2014-06-22 01:04:38

by Ben Hutchings

[permalink] [raw]

Subject: Latency histogram broken after "hrtimer: Set expiry time before switch_hrtimer_base()"

---

In an rt-kernel with CONFIG_MISSED_TIMER_OFFSETS_HIST enabled,
__hrtimer_start_range_ns() now crashes, as new_base is not assigned
before it is used.

I'm not sure how this should be fixed; is it:

--- a/kernel/hrtimer.c
+++ b/kernel/hrtimer.c
@@ -1108,7 +1108,7 @@ int __hrtimer_start_range_ns(struct hrti

#ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
{
- ktime_t now = new_base->get_time();
+ ktime_t now = base->get_time();

if (ktime_to_ns(tim) < ktime_to_ns(now))
timer->praecox = now;
--- END ---

or:

--- a/kernel/hrtimer.c
+++ b/kernel/hrtimer.c
@@ -1106,6 +1106,11 @@ int __hrtimer_start_range_ns(struct hrti
#endif
}

+ hrtimer_set_expires_range_ns(timer, tim, delta_ns);
+
+ /* Switch the timer base, if necessary: */
+ new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
+
#ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
{
ktime_t now = new_base->get_time();
@@ -1117,11 +1122,6 @@ int __hrtimer_start_range_ns(struct hrti
}
#endif

- hrtimer_set_expires_range_ns(timer, tim, delta_ns);
-
- /* Switch the timer base, if necessary: */
- new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
-
timer_stats_hrtimer_set_start_info(timer);

leftmost = enqueue_hrtimer(timer, new_base);
--- END ---

or something else?

Ben.

--
Ben Hutchings
73.46% of all statistics are made up.

---

Attachments:

signature.asc (811.00 B)
This is a digitally signed message part

2014-06-22 04:10:28

by Mike Galbraith

[permalink] [raw]

Subject: Re: Latency histogram broken after "hrtimer: Set expiry time before switch_hrtimer_base()"

---

On Sun, 2014-06-22 at 02:04 +0100, Ben Hutchings wrote:
> In an rt-kernel with CONFIG_MISSED_TIMER_OFFSETS_HIST enabled,
> __hrtimer_start_range_ns() now crashes, as new_base is not assigned
> before it is used.

Oh yeah, forgot about this.

> I'm not sure how this should be fixed; is it:

My (3.12-ish tree) merge resolution was the later, but it shouldn't
matter which you choose, what's stored where is unchanged pre/post.

> --- a/kernel/hrtimer.c
> +++ b/kernel/hrtimer.c
> @@ -1108,7 +1108,7 @@ int __hrtimer_start_range_ns(struct hrti
>
> #ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
> {
> - ktime_t now = new_base->get_time();
> + ktime_t now = base->get_time();
>
> if (ktime_to_ns(tim) < ktime_to_ns(now))
> timer->praecox = now;
> --- END ---
>
> or:
>
> --- a/kernel/hrtimer.c
> +++ b/kernel/hrtimer.c
> @@ -1106,6 +1106,11 @@ int __hrtimer_start_range_ns(struct hrti
> #endif
> }
>
> + hrtimer_set_expires_range_ns(timer, tim, delta_ns);
> +
> + /* Switch the timer base, if necessary: */
> + new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
> +
> #ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
> {
> ktime_t now = new_base->get_time();
> @@ -1117,11 +1122,6 @@ int __hrtimer_start_range_ns(struct hrti
> }
> #endif
>
> - hrtimer_set_expires_range_ns(timer, tim, delta_ns);
> -
> - /* Switch the timer base, if necessary: */
> - new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
> -
> timer_stats_hrtimer_set_start_info(timer);
>
> leftmost = enqueue_hrtimer(timer, new_base);
> --- END ---
>
> or something else?
>
> Ben.
>

2014-06-22 04:24:14

by Mike Galbraith

[permalink] [raw]

Subject: Re: Latency histogram broken after "hrtimer: Set expiry time before switch_hrtimer_base()"

---

(CCs stable -rt maintainer)

On Sun, 2014-06-22 at 06:10 +0200, Mike Galbraith wrote:
> On Sun, 2014-06-22 at 02:04 +0100, Ben Hutchings wrote:
> > In an rt-kernel with CONFIG_MISSED_TIMER_OFFSETS_HIST enabled,
> > __hrtimer_start_range_ns() now crashes, as new_base is not assigned
> > before it is used.
>
> Oh yeah, forgot about this.
>
> > I'm not sure how this should be fixed; is it:
>
> My (3.12-ish tree) merge resolution was the later, but it shouldn't
> matter which you choose, what's stored where is unchanged pre/post.
>
> > --- a/kernel/hrtimer.c
> > +++ b/kernel/hrtimer.c
> > @@ -1108,7 +1108,7 @@ int __hrtimer_start_range_ns(struct hrti
> >
> > #ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
> > {
> > - ktime_t now = new_base->get_time();
> > + ktime_t now = base->get_time();
> >
> > if (ktime_to_ns(tim) < ktime_to_ns(now))
> > timer->praecox = now;
> > --- END ---
> >
> > or:
> >
> > --- a/kernel/hrtimer.c
> > +++ b/kernel/hrtimer.c
> > @@ -1106,6 +1106,11 @@ int __hrtimer_start_range_ns(struct hrti
> > #endif
> > }
> >
> > + hrtimer_set_expires_range_ns(timer, tim, delta_ns);
> > +
> > + /* Switch the timer base, if necessary: */
> > + new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
> > +
> > #ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
> > {
> > ktime_t now = new_base->get_time();
> > @@ -1117,11 +1122,6 @@ int __hrtimer_start_range_ns(struct hrti
> > }
> > #endif
> >
> > - hrtimer_set_expires_range_ns(timer, tim, delta_ns);
> > -
> > - /* Switch the timer base, if necessary: */
> > - new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
> > -
> > timer_stats_hrtimer_set_start_info(timer);
> >
> > leftmost = enqueue_hrtimer(timer, new_base);
> > --- END ---
> >
> > or something else?
> >
> > Ben.
> >
>
>

2014-06-22 10:25:59

by Carsten Emde

[permalink] [raw]

Subject: Re: Latency histogram broken after "hrtimer: Set expiry time before switch_hrtimer_base()"

---

On 06/22/2014 06:24 AM, Mike Galbraith wrote:
> [..]
> On Sun, 2014-06-22 at 06:10 +0200, Mike Galbraith wrote:
>> On Sun, 2014-06-22 at 02:04 +0100, Ben Hutchings wrote:
>>> In an rt-kernel with CONFIG_MISSED_TIMER_OFFSETS_HIST enabled,
>>> __hrtimer_start_range_ns() now crashes, as new_base is not assigned
>>> before it is used.
>>
>> Oh yeah, forgot about this.
>>
>>> I'm not sure how this should be fixed; is it:
>>
>> My (3.12-ish tree) merge resolution was the later, but it shouldn't
>> matter which you choose, what's stored where is unchanged pre/post.
I can confirm that 3.12.22-rt34 systems with
CONFIG_MISSED_TIMER_OFFSETS_HIST enabled crash at an early boot stage.
After applying Ben's second patch proposal, all is back to normal.

>>
>>> --- a/kernel/hrtimer.c
>>> +++ b/kernel/hrtimer.c
>>> @@ -1108,7 +1108,7 @@ int __hrtimer_start_range_ns(struct hrti
>>>
>>> #ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
>>> {
>>> - ktime_t now = new_base->get_time();
>>> + ktime_t now = base->get_time();
>>>
>>> if (ktime_to_ns(tim) < ktime_to_ns(now))
>>> timer->praecox = now;
>>> --- END ---
>>>
>>> or:
Didn't use this one.

>>> --- a/kernel/hrtimer.c
>>> +++ b/kernel/hrtimer.c
>>> @@ -1106,6 +1106,11 @@ int __hrtimer_start_range_ns(struct hrti
>>> #endif
>>> }
>>>
>>> + hrtimer_set_expires_range_ns(timer, tim, delta_ns);
>>> +
>>> + /* Switch the timer base, if necessary: */
>>> + new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
>>> +
>>> #ifdef CONFIG_MISSED_TIMER_OFFSETS_HIST
>>> {
>>> ktime_t now = new_base->get_time();
>>> @@ -1117,11 +1122,6 @@ int __hrtimer_start_range_ns(struct hrti
>>> }
>>> #endif
>>>
>>> - hrtimer_set_expires_range_ns(timer, tim, delta_ns);
>>> -
>>> - /* Switch the timer base, if necessary: */
>>> - new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED);
>>> -
>>> timer_stats_hrtimer_set_start_info(timer);
>>>
>>> leftmost = enqueue_hrtimer(timer, new_base);
>>> --- END ---
>>> or something else?
Tested-by: Carsten Emde <[email protected]>

Thanks,
-Carsten.