Xen devel - Jun 2012 - [PATCH 0 of 2 v3] Sanity checking of scheduling parameters

Hi,

This small series achieves two goals:
 - check the return value of libxl_domain_sched_params_set() in
   libxl__build_post() and deal with the error, if that is the
   case (patch #1);
 - check and ensue we are passing along a meaningful set of sedf
   scheduling parameters when they come directly from the config file
   (patch #2)

Tested on both credit and sedf schedulers.

Changes from v2:
 * actually propagate the error correctly instead of always returning
   INVAL (in patch #1).

Changes from v1:
 * patch #1: it was not there at all in v1! :-P
 * patch #2: the if-s have been moved into an helper function. Also,
   they only happen if the domain is actually being scheduled with
   sedf (IanC, yes, I decided to do it... At the end of the day, it is
   simple enough I think).


Thanks and Regards,
Dario

-- 
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://retis.sssup.it/people/faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

So that the caller (e.g., libxl__build_post() ) knows and can deal with it.

Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>

diff --git a/tools/libxl/libxl_dom.c b/tools/libxl/libxl_dom.c
--- a/tools/libxl/libxl_dom.c
+++ b/tools/libxl/libxl_dom.c
@@ -173,9 +173,11 @@ int libxl__build_post(libxl__gc *gc, uin
     char *dom_path, *vm_path;
     xs_transaction_t t;
     char **ents, **hvm_ents;
-    int i;
+    int i, rc;
 
-    libxl_domain_sched_params_set(CTX, domid, &info->sched_params);
+    rc = libxl_domain_sched_params_set(CTX, domid, &info->sched_params);
+    if (rc)
+        return rc;
 
     libxl_cpuid_apply_policy(ctx, domid);
     if (info->cpuid != NULL)

As it happens in the implementation of `xl sched-sedf -d ...'', some
consistency checking is needed for the scheduling parameters when
they come from the config file.

Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>

diff --git a/tools/libxl/xl_cmdimpl.c b/tools/libxl/xl_cmdimpl.c
--- a/tools/libxl/xl_cmdimpl.c
+++ b/tools/libxl/xl_cmdimpl.c
@@ -550,6 +550,31 @@ vcpp_out:
     return rc;
 }
 
+static int sched_params_valid(libxl_domain_sched_params *scp)
+{
+    int has_weight = scp->weight != LIBXL_DOMAIN_SCHED_PARAM_WEIGHT_DEFAULT;
+    int has_period = scp->period != LIBXL_DOMAIN_SCHED_PARAM_PERIOD_DEFAULT;
+    int has_slice = scp->slice != LIBXL_DOMAIN_SCHED_PARAM_SLICE_DEFAULT;
+    libxl_domain_sched_params sci;
+
+    libxl_domain_sched_params_get(ctx, domid, &sci);
+
+    /* The sedf scheduler needs some more consistency checking */
+    if (sci.sched == LIBXL_SCHEDULER_SEDF) {
+        if (has_weight && (has_period || has_slice))
+            return 0;
+
+        if (has_weight) {
+            scp->slice = 0;
+            scp->period = 0;
+        }
+        if (has_period || has_slice)
+            scp->weight = 0;
+    }
+
+    return 1;
+}
+
 static void parse_config_data(const char *config_source,
                               const char *config_data,
                               int config_len,
@@ -644,6 +669,10 @@ static void parse_config_data(const char
         b_info->sched_params.latency = l;
     if (!xlu_cfg_get_long (config, "extratime", &l, 0))
         b_info->sched_params.extratime = l;
+    if (!sched_params_valid(&b_info->sched_params)) {
+        fprintf(stderr, "Invalid scheduling parameters\n");
+        exit(1);
+    }
 
     if (!xlu_cfg_get_long (config, "vcpus", &l, 0)) {
         b_info->max_vcpus = l;

Dario Faggioli writes ("[PATCH 2 of 2 v3] xl: check for meaningful
combination of sedf config file parameters"):
> As it happens in the implementation of `xl sched-sedf -d ...'',
some
> consistency checking is needed for the scheduling parameters when
> they come from the config file.
Thanks, I am happy with this from a libxl point of view.  I would like
to see an ack from George from a scheduler point of view.

Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>

Thanks,
Ian.

Dario Faggioli writes ("[PATCH 1 of 2 v3] libxl: propagete down the error
from libxl_domain_sched_params_set"):
> So that the caller (e.g., libxl__build_post() ) knows and can deal with it.
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>

On 08/06/12 12:10, Ian Jackson wrote:
> Dario Faggioli writes ("[PATCH 2 of 2 v3] xl: check for meaningful
combination of sedf config file parameters"):
>> As it happens in the implementation of `xl sched-sedf -d ...'',
some
>> consistency checking is needed for the scheduling parameters when
>> they come from the config file.
> Thanks, I am happy with this from a libxl point of view.  I would like
> to see an ack from George from a scheduler point of view.
I think Dario knows more about what sedf wants than I do; and since it 
only does anything when the domain is using sedf, I think it''s fine.

So FWIW:

Acked-by: George Dunlap <george.dunlap@eu.citrix.com>

George Dunlap writes ("Re: [Xen-devel] [PATCH 2 of 2 v3] xl: check for
meaningful combination of sedf config file
parameters"):
> On 08/06/12 12:10, Ian Jackson wrote:
> > Dario Faggioli writes ("[PATCH 2 of 2 v3] xl: check for
meaningful combination of sedf config file parameters"):
> >> As it happens in the implementation of `xl sched-sedf -d
...'', some
> >> consistency checking is needed for the scheduling parameters when
> >> they come from the config file.
> > Thanks, I am happy with this from a libxl point of view.  I would like
> > to see an ack from George from a scheduler point of view.
> I think Dario knows more about what sedf wants than I do; and since it 
> only does anything when the domain is using sedf, I think it''s
fine.
> 
> So FWIW:
> 
> Acked-by: George Dunlap <george.dunlap@eu.citrix.com>
Great, thanks.

Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>

Ian.

On Fri, 2012-06-08 at 15:26 +0100, Ian Jackson wrote:
> George Dunlap writes ("Re: [Xen-devel] [PATCH 2 of 2 v3] xl: check for
meaningful combination of sedf config file parameters"):
> > On 08/06/12 12:10, Ian Jackson wrote:
> > > Dario Faggioli writes ("[PATCH 2 of 2 v3] xl: check for
meaningful combination of sedf config file parameters"):
> > >> As it happens in the implementation of `xl sched-sedf -d
...'', some
> > >> consistency checking is needed for the scheduling parameters
when
> > >> they come from the config file.
> > > Thanks, I am happy with this from a libxl point of view.  I would
like
> > > to see an ack from George from a scheduler point of view.
> > I think Dario knows more about what sedf wants than I do; and since it
> > only does anything when the domain is using sedf, I think
it''s fine.
> > 
> > So FWIW:
> > 
> > Acked-by: George Dunlap <george.dunlap@eu.citrix.com>
> 
> Great, thanks.
> 
> Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Was it deliberate that you took this one but not the other patch from
this series of two, namely "libxl: propagete down the error from
libxl_domain_sched_params_set"?

Ian.

Ian Campbell writes ("Re: [Xen-devel] [PATCH 2 of 2 v3] xl: check for
meaningful combination of sedf config file
parameters"):
> Was it deliberate that you took this one but not the other patch from
> this series of two, namely "libxl: propagete down the error from
> libxl_domain_sched_params_set"?
No, sorry.

Dario Faggioli writes ("[Xen-devel] [PATCH 1 of 2 v3] libxl: propagete down
the error from libxl_domain_sched_params_set"):
> So that the caller (e.g., libxl__build_post() ) knows and can deal with it.
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>

On Thu, 2012-06-14 at 16:06 +0100, Ian Jackson wrote:
> Ian Campbell writes ("Re: [Xen-devel] [PATCH 2 of 2 v3] xl: check for
meaningful combination of sedf config file parameters"):
> > Was it deliberate that you took this one but not the other patch from
> > this series of two, namely "libxl: propagete down the error from
> > libxl_domain_sched_params_set"?
> 
> No, sorry.
> 
> Dario Faggioli writes ("[Xen-devel] [PATCH 1 of 2 v3] libxl: propagete
down the error from libxl_domain_sched_params_set"):
> > So that the caller (e.g., libxl__build_post() ) knows and can deal
with it.
> 
> Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
>
Cool. Thanks,
Dario

-- 
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://retis.sssup.it/people/faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)



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