, schema: add support for shared memory mapping
On Mon, Sep 15, 2014 at 12:40:48PM +0100, Daniel P. Berrange wrote: On Mon, Sep 15, 2014 at 01:19:25PM +0200, Martin Kletzander wrote: On Mon, Sep 15, 2014 at 10:20:01AM +0100, Daniel P. Berrange wrote: On Mon, Sep 15, 2014 at 09:47:45AM +0200, Martin Kletzander wrote: On Wed, Sep 10, 2014 at 04:46:27PM +0100, Daniel P. Berrange wrote: On Wed, Sep 10, 2014 at 05:36:58PM +0200, Ján Tomko wrote: On 09/08/2014 01:40 PM, Martin Kletzander wrote: Signed-off-by: Martin Kletzander mklet...@redhat.com --- docs/formatdomain.html.in | 7 +++- docs/schemas/domaincommon.rng | 5 +++ src/conf/cpu_conf.c| 25 +++- src/conf/cpu_conf.h| 7 ++-- .../qemuxml2argv-cpu-numa-memshared.xml| 28 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++ tests/qemuxml2xmltest.c| 2 + 7 files changed, 113 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 94236dd..b284d6e 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1105,7 +1105,7 @@ ... lt;numagt; lt;cell id='0' cpus='0-3' memory='512000'/gt; - lt;cell id='1' cpus='4-7' memory='512000'/gt; + lt;cell id='1' cpus='4-7' memory='512000' memShared='on'/gt; I wonder if shared='on' would be enough, avoiding the need for a multi-word attribute. Or how about access=shared|private ? I prepended the mem so that it is visible that it has something to do with the memory, not the whole node. But I'm OK with pushing shared= as well. Using access= seems too ambiguously worded to me, although if most of you agree... Sure, memAccess is fine with me. Is there any possibility of that option having another value (in the future)? Otherwise shared= seems more appropriate to me. Let's see what others think, so I can finally get rid of this problem :) I prefer the approach of having values reflect the usage, as 'shared' vs 'private' for the value is clearer than 'on' vs 'off' IMHO. OK, it makes more sense when you put it like that. I'll send a v2 to make sure everyone agrees on all three patches. Martin signature.asc Description: Digital signature -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Re: [libvirt] [PATCH 2/3] docs, conf, schema: add support for shared memory mapping
On 09/08/2014 01:40 PM, Martin Kletzander wrote: Signed-off-by: Martin Kletzander mklet...@redhat.com --- docs/formatdomain.html.in | 7 +++- docs/schemas/domaincommon.rng | 5 +++ src/conf/cpu_conf.c| 25 +++- src/conf/cpu_conf.h| 7 ++-- .../qemuxml2argv-cpu-numa-memshared.xml| 28 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++ tests/qemuxml2xmltest.c| 2 + 7 files changed, 113 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 94236dd..b284d6e 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1105,7 +1105,7 @@ ... lt;numagt; lt;cell id='0' cpus='0-3' memory='512000'/gt; - lt;cell id='1' cpus='4-7' memory='512000'/gt; + lt;cell id='1' cpus='4-7' memory='512000' memShared='on'/gt; I wonder if shared='on' would be enough, avoiding the need for a multi-word attribute. lt;/numagt; ... lt;/cpugt; @@ -1122,6 +1122,11 @@ assigned codeid/codes in the increasing order starting from 0. Mixing cells with and without the codeid/code attribute is not recommended as it may result in unwanted behaviour. + + span class='since'Since 1.2.9/span the optional attribute + codememShared/code can control whether the memory is to be + mapped as shared or not (values on/off). This is valid only + for hugepages-backed memory. /p p @@ -491,7 +491,7 @@ virCPUDefParseXML(xmlNodePtr node, goto error; } -ret = virStrToLong_ui(memory, NULL, 10, def-cells[cur_cell].mem); +ret = virStrToLong_ui(memory, NULL, 10, def-cells[cur_cell].mem); if (ret == -1) { virReportError(VIR_ERR_XML_ERROR, %s, _(Invalid 'memory' attribute in NUMA cell)); @@ -93,9 +93,10 @@ struct _virCPUFeatureDef { typedef struct _virCellDef virCellDef; typedef virCellDef *virCellDefPtr; struct _virCellDef { - virBitmapPtr cpumask; /* CPUs that are part of this node */ - char *cpustr; /* CPUs stored in string form for dumpxml */ - unsigned int mem; /* Node memory in kB */ +virBitmapPtr cpumask;/* CPUs that are part of this node */ +char *cpustr;/* CPUs stored in string form for dumpxml */ The comments would look nicer aligned. +unsigned int mem;/* Node memory in kB */ +virTristateSwitch memShared; }; typedef struct _virCPUDef virCPUDef; Please push these whitespace cleanups separately. Jan signature.asc Description: OpenPGP digital signature -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Re: [libvirt] [PATCH 2/3] docs, conf, schema: add support for shared memory mapping
On Wed, Sep 10, 2014 at 05:36:58PM +0200, Ján Tomko wrote: On 09/08/2014 01:40 PM, Martin Kletzander wrote: Signed-off-by: Martin Kletzander mklet...@redhat.com --- docs/formatdomain.html.in | 7 +++- docs/schemas/domaincommon.rng | 5 +++ src/conf/cpu_conf.c| 25 +++- src/conf/cpu_conf.h| 7 ++-- .../qemuxml2argv-cpu-numa-memshared.xml| 28 ++ .../qemuxml2argv-hugepages-shared.xml | 45 ++ tests/qemuxml2xmltest.c| 2 + 7 files changed, 113 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-numa-memshared.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hugepages-shared.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 94236dd..b284d6e 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1105,7 +1105,7 @@ ... lt;numagt; lt;cell id='0' cpus='0-3' memory='512000'/gt; - lt;cell id='1' cpus='4-7' memory='512000'/gt; + lt;cell id='1' cpus='4-7' memory='512000' memShared='on'/gt; I wonder if shared='on' would be enough, avoiding the need for a multi-word attribute. Or how about access=shared|private ? 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 :| -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list