week.
BTW, can you update my email address to raist...@linux.it ?
Regards,
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
signature.asc
Description: This is a dig
On Tue, 2017-12-05 at 21:58 +0530, Praveen Kumar wrote:
> On Tue, Sep 5, 2017 at 11:26 PM, Dario Faggioli
> wrote:
> >
> > I'm asking because I do have it half done myself, and it would not
> > take
> > too much time to me to finish it.
> >
> > I
> > - Passing the "urgent" indicator via parameter to idle (you didn't
> >like that)
> >
> > - Make urgent_count a plain percpu variable
>
> Named sched_urgent_count or some such, this latter option doesn
On Fri, 2019-08-23 at 18:16 -0700, Stefano Stabellini wrote:
> On Wed, 21 Aug 2019, Dario Faggioli wrote:
> >
> > Hey, Stefano, Julien,
> >
> > Here's another patch.
> >
> > Rather than a debug patch, this is rather an actual "proposed
> >
k it's worth to do it right now, out
of concerns of potential scalability issues for socket-scheduling.
We're discussing about merging core-scheduling, but keeping it disabled
for now. We're not even sure if/when people will start to use *it*. I
currently don't see much use ca
t; >
> > Given the response on an earlier similar question, I don't suppose
> > I could talk you into dropping the sched_ prefix here?
>
> I like it better with prefix. Any opinions by the scheduler
> maintainers?
>
I do like it with prefix better too.
next_slice.time);
> -trace_continue_running(next->vcpu_list);
> -return continue_running(prev->vcpu_list);
> + prev->next_time);
> }
>
Mmm... I'm sorry, but I'm not sure I understand what is going on here.
Do you mind explai
On Mon, 2019-09-09 at 11:33 +0200, Juergen Gross wrote:
> Instead of having a cpupool_dprintk() define just use debugtrace.
>
> Signed-off-by: Juergen Gross
>
Acked-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engine
+return runnable;
> > +}
>
> Especially for understanding the (correctness of the) credit1
> changes it would be rather helpful if once again this function
> actually iterated over all vCPU-s right away (even if there's
> only one per unit right now), to see
ically
introducing no functional (or at least behavioral) change, is this
right too?
- can you provide some more detail about the race. When/how it occurs
and how the changes done in credit and rt fix it?
Thanks and Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggi
;t
> > have time to review the series).
> >
Ok. I'm back from vacation and am now looking at the patches. I am
traveling next week, but I hope to be able to continue reviewing during
that.
> > No decisions are official until discussed on xen-devel; so the
> > decision
arlier, so
> do that just before assigning the cpus. Initialize free cpus with all
> online cpus at that time in order to be able to add the cpu notifier
> late, too.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about
.
Is this anything that you've seen or that you can reproduce?
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Rai
or Linux, but I think
this is the case for FreeBSD too. So, even if the patch would have a
negative impact (which again I don't think it's the case), we probably
won't see them.
On this grounds (and, of course, on the one that I've looked at the
code, and think it's corre
s with core-scheduling actually enabled.
With these patches applied, but cpu-scheduling selected at boot, fully
idle cores are still preferred, and the vCPUs will still be spread
among them (as soon as there's any available).
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtua
ILED case from cpu_schedule_callback().
>
> Note that we are calling now schedule_cpu_switch() in stop_machine
> context so we need to switch from spinlock_irq to spinlock_irqsave.
>
> Signed-off-by: Juergen Gross
> ---
> V2: rename cpupool_unassign_cpu_[epi|pro]logue() (Dario
>curr)
> >
> > By moving this a few lines down if could become an inline function
> > as it seems, if (see above) its use as an lvalue is not intended.
>
> I like that idea. Will change to inline function.
>
Yes, definitely better. With this (turning to inline fun
ode churn) is
> > indeed better than naming at least new variable instances properly
> > right away.
>
> Okay with me. What do scheduler maintainers think?
>
I do prefer 'sr'. As to whether this happens right now, but only for
new variables, or all at once in a fo
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Rename vcpu_schedule_[un]lock[_irq]() to
> unit_schedule_[un]lock[_irq]()
> and let it take a sched_unit pointer instead of a vcpu pointer as
> parameter.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario F
broken affinity" when any of its vcpus has the affinity_broken flag
> set.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Rename the scheduler related perf counters from vcpu* to unit* where
> appropriate.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtuali
Hopefully this gives some level of confidence to this series and the
> plan about including it into 4.13 [1]
>
Thanks a lot for doing this, and for letting us know.
Can I ask whether the tests were done using Credit2 (i.e., upstream
default) or Credit1, as scheduler?
Thanks again and
>
Reviewed-by: Dario Faggioli
(Stands with parameter constified as suggested.)
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Rai
en called" would
sound better.
It's, of course, not your fault, but I guess we can consider adjusting
it, since we are touching the line above it.
Or maybe you also want to queue it up for the cleanup series?
Regards
scheduler needs to remove the vcpu-data from a list).
>
> So instead xfree() make use of the sched_free_vdata() hook.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Softwa
Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Raistlin Majere)
signature.asc
De
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Switch null scheduler completely from vcpu to sched_unit usage.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
With, if possible, all the occurrences of 'item' in comments actually
replaced w
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Switch rt scheduler completely from vcpu to sched_unit usage.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
With just one additional note.
> * This scheduler follows the Preemptive Global Earliest D
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Switch arinc653 scheduler completely from vcpu to sched_unit usage.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Similar considerations to the ones I made on sched_rt applies here, I
guess.
But again, as
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Switch credit scheduler completely from vcpu to sched_unit usage.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engine
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Switch credit2 scheduler completely from vcpu to sched_unit usage.
>
> As we are touching lots of lines remove some white space at the end
> of
> the line, too.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dar
nit->vcpu_list);
> +}
> +
> +static inline void sched_unit_unpause(struct sched_unit *unit)
> +{
> +vcpu_unpause(unit->vcpu_list);
> +}
> +
So, isn't this another one of those places where we could have the
for_each_unit_vcpu() loop already?
Regards
--
Dario Fa
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Let the schedulers put a sched_unit pointer into struct task_slice
> instead of a vcpu pointer.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me
;
> At the same time introduce a state_entry_time field in struct
> sched_unit being updated whenever the is_running indicator is
> changed.
> Use that new field in the schedulers instead of the similar vcpu
> field.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Fag
On Wed, 2019-09-25 at 16:19 +0100, Julien Grall wrote:
> (+Juergen)
>
> Hi Dario,
>
Hi,
> On 11/09/2019 14:53, Dario Faggioli wrote:
> > On Fri, 2019-08-23 at 18:16 -0700, Stefano Stabellini wrote:
> > Ok, thanks again for testing, and good to know.
> >
o.
>
> Signed-off-by: Juergen Gross
> Acked-by: Andrew Cooper
> Acked-by: Julien Grall
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Use sched_units instead of vcpus in schedule(). This includes the
> introduction of sched_unit_runstate_change() as a replacement of
> vcpu_runstate_change() in schedule().
>
> Signed-off-by: Juergen Gross
>
Reviewed-
On Wed, 2019-09-25 at 14:04 +0200, Jürgen Groß wrote:
> On 25.09.19 00:33, Dario Faggioli wrote:
> >
> > > As it is needed anyway call vcpu_sync_execstate() for each
> > > vcpu of the unit when changing processors.
> > >
> > Again, what do you mean with
...
>
> > I can merge this patch into patch 3 if you like that better.
>
> ... yes, personally I'd prefer this, but in the end it's the call
> of the scheduler maintainers.
>
Yes, I think I like it better too for the patches to be merged.
Regards
--
Dario Faggio
; the function afaics, and hence "ops" could be used here.
>
> "sd" is now "sr" everywhere.
>
> And yes, using ops here seems okay.
>
> > Preferably with the suggested adjustments (but final judgement is
> > with the scheduler maintainers any
On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> Having a pointer to struct cpupool in struct sched_resource instead
> of per cpu is enough.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
socket scheduling.
>
> Signed-off-by: Juergen Gross
> Acked-by: Jan Beulich
>
Acked-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
--
ividual cpus are offlined.
>
> Disabling the scheduler is done by replacing the softirq handler
> with a dummy scheduling routine only enabling tasklets to run.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me
ng. In fact, as said elsewhere
already, I really don't expect anyone to want to use either die-
scheduling, socket-scheduling or anything different than core-
scheduling anytime soon.
I'll look into the Credit2 runq
(x86 only)
> "sched-gran".
>
> According to the selected granularity sched_granularity is set after
> all cpus are online.
>
> A test is added for all sched resources holding the same number of
> cpus. Fall back to core- or cpu-scheduling in that case.
>
> Si
es should be replaced by the master_cpu number of the
> local
> sched_resource instead.
>
> Add a helper to get that sched_resource master_cpu and modify the
> schedulers to use it in a correct way.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
directly set the different units in struct
> sched_unit instead of using an on-stack copy for returning the data.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Softwa
On Thu, 2019-09-26 at 06:51 +0200, Jürgen Groß wrote:
> On 25.09.19 18:37, Dario Faggioli wrote:
> > > This is done by vcpu_sync_execstate(). Without the explicit call
> > > it would be done either when the vcpu is being scheduled on the
> > > new
> > > c
eems to be more expensive as the current variant, too.
>
> It's not this much more expensive, and it eliminates unexpected
> (as I would call it) behavior, so I think I'd go this route.
>
So, I honestly like the way it's currently done in Juergen's pat
On Thu, 2019-09-26 at 14:37 +0200, Jürgen Groß wrote:
> On 26.09.19 11:46, Dario Faggioli wrote:
> > On Sat, 2019-09-14 at 10:52 +0200, Juergen Gross wrote:
> > > Add a scheduling granularity enum ("cpu", "core", "socket") for
> > > specifi
On Thu, 2019-09-26 at 16:40 +0200, Jürgen Groß wrote:
> On 26.09.19 15:53, Dario Faggioli wrote:
> > However, I'm not sure I understand what it is the issue that Jan
> > thinks
> > that has, and in what sense the code/behavior is regarded as
> > "unexpect
. In that
> case we need the sched_wake() call even if the unit is still running.
>
TBH, I think it is ok for this call to be unconditional. Indeed it
looks a bit weird when you compare this to the sched_sleep() calls in
vcpu_sleep_nosync_locked(), as they are conditional, but I think a
ic physical cpu
>
> Signed-off-by: Juergen Gross
>
With that fixed:
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
-
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Raistlin Majere)
signature.asc
Description: Th
asily. Note that
> this functionality is x86 only.
>
> Signed-off-by: Juergen Gross
> Acked-by: Julien Grall
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Raistlin Majere)
signature.asc
Descript
scheduler() use it.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
One thing I spotted is that, here...
> --- a/xen/common/schedule.c
> +++ b/xen/common/schedule.c
> @@ -157,26 +157,32 @@ static inline struct scheduler
> *dom_scheduler(const struct domain *d)
scheduled on the old
> processor).
>
> vcpu_move_locked() is switched to schedule unit, too.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Sof
it return a cpumask with only
> one bit set per scheduling resource.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, S
DULE_SOFTIRQ events.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Raistlin Majere)
signature.asc
Description: This is a
() as that relies on for_each_sched_unit_vcpu() which in
> turn needs the vcpu already as a member of the domain's vcpu linked
> list, which isn't the case.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualiz
, but with core
> scheduling
> it will be possible to have higher values, too.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE L
d by other domains (as far as possible). Softirqs are considered
> to
> be save.
>
> In order to avoid livepatching when going to guest idle another
> variant of reset_stack_and_jump() not calling
> check_for_livepatch_work
> is needed.
>
> Signed-off-by: Juergen Gr
tic as it is used in
> schedule.c only.
>
> Signed-off-by: Juergen Gross
>
I'm ok with the code in this patch.
I'd just like to see the comment(s) around the asymmetry between
vcpu_sleep_xxx() and vcpu_wake() added, as agreed upon this morning (in
the V3 thread).
Regar
> Move the granularity related variables to cpupool.c as they are now
> used form there only.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
pus as the idle scheduler doesn't need such data.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE La
ng resources. This in turn requires to modify
> the cpu <-> scheduling resource relation. In order to be able to free
> unused resources protect struct sched_resource via RCU. This ensures
> there are no users left when freeing such a resource.
>
> Signed-off-by: Juergen Gro
gen Gross
>
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Raistlin Majere)
signature.asc
Descri
ing resource with more than one cpu.
>
> Handle that by having arrays of old/new pdata and vdata and loop over
> those where appropriate.
>
> Additionally the scheduling resource(s) must either be merged or
> split.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dari
On Fri, 2019-09-27 at 09:00 +0200, Juergen Gross wrote:
> With core scheduling active it is necessary to move multiple cpus at
> the same time to or from a cpupool in order to avoid split scheduling
> resources in between.
>
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario
I've asked more comments this morning.
If the patch have to go in by today, I guess I'm fine with that patch
going in as it is (i.e., committers can add my `Reviewed-by:`), as far
as the additional comments come in a follow-up patch as soon as
possible.
Thanks and Regards,
Dario
--
a new helper unit_runnable_state() which
> will save the new runstate of all tested vcpus in a new field of the
> vcpu struct.
>
> Signed-off-by: Juergen Gross
> ---
> RFC V2:
> - new patch
> V3:
> - add vcpu loop to unit_runnable_state() right now instead of doing
ke cpu parameter unsigned int (Jan Beulich)
> - comments (Jan Beulich)
> - use true instead 1 for setting bool (Jan Beulich)
> - const parameter (Jan Beulich)
> V5:
> - add comments (Dario Faggioli)
>
Reviewed-by: Dario Faggioli
Thanks and Regards
--
Dario Faggioli, Ph.D
http
d by other domains (as far as possible). Softirqs are considered
> to
> be save.
>
> In order to avoid livepatching when going to guest idle another
> variant of reset_stack_and_jump() not calling
> check_for_livepatch_work
> is needed.
>
> Signed-off-by: Juergen Gr
g.
>
Mmm... I'm not an expert of these things, but doesn't this means we
need a "Signed-off-by: Xi Xiong " then? Cc-ing Lars...
> Signed-off-by: Eslam Elnikety
> Reviewed-by: Leonard Foerster
> Reviewed-by: Petre Eftime
>
About the patch itself:
Acked-by
t; +s->free_pdata(s, data, cpu);
> +else
> +/*
> + * Check that if there isn't a free_pdata hook, we haven't
> got any
> + * data we're expected to deal with.
> + */
> + ASSERT(!data);
> +}
>
Doing,
hedule_tail()
> with ASSERT() which is the development (debug) build guard.
>
> Signed-off-by: Andrii Anisov
>
FWIW,
Reviewed-by: Dario Faggioli
Thanks and Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE
tity* with which the scheduler
deals.
> I realize changing it here probably means also changing it in the
> next
> 43 patches as well.
>
Indeed! :-O
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://w
ne length, indentation) manually afterwards.
>
Ah, right, I didn't spot that 'item' and 'unit' are the same length.
Makes sense.
> So Dario,
> would you be fine with "unit"? This might save me some hours of work.
>
I guess I am, yes. I like it more than
ests.
>
+1
> Hmm, thinking more about it: why don't we just drop those
> ASSERT/BUG_ON
> for mandatory functions and test them when doing the global_init()
> loop
> over all schedulers. We could just reject schedulers with missing
> functions.
>
+10
:-D
Regards
--
sh I managed to put a cpupool test in OSSTest. I guess I can
pickup my old draft, but I need some time to put together an OSSTest
dev environment again. :-/
Anyway...
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
Thanks and Regards
--
Dario Faggioli, Ph.D
http://about.me/dar
portant.
>
I agree.
Isn't it the case that (but note that I'm just thinking out loud here),
if we make smt= and sched-gran= per-cpupool, the user gains the chance
to use both, if he/she wants (e.g., for testing)?
If yes, is such a thing valuable enough that it'd it make sense to wo
On Fri, 2019-05-31 at 14:31 +0800, Baodong Chen wrote:
> when 'periodic_period' is zero, there is no need to initialize 'now'.
>
> Signed-off-by: Baodong Chen
> ---
> xen/common/schedule.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
Ack
domain *d);
> int cpupool_move_domain(struct domain *d, struct cpupool *c);
> int cpupool_do_sysctl(struct xen_sysctl_cpupool_op *op);
> +#ifdef CONFIG_HAS_KEYHANDLER
> void schedule_dump(struct cpupool *c);
> extern void dump_runq(unsigned char
nding something that is short and, at the same time, clear enough),
or we change how the variable is using.
Like, e.g., in csched_schedule(), we first set it to 0, and then we
update it with `now` for `prev` if `prev != next && !is_idle(prev)` (or
something like that)
The rest of the patch looks
s to be as simple as possible, even simpler than null...
And I think we can easily see why that's the case, just by looking at
what code this patch let us remove (e.g., the need for some of the
checking of `system_state` in cpupool or scheduler code, just to
men
ake of not making rebasing 60 patches more complicated than
it must be already, for you :-D).
Let me know what you think.
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
--
On Mon, 2019-06-10 at 18:16 +0300, Andrii Anisov wrote:
> On 31.05.19 16:24, Dario Faggioli wrote:
> > --- a/xen/common/sched_credit.c
> > +++ b/xen/common/sched_credit.c
> > @@ -175,6 +175,9 @@ struct csched_vcpu {
> > atomic_t credit;
>
optimize this write away.
> >
> > So, I guess I'd be OK with this particular hunk. Dario, any
> > opinions?
>
I'm ok with it too, but...
> I'd rather switch to PTR_ERR() here dropping the perr parameter.
>
... I'd be even more ok with this! :-)
thread already, I personally do like this
(and things like to this) quite a bit... Thanks for doing it!
Reviewed-by: Dario Faggioli
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE ht
uergen Gross
>
Reviewed-by: Dario Faggioli
Thanks and Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Raistlin Majere)
signature.
is change, we don't any longer.
That being said, I've tried to think about how this could be a problem,
but failed at imagining such a scenario, so:
> Signed-off-by: Juergen Gross
>
Reviewed-by: Dario Faggioli
I'm wondering whether it make sense for the above to be quickl
an appropriate error code for
> > this new case. -EPERM, -EACCES, or -EIO would all seem more
> > appropriate to me (and perhaps there are further ones).
>
> I think -EIO or -EBUSY would be the best fit.
>
I agree, with mild preference for EBUSY.
Regards
--
Dario Faggioli,
On Wed, 2019-06-12 at 10:06 +0200, Juergen Gross wrote:
> On 12.06.19 09:40, Jan Beulich wrote:
> > > > > On 11.06.19 at 18:55, wrote:
> > > On Tue, 2019-05-28 at 12:32 +0200, Juergen Gross wrote:
> > > > Signed-off-by: Juergen Gross
> > > >
however, I think the comment needs to be adjusted too, and
the commit message needs to briefly mention this new change we're
doing.
Maybe, for the comment, add something like:
"The smp_wmb() corresponds to the barrier implicit in successfully
taking the lock."
And, for the chang
also:
> - turn last_sched_time into s_time_t, which is more appropriate.
> - properly const-ify related argument of
> __csched_vcpu_is_cache_hot().
>
> Signed-off-by: Andrii Anisov
>
Reviewed-by: Dario Faggioli
Thanks and Regards
--
Dario Faggiol
ponding *_schedule_unlock_*() ones, of course. ;-P
Regards
--
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
---
<> (Raistlin Majere)
signature.asc
gen Gross
>
Acked-by: Dario Faggioli
Thanks,
Dario
--
<> (Raistlin Majere)
-----
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Software Engineer @ SUSE https://www.suse.com/
signature.asc
Description: Th
Depends on enabled
> > > > schedulers.
> > The Kconfig default is credit2.
>
> Only when credit2 is compiled in, which isn't even the the case for
> all
> Kconfig fragments we have in tree.
>
So, basically, what you're saying is that we need a patch t
1 - 100 of 789 matches
Mail list logo