Xen devel - Dec 2010 - [PATCH 0 of 5] support of NUMA topology in xl

Patch 1 adds a libxl function to retireve topology information from hypervisor
Patch 2 adds support of -n option to xl info
Patch 3 adds possibility to specify complete nodes instead of cpus for cpupools
Patch 4 adds possibility to rename a cpupool (libxl and xl command)
Patch 5 adds a new xl command cpupool-numa-split to create one cpupool per numa
node

9 files changed, 536 insertions(+), 21 deletions(-)
tools/libxl/libxl.c               |  167 ++++++++++++++++++++
tools/libxl/libxl.h               |   11 +
tools/libxl/libxl.idl             |    7 
tools/libxl/libxl_utils.c         |   28 +++
tools/libxl/libxl_utils.h         |    2 
tools/libxl/xl.h                  |    2 
tools/libxl/xl_cmdimpl.c          |  300 +++++++++++++++++++++++++++++++++++--
tools/libxl/xl_cmdtable.c         |   16 +
tools/python/xen/lowlevel/xl/xl.c |   24 ++

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Added new function libxl_get_topologyinfo() to obtain this information from
hypervisor.

Signed-off-by: juergen.gross@ts.fujitsu.com


6 files changed, 118 insertions(+)
tools/libxl/libxl.c               |   53 +++++++++++++++++++++++++++++++++++++
tools/libxl/libxl.h               |    8 +++++
tools/libxl/libxl.idl             |    7 ++++
tools/libxl/libxl_utils.c         |   24 ++++++++++++++++
tools/libxl/libxl_utils.h         |    2 +
tools/python/xen/lowlevel/xl/xl.c |   24 ++++++++++++++++




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Adds option -n/--numa to xl info command to print topology information.
No numa information up to now, as I''ve no machine which will give this
info
via xm info (could be a bug in xm, however).

Signed-off-by: juergen.gross@ts.fujitsu.com


2 files changed, 50 insertions(+), 11 deletions(-)
tools/libxl/xl_cmdimpl.c  |   59 +++++++++++++++++++++++++++++++++++++--------
tools/libxl/xl_cmdtable.c |    2 -




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

The user interfaces for cpupools are extended to support numa machines:
- xl cpupool-create supports now specifying a node list instead of a cpu list.
  The new cpupool will be created with all free cpus of the specified numa
  nodes.
- xl cpupool-cpu-remove and xl cpupool-cpu-add can take a node number instead
  of a cpu number. Using ''node:1'' for the cpu parameter will,
depending on
  the operation, either remove all cpus of node 1 in the specified cpupool,
  or add all free cpus of node 1 to the cpupool.

libxl is extended with the following functions to support this feature:
int libxl_cpupool_cpuadd_node(libxl_ctx *ctx, uint32_t poolid, int node, int
*cpus)
int libxl_cpupool_cpuremove_node(libxl_ctx *ctx, uint32_t poolid, int node, int
*cpus)

Signed-off-by: juergen.gross@ts.fujitsu.com


4 files changed, 155 insertions(+), 8 deletions(-)
tools/libxl/libxl.c       |   74 ++++++++++++++++++++++++++++++++++++++++
tools/libxl/libxl.h       |    2 +
tools/libxl/xl_cmdimpl.c  |   83 +++++++++++++++++++++++++++++++++++++++++----
tools/libxl/xl_cmdtable.c |    4 +-




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Add a new library function libxl_cpupool_rename() and a new xl command
xl cpupool-rename to support renaming of cpupools.

Signed-off-by: juergen.gross@ts.fujitsu.com


6 files changed, 90 insertions(+), 2 deletions(-)
tools/libxl/libxl.c       |   40 ++++++++++++++++++++++++++++++++++++++++
tools/libxl/libxl.h       |    1 +
tools/libxl/libxl_utils.c |    4 ++--
tools/libxl/xl.h          |    1 +
tools/libxl/xl_cmdimpl.c  |   41 +++++++++++++++++++++++++++++++++++++++++
tools/libxl/xl_cmdtable.c |    5 +++++




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

New xl command cpupool-numa-split which will create one cpupool for each
numa node of the machine. Can be called only if no other cpupools than Pool 0
are defined. After creation the cpupools can be managed as usual.

Signed-off-by: juergen.gross@ts.fujitsu.com


3 files changed, 123 insertions(+)
tools/libxl/xl.h          |    1 
tools/libxl/xl_cmdimpl.c  |  117 +++++++++++++++++++++++++++++++++++++++++++++
tools/libxl/xl_cmdtable.c |    5 +




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Thu, 2010-12-09 at 10:51 +0000, Juergen Gross wrote:
> 
> diff -r b979d430eab7 -r fc3f33c7faa7 tools/libxl/libxl.c
> --- a/tools/libxl/libxl.c       Thu Dec 09 11:26:37 2010 +0100
> +++ b/tools/libxl/libxl.c       Thu Dec 09 11:29:25 2010 +0100
> @@ -3809,6 +3809,46 @@ out:
>      return rc;
>  }
>  
> +int libxl_cpupool_rename(libxl_ctx *ctx, const char *name, uint32_t
> poolid)
> +{
> +    libxl__gc gc = LIBXL_INIT_GC(ctx);
> +    xs_transaction_t t;
> +    xc_cpupoolinfo_t *info;
> +    int rc;
> +
> +    info = xc_cpupool_getinfo(ctx->xch, poolid);
> +    if (info == NULL)
> +        return ERROR_NOMEM;
> +
> +    rc = ERROR_INVAL;
> +    if (info->cpupool_id != poolid)
> +        goto out;
> +
> +    rc = 0;
> +
> +    for (;;) {
> +        t = xs_transaction_start(ctx->xsh);
> +
> +        libxl__xs_write(&gc, t,
> +                        libxl__sprintf(&gc,
"/local/pool/%d/name",
> poolid),
> +                        "%s", name);
> +
> +        if (xs_transaction_end(ctx->xsh, t, 0))
> +            break;
> +
> +        if (errno == EAGAIN)
> +            continue;
> +
> +        rc = ERROR_FAIL;
> +        break;
> +    }
> +
> +out:
> +    xc_cpupool_infofree(ctx->xch, info);
You need:
	libxl__free_all(&gc);



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Other than my one comment to 4/5 and Gianni''s python issue the series
looks good to me so assuming those are fixed all:

Acked-by: Ian Campbell <ian.campbell@citrix.com>

Thanks!

On Thu, 2010-12-09 at 10:51 +0000, Juergen Gross wrote:
> Patch 1 adds a libxl function to retireve topology information from
hypervisor
> Patch 2 adds support of -n option to xl info
> Patch 3 adds possibility to specify complete nodes instead of cpus for
cpupools
> Patch 4 adds possibility to rename a cpupool (libxl and xl command)
> Patch 5 adds a new xl command cpupool-numa-split to create one cpupool per
numa node
> 
> 9 files changed, 536 insertions(+), 21 deletions(-)
> tools/libxl/libxl.c               |  167 ++++++++++++++++++++
> tools/libxl/libxl.h               |   11 +
> tools/libxl/libxl.idl             |    7 
> tools/libxl/libxl_utils.c         |   28 +++
> tools/libxl/libxl_utils.h         |    2 
> tools/libxl/xl.h                  |    2 
> tools/libxl/xl_cmdimpl.c          |  300
+++++++++++++++++++++++++++++++++++--
> tools/libxl/xl_cmdtable.c         |   16 +
> tools/python/xen/lowlevel/xl/xl.c |   24 ++
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

This is V3 of my patch series to support NUMA in xl with cpupools.

Patch 1 adds a libxl function to retireve topology information from hypervisor
Patch 2 adds support of -n option to xl info
Patch 3 adds possibility to specify complete nodes instead of cpus for cpupools
Patch 4 adds possibility to rename a cpupool (libxl and xl command)
Patch 5 adds a new xl command cpupool-numa-split to create one cpupool per numa
node

Changes since last version:
- added Py_INCREF when appending Py_None to list
- added libxl__free_all(&gc) in libxl_cpupool_rename()

9 files changed, 540 insertions(+), 21 deletions(-)
tools/libxl/libxl.c               |  170 ++++++++++++++++++++
tools/libxl/libxl.h               |   11 +
tools/libxl/libxl.idl             |    7 
tools/libxl/libxl_utils.c         |   28 +++
tools/libxl/libxl_utils.h         |    2 
tools/libxl/xl.h                  |    2 
tools/libxl/xl_cmdimpl.c          |  300 +++++++++++++++++++++++++++++++++++--
tools/libxl/xl_cmdtable.c         |   16 +
tools/python/xen/lowlevel/xl/xl.c |   25 +++

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Added new function libxl_get_topologyinfo() to obtain this information from
hypervisor.

Signed-off-by: juergen.gross@ts.fujitsu.com


6 files changed, 119 insertions(+)
tools/libxl/libxl.c               |   53 +++++++++++++++++++++++++++++++++++++
tools/libxl/libxl.h               |    8 +++++
tools/libxl/libxl.idl             |    7 ++++
tools/libxl/libxl_utils.c         |   24 ++++++++++++++++
tools/libxl/libxl_utils.h         |    2 +
tools/python/xen/lowlevel/xl/xl.c |   25 +++++++++++++++++




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Adds option -n/--numa to xl info command to print topology information.
No numa information up to now, as I''ve no machine which will give this
info
via xm info (could be a bug in xm, however).

Signed-off-by: juergen.gross@ts.fujitsu.com


2 files changed, 50 insertions(+), 11 deletions(-)
tools/libxl/xl_cmdimpl.c  |   59 +++++++++++++++++++++++++++++++++++++--------
tools/libxl/xl_cmdtable.c |    2 -




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

The user interfaces for cpupools are extended to support numa machines:
- xl cpupool-create supports now specifying a node list instead of a cpu list.
  The new cpupool will be created with all free cpus of the specified numa
  nodes.
- xl cpupool-cpu-remove and xl cpupool-cpu-add can take a node number instead
  of a cpu number. Using ''node:1'' for the cpu parameter will,
depending on
  the operation, either remove all cpus of node 1 in the specified cpupool,
  or add all free cpus of node 1 to the cpupool.

libxl is extended with the following functions to support this feature:
int libxl_cpupool_cpuadd_node(libxl_ctx *ctx, uint32_t poolid, int node, int
*cpus)
int libxl_cpupool_cpuremove_node(libxl_ctx *ctx, uint32_t poolid, int node, int
*cpus)

Signed-off-by: juergen.gross@ts.fujitsu.com


4 files changed, 155 insertions(+), 8 deletions(-)
tools/libxl/libxl.c       |   74 ++++++++++++++++++++++++++++++++++++++++
tools/libxl/libxl.h       |    2 +
tools/libxl/xl_cmdimpl.c  |   83 +++++++++++++++++++++++++++++++++++++++++----
tools/libxl/xl_cmdtable.c |    4 +-




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Add a new library function libxl_cpupool_rename() and a new xl command
xl cpupool-rename to support renaming of cpupools.

Signed-off-by: juergen.gross@ts.fujitsu.com


6 files changed, 93 insertions(+), 2 deletions(-)
tools/libxl/libxl.c       |   43 +++++++++++++++++++++++++++++++++++++++++++
tools/libxl/libxl.h       |    1 +
tools/libxl/libxl_utils.c |    4 ++--
tools/libxl/xl.h          |    1 +
tools/libxl/xl_cmdimpl.c  |   41 +++++++++++++++++++++++++++++++++++++++++
tools/libxl/xl_cmdtable.c |    5 +++++




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

New xl command cpupool-numa-split which will create one cpupool for each
numa node of the machine. Can be called only if no other cpupools than Pool 0
are defined. After creation the cpupools can be managed as usual.

Signed-off-by: juergen.gross@ts.fujitsu.com


3 files changed, 123 insertions(+)
tools/libxl/xl.h          |    1 
tools/libxl/xl_cmdimpl.c  |  117 +++++++++++++++++++++++++++++++++++++++++++++
tools/libxl/xl_cmdtable.c |    5 +




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Thu, 2010-12-09 at 12:39 +0000, Juergen Gross wrote:
> 
> +PyObject *attrib__libxl_cpuarray_get(libxl_cpuarray *pptr)
> +{
> +    PyObject *list = NULL;
> +    int i;
> +
> +    list = PyList_New(0);
> +    for (i = 0; i < pptr->entries; i++) {
> +        if (pptr->array[i] == LIBXL_CPUARRAY_INVALID_ENTRY) {
> +            PyList_Append(list, Py_None);
> +            Py_INCREF(Py_None); 

Actually I think I am the one who is confused. It is the other way
around. I need to fix other places that leak references due to this.

IOW, this was good to go first time around, tried to respond earlier but
my mail client barfed the message.

Sorry for the noise! :(

Gianni


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Ian,

could you take my previous version of this patch (1 of 5), or would you like
me to send another complete patch series?


Juergen

On 12/09/10 14:11, Gianni Tedesco wrote:
> On Thu, 2010-12-09 at 12:39 +0000, Juergen Gross wrote:
>>
>> +PyObject *attrib__libxl_cpuarray_get(libxl_cpuarray *pptr)
>> +{
>> +    PyObject *list = NULL;
>> +    int i;
>> +
>> +    list = PyList_New(0);
>> +    for (i = 0; i<  pptr->entries; i++) {
>> +        if (pptr->array[i] == LIBXL_CPUARRAY_INVALID_ENTRY) {
>> +            PyList_Append(list, Py_None);
>> +            Py_INCREF(Py_None);
>
>
> Actually I think I am the one who is confused. It is the other way
> around. I need to fix other places that leak references due to this.
>
> IOW, this was good to go first time around, tried to respond earlier but
> my mail client barfed the message.
>
> Sorry for the noise! :(
-- 
Juergen Gross                 Principal Developer Operating Systems
TSP ES&S SWE OS6                       Telephone: +49 (0) 89 3222 2967
Fujitsu Technology Solutions              e-mail: juergen.gross@ts.fujitsu.com
Domagkstr. 28                           Internet: ts.fujitsu.com
D-80807 Muenchen                 Company details: ts.fujitsu.com/imprint.html

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Juergen Gross writes ("Re: [Xen-devel] [PATCH 1 of 5] Support getting
topology info in libxl"):
> could you take my previous version of this patch (1 of 5), or would you
like
> me to send another complete patch series?
I''ll take the previous version, thanks.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Ian,

any reason you didn''t apply my series up to now?
Did I miss any objection?


Juergen

On 12/09/10 13:39, Juergen Gross wrote:
> This is V3 of my patch series to support NUMA in xl with cpupools.
>
> Patch 1 adds a libxl function to retireve topology information from
hypervisor
> Patch 2 adds support of -n option to xl info
> Patch 3 adds possibility to specify complete nodes instead of cpus for
cpupools
> Patch 4 adds possibility to rename a cpupool (libxl and xl command)
> Patch 5 adds a new xl command cpupool-numa-split to create one cpupool per
numa node
>
> Changes since last version:
> - added Py_INCREF when appending Py_None to list
> - added libxl__free_all(&gc) in libxl_cpupool_rename()
>
> 9 files changed, 540 insertions(+), 21 deletions(-)
> tools/libxl/libxl.c               |  170 ++++++++++++++++++++
> tools/libxl/libxl.h               |   11 +
> tools/libxl/libxl.idl             |    7
> tools/libxl/libxl_utils.c         |   28 +++
> tools/libxl/libxl_utils.h         |    2
> tools/libxl/xl.h                  |    2
> tools/libxl/xl_cmdimpl.c          |  300
+++++++++++++++++++++++++++++++++++--
> tools/libxl/xl_cmdtable.c         |   16 +
> tools/python/xen/lowlevel/xl/xl.c |   25 +++
-- 
Juergen Gross                 Principal Developer Operating Systems
TSP ES&S SWE OS6                       Telephone: +49 (0) 89 3222 2967
Fujitsu Technology Solutions              e-mail: juergen.gross@ts.fujitsu.com
Domagkstr. 28                           Internet: ts.fujitsu.com
D-80807 Muenchen                 Company details: ts.fujitsu.com/imprint.html

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Juergen Gross writes ("Re: [Xen-devel] [PATCH 0 of 5] support of NUMA
topology in xl"):
> any reason you didn''t apply my series up to now?
> Did I miss any objection?
No, there is no problem with your series.  However, our tests are
failing a lot and I wanted to wait until we have a test pass and a
push before dumping a lot more stuff into xen-unstable.

So I''m holding off on your series as well as a few others.

Keir, it would be nice if we could try to restrict stuff going into
xen-unstable to obvious bugfixes, for a bit.

I suggest that our rule for the end-of-December freeze should be that
patches which were posted in their final form or something very like
it by (say) Christmas Eve should go into 4.1, even if we have held off
actually committing them until early January.

Alternatively I can throw everything in now and try to fix it up later
but it does make it harder to pin down what went wrong when things
break.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On 17/12/2010 17:46, "Ian Jackson" <Ian.Jackson@eu.citrix.com>
wrote:
> Juergen Gross writes ("Re: [Xen-devel] [PATCH 0 of 5] support of NUMA
topology
> in xl"):
>> any reason you didn''t apply my series up to now?
>> Did I miss any objection?
> 
> No, there is no problem with your series.  However, our tests are
> failing a lot and I wanted to wait until we have a test pass and a
> push before dumping a lot more stuff into xen-unstable.
> 
> So I''m holding off on your series as well as a few others.
> 
> Keir, it would be nice if we could try to restrict stuff going into
> xen-unstable to obvious bugfixes, for a bit.
> 
> I suggest that our rule for the end-of-December freeze should be that
> patches which were posted in their final form or something very like
> it by (say) Christmas Eve should go into 4.1, even if we have held off
> actually committing them until early January.
> 
> Alternatively I can throw everything in now and try to fix it up later
> but it does make it harder to pin down what went wrong when things
> break.
That sounds okay. I have nothing in my queue to be applied now anyway.

 -- Keir



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

This patch fixes a compilation warning (uninitialized parameter before 
use) on libxl.c. Xen failed to compile without it. The warning was 
reported on gcc 4.1.2.

Signed-off-by: Wei Huang <wei.huang2@amd.com>




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Juergen Gross writes ("[Xen-devel] [PATCH 1 of 5] Support getting topology
info in libxl"):
> Added new function libxl_get_topologyinfo() to obtain this information from
> hypervisor.
Thanks, I''ve applied this version (Thu, 09 Dec 2010 10:56:35 GMT) of
patch 1, and the latest version (Thu, 09 Dec 2010 12:49:22 GMT etc.)
of 2-5.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Wei Huang writes ("[Xen-devel] [PATCH] xl: fix compilation warning on
libxl.c"):
> This patch fixes a compilation warning (uninitialized parameter before 
> use) on libxl.c. Xen failed to compile without it. The warning was 
> reported on gcc 4.1.2.
Zhang, Yang Z writes ("[Xen-devel] [PATCH] fix build error in
libxl"):
> Move free(disks) into the bracket, or it will show error
''disks'' used uninitialized when build libxl.
Thanks.  I have applied Yang Z Zhang''s patch.

Yang, since your patch was only one line I applied it anyway, but in
future could you please provide a Signed-Off-By, confirming the
copyright status according to the Developer''s Certificate of Origin ?

Thanks,
Ian.
>From Documentation/SubmittingPatches in the Linux kernel tree:
       Developer''s Certificate of Origin 1.1

       By making a contribution to this project, I certify that:

       (a) The contribution was created in whole or in part by me and I
           have the right to submit it under the open source license
           indicated in the file; or

       (b) The contribution is based upon previous work that, to the best
           of my knowledge, is covered under an appropriate open source
           license and I have the right under that license to submit that
           work with modifications, whether created in whole or in part
           by me, under the same open source license (unless I am
           permitted to submit under a different license), as indicated
           in the file; or

       (c) The contribution was provided directly to me by some other
           person who certified (a), (b) or (c) and I have not modified
           it.

       (d) I understand and agree that this project and the contribution
           are public and that a record of the contribution (including all
           personal information I submit with it, including my sign-off) is
           maintained indefinitely and may be redistributed consistent with
           this project or the open source license(s) involved.

--

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel