ize, if it were me doing this, I'd sync either in
vmx_vcpu_destroy() or in complete_domain_destroy(). But (for what it's
worth) I'm fine with it happening in rcu_do_batch().
Regards,
Dario
--
<> (Raistlin Majere)
ough, I guess it will have, at
least potentially, an impact on performance (as other solutions
envisioned in the thread). Any idea how big the hit could be?
Regards,
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D
On Thu, 2017-11-09 at 07:49 -0700, Jan Beulich wrote:
> See the code comment being added for why we need this.
>
> Reported-by: Igor Druzhinin
> Signed-off-by: Jan Beulich
>
Reviewed-by: Dario Faggioli
Regards,
Dario
--
<
look.
What about the effect on performance, though?
I mean, assuming that lazy context switch does a good job, with respect
to that, by avoiding synching in enough case where it is actually not
necessary, how do things change if we start to sync at any softirq,
even when the handler would have not re
think it is indeed more general, as in, without even the need to
consider a second pCPU:
pCPU1
=
current == vCPU1
context_switch(next == idle)
!! __context_switch() is skipped
vcpu_migrate(vCPU1)
anything_that_uses_or_touches_context()
So, it must be anything_that_uses_or_touches_context
ble.
>
> That's not what Igor's diagram says - it's indicating the fact that
> is_running is being set to 1 in the process of context switching
> into vCPUx.
>
Ah, ok. So I was right: I indeed was misunderstanding something, i.e.,
the diagram itself. :-)
Now I ge
ioned, and it was being considered to treat that in a special way.
But it looks to me that something very similar may, at least in theory,
happen any time a lazy context switch occurs, no matter whether the
pCPU has become idle because the previously running vCPU wants to move,
or because it blo
ecide to pick vCPUx from the
runqueue and switch to it _only_ if it has gone through vcpu_wake(),
which must actually have woken up it, which happens if _VPF_migrating
has been cleared, which means is_running was 0 already.
Dario
--
<> (Raistlin Majere)
--
As soft-affinity and caps will be available in Xen 4.10.
Signed-off-by: Dario Faggioli
---
Cc: George Dunlap
Cc: Julien Grall
Cc: Lars Kurth
---
Julien, doc change, so no risk.
Sorry I missed doing this at the same time of the changes themselves.
---
docs/features/sched_credit2.pandoc
On Wed, 2017-10-11 at 14:02 -0400, Meng Xu wrote:
> Change repl_budget event output for xentrace formats and xenalyze
>
> Signed-off-by: Meng Xu
>
I'd say:
Reviewed-by: Dario Faggioli
However...
> diff --git a/tools/xentrace/xenalyze.c b/tools/xentrace/xenalyze.c
>
On Tue, 2017-10-17 at 09:26 +0200, Dario Faggioli wrote:
> On Thu, 2017-10-12 at 10:34 -0400, Meng Xu wrote:
> > On Thu, Oct 12, 2017 at 5:02 AM, Wei Liu
> > wrote:
> > >
> > > FYI all patches except the xentrace one were committed yesterday.
> >
> >
nks and Regards,
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
signature.asc
Description: This is a digitally signed message part
___
Xen-dev
", budget = %"PRIu64"\n",
> + ri->dump_header, r->domid, r->vcpuid,
> + r->priority_level, r->cur_dl, r->cur_bg);
> }
> break;
> case TRC_SCHED_CLASS_
On Tue, 2017-10-10 at 19:17 -0400, Meng Xu wrote:
> Modify libxl_vcpu_sched_params_get/set and sched_rtds_vcpu_get/set
> functions to support per-VCPU extratime flag
>
> Signed-off-by: Meng Xu
>
Reviewed-by: Dario Faggioli
Regards,
Dario
--
<
ties.
> Depleted queue holds VCPUs which have extratime flag cleared and
> depleted budget.
> Replenished queue is not modified.
>
> Distribution of spare bandwidth
> Spare bandwidth is distributed among all VCPUs with extratime flag
> set,
> proportional to these VCP
On Tue, 2017-10-10 at 19:17 -0400, Meng Xu wrote:
> Revise xl tool use case by adding -e option
> Remove work-conserving from TODO list
>
> Signed-off-by: Meng Xu
>
Reviewed-by: Dario Faggioli
Regards,
Dario
--
<
On Tue, 2017-10-10 at 19:17 -0400, Meng Xu wrote:
> Change main_sched_rtds and related output functions to support
> per-VCPU extratime flag.
>
> Signed-off-by: Meng Xu
>
Reviewed-by: Dario Faggioli
Regards,
Dario
--
<
On Mon, 2017-10-09 at 12:13 -0400, Meng Xu wrote:
> On Wed, Sep 13, 2017 at 8:51 PM, Dario Faggioli
> wrote:
> >
> > On Fri, 2017-09-01 at 11:58 -0400, Meng Xu wrote:
> > > diff --git a/tools/xl/xl_cmdtable.c b/tools/xl/xl_cmdtable.c
> > > index ba0159d.
r sending v4.
>
Which, AFAICS, you've done, and checked in the series.
Thanks!
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
signature.asc
Descript
On Thu, 2017-10-05 at 13:19 -0400, Meng Xu wrote:
> On Thu, Oct 5, 2017 at 10:28 AM, Dario Faggioli
> wrote:
> >
> > Soon I won't have access to dario.faggi...@citrix.com email
> > address.
>
> It's sad to hear this. :(
>
Things change. It's lif
Replace, in the 'M:' fields of the components I co-maintain
('CPU POOLS', 'SCHEDULING' and 'RTDS SCHEDULER'), the Citrix
email, to which I don't have access any longer, with my
personal email.
Signed-off-by: Dario Faggioli
---
Cc: Andrew Cooper
TDS entry,
from 'S' to 'M'.
Regards,
Dario
---
Dario Faggioli (1):
MAINTAINERS: update entries to new email address.
MAINTAINERS |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
--
<< ... ... ... >>
----
On Wed, 2017-10-04 at 14:23 +0100, George Dunlap wrote:
> On 09/15/2017 06:35 PM, Dario Faggioli wrote:
> >
> > diff --git a/xen/common/sched_credit.c b/xen/common/sched_credit.c
> > index 3efbfc8..35d0c98 100644
> > --- a/xen/common/sched_credit.c
> >
ebugging-a-flaky-cpu-steal-time-counter-on-a-parav
irtualized-xen-guest/
What's the version of your guest kernel?
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engine
one with w=256).
This is kind of similar with what Linux does with scheduling classes,
but even more flexible.
I am not working on implementing this right now, because I'm busy with
other things, but I would like to do that at some point. And if you're
up for helping, that would be grea
s more extensive use of the c2rqd() helper,
making the code more uniform and easier to read;
* requires the hook csched2_{alloc,free}_pdata to be
defined for Credit2;
* requires the assignment of new values to 'scheduler'
and 'schedule_data.sched_priv' per-cpu variables
On Thu, 2017-09-28 at 18:01 +0200, Dario Faggioli wrote:
> On Thu, 2017-09-28 at 12:18 +0300, Andrii Anisov wrote:
> > > >
> The GP scheduler --like Credit1 or Credit2, if you are on Xen-- will
> give you fairness, but without any precise temporal guarantee. This
> means
header);
---
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)
signature.asc
Description: This is a digitally signed message part
___
And now take 3.
v1:
https://lists.xen.org/archives/html/xen-devel/2017-09/msg01855.html
v2:
https://lists.xen.org/archives/html/xen-devel/2017-09/msg03515.html
I've just took care of Jan's comments to v2. Details in single patches.
Thanks and Regards,
Dario
---
Dario Faggioli (3):
handle the TIMER_SOFTIRQ
(which will perform the timer deactivation and run its handler).
Signed-off-by: Dario Faggioli
---
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Jan Beulich
Cc: Stefano Stabellini
Cc: Julien Grall
Cc: Tim Deegan
---
Changes from v2:
- improved comment in rcu_idle_
rlier next time).
The goal is to let the period timer sefl-adjust to a
number of 'misses', of the order of 1%.
Signed-off-by: Dario Faggioli
Suggested-by: George Dunlap
Reviewed-by: Jan Beulich
---
Cc: Julien Grall
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Stefano Stabellini ,
Cc: Tim
10, 4 or 1 ms. Default valus (used
when this parameter is not specified) is 10ms. Maximum is
100ms.
Signed-off-by: Dario Faggioli
---
Cc: Jan Beulich
Cc: Julien Grall
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Stefano Stabellini ,
Cc: Tim Deegan
---
Changes from v2:
- use '!= 0
g to a whitepaper or in general a place where
a full description of the solution can be found), or it's probably
better to drop the 'close-loop' reference, and explain how MBA can be
useful in another way.
Regards,
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D,
URE.
>
...this is just my opinion (as one of the ones which was part of that
effort, and that plans to get back to finish it at some point). You're
the maintainer, your rules. :-P :-D
Regards,
Dario
--
<> (Raistlin Majere)
giving d2v0 and d2v1 budget and period of 80ms
and 100ms, respectively, will result in a schedulable system, even
though you have 2 CPUs, and 2/10+2/10+80/100+80/100=2 (for Meng, I'm
talking in general, about the fact that, under gEDF, with M CPUs, U> (Raistlin Majere)
-
r_period_ms now isn't
> really needed outside of this function. I'd prefer if you moved it
> and the integer_param() into this function, to limit their scopes
> as much as possible.
>
Ok. idle_timer_period_ms still wants to go into __initdata, right?
Regards,
Dario
--
<> (R
ne, where you pass it a
cached value of NOW().
But I don't feel like arguing too much about this (especially now that
this is patch is the only use case).
If the problem is "just" the parameter (or maybe both the parameter's
and the function's) name(s), I 'd be
Thanks and Regards,
Dario
---
Dario Faggioli (3):
xen: RCU: let the RCU idle timer handler run
xen: RCU: make the period of the idle timer configurable.
xen: RCU: make the period of the idle timer adaptive.
docs/misc/xen-command-line.markdown | 10 +
xen/common/rcupda
10, 4 or 1 ms. Default valus (used
when this parameter is not specified) is 10ms. Maximum is
100ms.
Signed-off-by: Dario Faggioli
---
Cc: Jan Beulich
Cc: Julien Grall
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Stefano Stabellini ,
Cc: Tim Deegan
---
Changes from v1:
- "-" instead of
rlier next time).
The goal is to let the period timer sefl-adjust to a
number of 'misses', of the order of 1%.
Signed-off-by: Dario Faggioli
Suggested-by: George Dunlap
---
Cc: Jan Beulich
Cc: Julien Grall
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Stefano Stabellini ,
Cc: Tim De
handle the TIMER_SOFTIRQ
(which will perform the timer deactivation and run its handler).
Signed-off-by: Dario Faggioli
---
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Jan Beulich
Cc: Stefano Stabellini
Cc: Julien Grall
Cc: Tim Deegan
---
Changes from v1:
- logic changed completely: instead o
On Wed, 2017-09-27 at 15:39 +0200, Dario Faggioli wrote:
> On Wed, 2017-09-27 at 04:30 -0600, Jan Beulich wrote:
> > Or wait -
> > wouldn't all you need be to avoid calling stop_timer() in the
> > call tree above, if the timer's expiry has passed (suitably
> >
g nature and the fact that these CPUs
are slow-- Dom0 is slow in dealing with the backends, to the point that
OSSTest times out.
Then, after the "load spike", everything goes back to normal, the
system is responsive, the logs (like the runqueue dump you posted)
depicts a normal semi-idle sys
On Wed, 2017-09-27 at 08:57 -0400, Robert VanVossen wrote:
> On 9/26/2017 3:12 AM, Dario Faggioli wrote:
> > [Cc-list modified by removing someone and adding someone else]
> >
> > Actually, the best candidate for gaining security support, is IMO
> > ARINC. Code is als
For the reason stated above, I addressed the problem at the
generic code level. If that doesn't fly, I'll do like this. I had
thought about that, and although I haven't tried, I think it works for
this case.
Thanks and Regards,
Dario
--
<> (Raistlin Majere)
---
e often that I would want
to. :-(
I'll try to be more careful.
> Signed-off-by: Wei Liu
> ---
> Cc: George Dunlap
> Cc: Dario Faggioli
> Cc: Marek Marczykowski-Górecki
>
FWIW:
Reviewed-by: Dario Faggioli
Regards,
Dario
--
<> (Raistlin Majere)
--
n/common/sched_credit2.c
> @@ -2831,6 +2831,7 @@ csched2_dom_cntl(
> if ( op->u.credit2.cap > 100 * sdom->nr_vcpus )
> {
> rc = -EINVAL;
> +write_unlock_irqrestore(&prv->lock, flags);
>
Indeed... :-/
Reviewed-b
he first, and for now only, timer that is stopped on
the CPU wake-up-from-idle path).
It is important (in the new version of this patch) for deactivation to
be skipped only in stop_timer(), and not, e.g., in kill_timer(). As, if
someone, in future, will want to kill and free the timer during the
wind
On Tue, 2017-09-26 at 18:28 +0100, Julien Grall wrote:
> On 09/26/2017 08:33 AM, Dario Faggioli wrote:
> > >
> > Here's the logs:
> > http://logs.test-lab.xenproject.org/osstest/logs/113816/test-armhf-
> > armhf-xl-rtds/info.html
>
> It does not seem to b
Yes, personally, I think that, whenever it is that we figure out that
we missed handling a timer interrupt, we should run it then.
Unfortunately, for achieving this, e.g., in the set_timer() case, I
don't see much alternatives to call execute_timer() right in there. But
that would violate
}
>
> Pointless braces. And do you really need min_t()/max_t() here,
> rather than just min()/max()?
>
I probably can. This must be a leftover from a version when the time of
idle_timer_period, and of the IDLE_TIMER_PERIOD_{MIN,MAX} constants
wasn't matching.
Sorry!
Dario
--
&l
d be
> a rather bad choice?
>
Indeed I should enforce a meaningful min too (I'll take it from patch 3
and put it here).
> Also can you please switch to - as the separator in the command
> line argument name?
>
Ah, ok.
Thanks and Regards,
Dario
--
<> (Raistlin Majere)
--
On Mon, 2017-09-25 at 17:23 +0100, Julien Grall wrote:
> On 09/25/2017 03:07 PM, Dario Faggioli wrote:
> > I don't see much in the logs, TBH, but both `xl vcpu-list' and the
> > 'r'
> > debug key seem to suggest that vCPU 0 is running, while the other
>
gaining security support, is IMO
ARINC. Code is also rather simple and "stable" (hasn't changed in the
last... years!) and it's used by DornerWorks' people for some of their
projects (I think?). It's also not tested in OSSTest, though, and
considering how special pu
On Mon, 2017-09-25 at 17:23 +0100, Julien Grall wrote:
> On 09/25/2017 03:07 PM, Dario Faggioli wrote:
> > Hey,
>
> Hi Dario,
>
Hi!
> > I don't see much in the logs, TBH, but both `xl vcpu-list' and the
> > 'r'
> > debug key seem to s
e need?
>
What we have right now, does not build, e.g., with gcc 7.2:
https://pastebin.com/xgeJHkdL
Which, I agree, should be mentioned/hinted at in the changelog.
Regards,
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, P
ubietruck/xen-unstable
So... ARM people? Does this ring any bell? Is this something known, or
easy to explain? What can I do for help?
Regards,
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggiol
tions, when IRQs are enabled.
Signed-off-by: Dario Faggioli
---
Cc: osstest service owner
Cc: George Dunlap
Cc: Wei Liu
Cc: Roger Pau Monné
Cc: Jan Beulich
---
This was spotted by OSSTest's flight 113562:
http://logs.test-lab.xenproject.org/osstest/logs/113562/
http://logs.test-lab.xe
I'll send a patch for that
Thanks and Regards,
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd
7;t
> anticipate the budget replenishment ever to need to walk the domain
> list, but should that change, this would be a bear of a bug to find.
>
Indeed it's rather unlikely for the replenishment handler to have to
use sdom_elem to go through the list of domains.
IAC, if you're conce
On Fri, 2017-09-15 at 12:01 -0400, Meng Xu wrote:
> On Wed, Sep 13, 2017 at 8:16 PM, Dario Faggioli
> wrote:
> >
> > > I'm ok with what it is in this patch, although I feel that we can
> > > kill the
> > > if (scinfo->extratime !=
> > > L
wledge to update it, then fine.
But, technically, the Wiki can be updated and, if possible, it should
be.
Regards,
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senio
On Mon, 2017-09-18 at 20:42 +0100, Anshul Makkar wrote:
>
> On 9/15/17 6:35 PM, Dario Faggioli wrote:
> > Hello,
> >
> > Dario Faggioli (4):
> > xen: sched: introduce 'adjust_affinity' hook.
> > xen: sched: optimize exclusive pi
tions, when IRQs are enabled.
Signed-off-by: Dario Faggioli
---
Cc: osstest service owner
Cc: George Dunlap
Cc: Wei Liu
---
This was spotted by OSSTest's flight 113562:
http://logs.test-lab.xenproject.org/osstest/logs/113562/
http://logs.test-lab.xenproject.org/osstest/logs/113562/te
being inside the critical section (which
> disables
> irqs); there's actually an xfree() right in that function outside the
> critical section.
>
So, during the afternoon, there was an glitch here, in the local
network/NAS of my home office.
It took me a bit to fix it, and that
timers stopped during the wakeup
from idle (e.g., C-states, on x86) path.
To fix that, don't deactivate a timer, while stopping it,
if it has expired. On the contrary, when that happens,
(re-)raise the timer softirq, to make sure the handler is
invoked.
Signed-off-by: Dario Faggioli
---
Cc: A
10, 4 or 1 ms. Default valus (used
when this parameter is not specified) is 10ms. Maximum is
100ms.
Signed-off-by: Dario Faggioli
---
Cc: Jan Beulich
Cc: Julien Grall
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Stefano Stabellini ,
Cc: Tim Deegan
---
docs/misc/xen-command-line.markdown |9
this series, as, without it, the RCU idle timer
handler is never run (and hence, George's algorithm can't work :-/).
Regards,
Dario
---
Dario Faggioli (3):
xen: timers: don't miss a timer event because of stop_timer()
xen: RCU: make the period of the idle timer configurable.
rlier next time).
The goal is to let the period timer sefl-adjust to a
number of 'misses', of the order of 1%.
Signed-off-by: Dario Faggioli
Suggested-by: George Dunlap
---
Cc: Jan Beulich
Cc: Julien Grall
Cc: Andrew Cooper
Cc: George Dunlap
Cc: Stefano Stabellini ,
Cc: Tim Deegan
-
line pCPUs and/or of what pCPUs are part
of the cpupool where the vCPU lives, and that's rather
impractical to store in a per-vCPU flag. Still the
overhead is reduced to "just" one cpumask_subset() (and
only if the newly introduced flag is 'true')!
Signed-off-by: Dario
enough, so go
for it.
Signed-off-by: Dario Faggioli
---
Cc: George Dunlap
Cc: Anshul Makkar
---
Changes from v1:
- use a flag during runtime, as suggested during review;
- make use of the affinity-change hook, introduced in pevious patch.
---
xen/common/sched_credit.c| 35
necessary to make the other twos possible.
Regards,
Dario
---
Dario Faggioli (4):
xen: sched: introduce 'adjust_affinity' hook.
xen: sched: optimize exclusive pinning case (Credit1 & 2)
xen: sched: improve checking soft-affinity
xen: sched: simplify (and spe
h the only exception _csched_cpu_pick()
(in Credit1 code).
Signed-off-by: Dario Faggioli
---
Cc: George Dunlap
Cc: Anshul Makkar
---
xen/common/sched_credit.c | 79 +---
xen/common/sched_credit2.c | 10 ++
xen/common/sched_null.c|8 ++--
xen/i
tions that, despite the fact
that they rarely change, are right now checked very
frequently, even in hot paths.
Note also that this, in future, may turn out as a useful
mean for, e.g., having the schedulers vet, ack or nack
the changes themselves.
Signed-off-by: Dario Faggioli
---
Cc: George Dunlap
INE_XEN_GUEST_HANDLE(xen_sysctl_credit_schedule_t);
> -
> -struct xen_sysctl_credit2_schedule {
> - unsigned ratelimit_us;
> -};
> -typedef struct xen_sysctl_credit2_schedule
> xen_sysctl_credit2_schedule_t;
> -DEFINE_XEN_GUEST_HANDLE(xen_sysctl_credit2_schedule_t);
>
e with this patch at the front), this patch looks fine to me.
Regards,
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)
sign
am_t param;
> >
> > static char *kwd_list[] = { "pool", "sched", NULL };
>
> [..]
> Another possibility would be to drop the cpupool python bindings
> completely (which I would prefer, TBH).
>
+1
Regards,
Dario
--
<> (Raistlin Majere)
>
> +static inline bool has_extratime(const struct rt_vcpu *svc)
> +{
> +return (svc->flags & RTDS_extratime) ? 1 : 0;
> +}
> +
'true' and 'false'. But I think
return svc->flags & RTDS_extratime
is just fine already, without any need for the
On Fri, 2017-09-01 at 11:58 -0400, Meng Xu wrote:
> Revise xl tool use case by adding -e option
> Remove work-conserving from TODO list
>
> Signed-off-by: Meng Xu
>
Reviewed-by: Dario Faggioli
Regards,
Dario
--
<
xl/xl_sched.c| 56 ++
>
>
Oh, and this patch must update docs/man/xl.pod.1.in as well.
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&
08x, cur_budget = 0x%(6)08x%(5)08x ]
>
Why 0x%(2)08d, and not just %(2)d ?
Have you tested it? What does it print?
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer,
878,7 @@ int main_sched_rtds(int argc, char **argv)
> xmalloc(sizeof(libxl_sched_params));
> scinfo.vcpus[0].period = periods[0];
> scinfo.vcpus[0].budget = budgets[0];
> +scinfo.vcpus[0].extratime = extratimes[0] ? 1 : 0;
>
But does thes
27;m ok with what it is in this patch, although I feel that we can
> kill the
> if (scinfo->extratime != LIBXL_DOMAIN_SCHED_PARAM_EXTRATIME_DEFAULT)
> because LIBXL_DOMAIN_SCHED_PARAM_EXTRATIME_DEFAULT is -1.
>
No, sorry, I don't understand what you mean here...
Dario
--
<> (Ra
scinfo->vcpus[i].extratime = !!(vcpus[i].u.rtds.flags &
XEN_DOMCTL_SCHEDRT_extra);
Regards,
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Syst
The patch introduces yet another runqueue arrangement option
for Credit2. In fact, it allows the user to specify, explicitly
and precisely, what pCPUs should belong to which runqueue.
Signed-off-by: Dario Faggioli
Signed-off-by: Praveen Kumar
---
Cc: George Dunlap
Cc: Praveen Kumar
---
This
gt; mechanical.
>
Yes, this is on my radar. I'll be able to look at it shortly.
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd.,
kes up, even on a totally
idle system, while ARM doesn't. But this should not be a problem as, in
both cases, if there's pages to scrub, the CPU just won't go to sleep.
If frequent wakeups is what causes the scrubbing to work ok on the x86,
and not on ARM, it IMO means that the
pe,
> - make x86's paging_domctl() and descendants take a properly typed
> handle,
> - add const in a few places.
>
> Signed-off-by: Jan Beulich
>
Acked-by: Dario Faggioli
Regards,
Dario
--
<> (Raistlin Majere)
--------
On Tue, 2017-09-12 at 09:10 -0600, Jan Beulich wrote:
> By virtue of the struct xen_sysctl container structure, most of them
> are really just cluttering the name space.
>
> Signed-off-by: Jan Beulich
>
Acked-by: Dario Faggioli
Regards,
Dario
--
<
step would be adding it to OSSTest, at which point, we can start
to think about calling it supported.
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)
signature.asc
Descr
help.
>
You being a beginner is not a problem. :-)
The biggest problem I have right now, while trying to help you, is that
I can't see what you are doing, and how you're changing the code.
Regards,
Dario
--
<> (Raistlin Majere)
-
Dar
le something
> that human cannot read.
>
> I think, literally, the 'Faulting linear address' is key point
> because I heard that it represents bad address that I should never
> access.
>
It is, but the only way of understanding why you hit such an access
violation, is underst
On Wed, 2017-09-06 at 12:29 -0700, Stefano Stabellini wrote:
> On Wed, 6 Sep 2017, Dario Faggioli wrote:
> >
> > Or, in general, make sense out of the fact that the stack pointer
> > register changes in such a way that, when we get back in
> > do_softirq(),
> >
On Tue, 2017-08-29 at 17:30 +0100, George Dunlap wrote:
> On 08/18/2017 07:04 PM, Dario Faggioli wrote:
> >
> > What we're trying to avoid is one of those idle CPUs to
> > wake up, only to discover that the grace period is still
> > running, and that it
On Tue, 2017-09-05 at 15:06 -0700, Stefano Stabellini wrote:
> On Tue, 5 Sep 2017, Dario Faggioli wrote:
> >
> > Re-checking things now, I actually do see that context_switch() on
> > ARM
> > is not 'terminal'. It call schedule_tail(), which on x86 does not
On Fri, 2017-06-09 at 18:41 +0200, Dario Faggioli wrote:
> Hey Praveen,
>
Hey, hello again!
> Here we are, sorry for the delay.
>
So, about this patch... I haven't seen a new version (or did I perhaps
miss it?).
I'm asking because I do have it half done myself, and it w
On Thu, 2017-08-24 at 20:42 +0100, Anshul Makkar wrote:
> On 8/18/17 4:50 PM, Dario Faggioli wrote:
> >
> > @@ -1515,7 +1633,16 @@ static void reset_credit(const struct
> > scheduler *ops, int cpu, s_time_t now,
> > * that the credit it has
> > +#define RCU_IDLE_TIMER_PERIOD MILLISECS(10)
> > >
> > > With you even mentioning that the original Linux code has ways
> > > to use different values, wouldn't it be worth allowing this to be
> > > command line controllable?
> >
> > Da
orrect?
Also, mostly out of curiosity, still looking at ARM code, I'm not
getting at all how continue_new_vcpu() works (e.g., when/how is it
invoked?).
Thanks and Regards,
Dario
--
<> (Raistlin Majere)
-
Dario Faggioli, Ph.D,
Il 29 Ago 2017 4:03 PM, George Dunlap ha scritto:
On Wed, Aug 16, 2017 at 5:45 PM, Dario Faggioli
wrote:
> In fact, right now, we read it at every iteration of the loop.
> The reason it's done like this is how context switch was handled
> on IA64 (see commit ae9bfcdc, "[X
1 - 100 of 2287 matches
Mail list logo