On Wed, Feb 13, 2019 at 7:45 AM Peter Zijlstra wrote:
>
> Before that, x86_64 switch_to() read like (much simplified):
>
> asm volatile ( /* do RSP twiddle */
> : /* output */
> : /* input */
> : "memory", "cc", "flags");
>
> (see __EXTRA_CLOBBER)
>
>
> On Feb 13, 2019, at 2:21 PM, Peter Zijlstra wrote:
>
> On Wed, Feb 13, 2019 at 10:51:24AM -0800, Andy Lutomirski wrote:
>>> On Feb 13, 2019, at 7:45 AM, Peter Zijlstra wrote:
>
>>> Which I suppose means that GCC generates the PUSHF/POPF to preserve the
>>> EFLAGS, since we mark those
On Wed, Feb 13, 2019 at 10:51:24AM -0800, Andy Lutomirski wrote:
> > On Feb 13, 2019, at 7:45 AM, Peter Zijlstra wrote:
> > Which I suppose means that GCC generates the PUSHF/POPF to preserve the
> > EFLAGS, since we mark those explicitly clobbered.
> >
>
> Not quite. A flags clobber doesn’t
On Wed, Feb 13, 2019 at 04:45:32PM +0100, Peter Zijlstra wrote:
> On Wed, Feb 13, 2019 at 03:41:45PM +0100, Peter Zijlstra wrote:
> > On Wed, Feb 13, 2019 at 02:39:22PM +, Julien Thierry wrote:
> > > Hi Peter,
> > >
> > > On 13/02/2019 14:25, Peter Zijlstra wrote:
> > > > On Wed, Feb 13, 2019
On Wed, Feb 13, 2019 at 03:41:45PM +0100, Peter Zijlstra wrote:
> On Wed, Feb 13, 2019 at 02:39:22PM +, Julien Thierry wrote:
> > Hi Peter,
> >
> > On 13/02/2019 14:25, Peter Zijlstra wrote:
> > > On Wed, Feb 13, 2019 at 02:00:26PM +, Will Deacon wrote:
> > >> The difference is because
On Wed, Feb 13, 2019 at 03:40:00PM +0100, Peter Zijlstra wrote:
> So far, all 6 in-tree users are indeed free of dodgy code, but *groan*.
because of this, there must also not be tracepoints (even implicit ones
like function-trace) between user_access_{begin,end}().
And while that is unlikely
On Wed, Feb 13, 2019 at 02:39:22PM +, Julien Thierry wrote:
> Hi Peter,
>
> On 13/02/2019 14:25, Peter Zijlstra wrote:
> > On Wed, Feb 13, 2019 at 02:00:26PM +, Will Deacon wrote:
> >> The difference is because getting preempted in the sequence above is
> >> triggered off the back of an
On Wed, Feb 13, 2019 at 02:24:34PM +, Julien Thierry wrote:
> On 13/02/2019 14:17, Peter Zijlstra wrote:
> > On Wed, Feb 13, 2019 at 02:00:26PM +, Will Deacon wrote:
> >>> This; how is getting preempted fundamentally different from scheduling
> >>> ourselves?
> >>
> >> The difference is
Hi Peter,
On 13/02/2019 14:25, Peter Zijlstra wrote:
> On Wed, Feb 13, 2019 at 02:00:26PM +, Will Deacon wrote:
>> The difference is because getting preempted in the sequence above is
>> triggered off the back of an interrupt. On arm64, and I think also on x86,
>> the user access state (SMAP
On 13/02/2019 14:17, Peter Zijlstra wrote:
> On Wed, Feb 13, 2019 at 02:00:26PM +, Will Deacon wrote:
>>> This; how is getting preempted fundamentally different from scheduling
>>> ourselves?
>>
>> The difference is because getting preempted in the sequence above is
>> triggered off the
On Wed, Feb 13, 2019 at 02:00:26PM +, Will Deacon wrote:
> The difference is because getting preempted in the sequence above is
> triggered off the back of an interrupt. On arm64, and I think also on x86,
> the user access state (SMAP or PAN) is saved and restored across exceptions
> but not
On Wed, Feb 13, 2019 at 02:00:26PM +, Will Deacon wrote:
> > This; how is getting preempted fundamentally different from scheduling
> > ourselves?
>
> The difference is because getting preempted in the sequence above is
> triggered off the back of an interrupt. On arm64, and I think also on
On 13/02/2019 14:00, Will Deacon wrote:
> Hi Peter,
>
> On Wed, Feb 13, 2019 at 02:17:20PM +0100, Peter Zijlstra wrote:
>> On Wed, Feb 13, 2019 at 10:50:21AM +, Julien Thierry wrote:
>>> On 13/02/2019 10:35, Peter Zijlstra wrote:
On Tue, Feb 12, 2019 at 09:15:13AM +, Julien
Hi Peter,
On Wed, Feb 13, 2019 at 02:17:20PM +0100, Peter Zijlstra wrote:
> On Wed, Feb 13, 2019 at 10:50:21AM +, Julien Thierry wrote:
> > On 13/02/2019 10:35, Peter Zijlstra wrote:
> > > On Tue, Feb 12, 2019 at 09:15:13AM +, Julien Thierry wrote:
> > >
> > > diff --git
On Wed, Feb 13, 2019 at 02:17:20PM +0100, Peter Zijlstra wrote:
> On Wed, Feb 13, 2019 at 10:50:21AM +, Julien Thierry wrote:
> > On 13/02/2019 10:35, Peter Zijlstra wrote:
> > > On Tue, Feb 12, 2019 at 09:15:13AM +, Julien Thierry wrote:
> > >
> > > diff --git a/kernel/sched/core.c
On Wed, Feb 13, 2019 at 10:50:21AM +, Julien Thierry wrote:
> On 13/02/2019 10:35, Peter Zijlstra wrote:
> > On Tue, Feb 12, 2019 at 09:15:13AM +, Julien Thierry wrote:
> >
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index a674c7db..b1bb7e9 100644
> > ---
On 13/02/2019 10:35, Peter Zijlstra wrote:
> On Tue, Feb 12, 2019 at 09:15:13AM +, Julien Thierry wrote:
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index a674c7db..b1bb7e9 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -3289,6
On Tue, Feb 12, 2019 at 09:15:13AM +, Julien Thierry wrote:
> >>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> >>> index a674c7db..b1bb7e9 100644
> >>> --- a/kernel/sched/core.c
> >>> +++ b/kernel/sched/core.c
> >>> @@ -3289,6 +3289,14 @@ static inline void schedule_debug(struct
* Julien Thierry wrote:
> I guess I'll drop the might_resched() part of this patch if that sounds
> alright.
That sounds perfect - that bit was pretty much the only problem I had
with the series.
Thanks,
Ingo
On 11/02/2019 13:51, Peter Zijlstra wrote:
> On Mon, Feb 11, 2019 at 02:45:27PM +0100, Ingo Molnar wrote:
>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>>> index a674c7db..b1bb7e9 100644
>>> --- a/kernel/sched/core.c
>>> +++ b/kernel/sched/core.c
>>> @@ -3289,6 +3289,14 @@ static
On Mon, Feb 11, 2019 at 02:45:27PM +0100, Ingo Molnar wrote:
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index a674c7db..b1bb7e9 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -3289,6 +3289,14 @@ static inline void schedule_debug(struct task_struct
>
* Julien Thierry wrote:
> While running a user_access regions, it is not supported to reschedule.
> Add an overridable primitive to indicate whether a user_access region is
> active and check that this is not the case when calling rescheduling
> functions.
>
> These checks are only performed
On 30/01/2019 16:58, Valentin Schneider wrote:
> On 15/01/2019 13:58, Julien Thierry wrote:
> [...]> @@ -6151,6 +6159,20 @@ void ___might_sleep(const char *file, int line,
> int preempt_offset)
>> EXPORT_SYMBOL(___might_sleep);
>> #endif
>>
>> +#ifdef CONFIG_DEBUG_UACCESS_SLEEP
>> +void
On 15/01/2019 13:58, Julien Thierry wrote:
[...]> @@ -6151,6 +6159,20 @@ void ___might_sleep(const char *file, int line,
int preempt_offset)
> EXPORT_SYMBOL(___might_sleep);
> #endif
>
> +#ifdef CONFIG_DEBUG_UACCESS_SLEEP
> +void __might_resched(const char *file, int line)
> +{
> + if
While running a user_access regions, it is not supported to reschedule.
Add an overridable primitive to indicate whether a user_access region is
active and check that this is not the case when calling rescheduling
functions.
These checks are only performed when DEBUG_UACCESS_SLEEP is selected.
25 matches
Mail list logo
