On Mon, Feb 23, 2015 at 06:43:23PM +0100, Jan Kiszka wrote:
> On 2015-02-23 18:21, Gilles Chanteperdrix wrote:
> > On Mon, Feb 23, 2015 at 06:12:50PM +0100, Gilles Chanteperdrix wrote:
> >> On Mon, Feb 23, 2015 at 06:01:54PM +0100, Philippe Gerum wrote:
> >>> On 02/23/2015 05:55 PM, Gilles Chanteperdrix wrote:
> >>>> On Mon, Feb 23, 2015 at 05:50:13PM +0100, Jan Kiszka wrote:
> >>>>> On 2015-02-23 17:37, Gilles Chanteperdrix wrote:
> >>>>>> On Mon, Feb 23, 2015 at 05:32:56PM +0100, Philippe Gerum wrote:
> >>>>>>> On 02/23/2015 05:06 PM, Jan Kiszka wrote:
> >>>>>>>> On 2015-02-20 20:52, Philippe Gerum wrote:
> >>>>>>>>> On 02/20/2015 08:47 PM, Philippe Gerum wrote:
> >>>>>>>>>> On 02/20/2015 08:44 PM, Philippe Gerum wrote:
> >>>>>>>>>>> On 02/20/2015 07:17 PM, Jan Kiszka wrote:
> >>>>>>>>>>>> On 2015-02-20 19:03, Jan Kiszka wrote:
> >>>>>>>>>>>>> Hi Gilles,
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> analyzing a lockdep warning on 3.16 with I-pipe enabled, I dug
> >>>>>>>>>>>>> deeper
> >>>>>>>>>>>>> into the hard and virtual interrupt state management during
> >>>>>>>>>>>>> exception
> >>>>>>>>>>>>> handling on ARM. I think there are several issues:
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> - ipipe_fault_entry should not fiddle with the root irq state
> >>>>>>>>>>>>> if run
> >>>>>>>>>>>>> over head, only when invoked over root.
> >>>>>>>>>>>>> - ipipe_fault_exit must not change the root state unless we
> >>>>>>>>>>>>> entered over
> >>>>>>>>>>>>> head and are about to leave over root - see x86. The current
> >>>>>>>>>>>>> code may
> >>>>>>>>>>>>> keep root incorrectly stalled after an exception, though this
> >>>>>>>>>>>>> will
> >>>>>>>>>>>>> probably be fixed up again in practice quickly.
> >>>>>>>>>>>>
> >>>>>>>>>>>> And the adjustment of the root irq state after migration has to
> >>>>>>>>>>>> happen
> >>>>>>>>>>>> before Linux starts to handle the event. It would basically be a
> >>>>>>>>>>>> late
> >>>>>>>>>>>> ipipe_fault_entry.
> >>>>>>>>>>>>
> >>>>>>>>>>>>> - do_sect_fault is only called by do_DataAbort and
> >>>>>>>>>>>>> do_PrefetchAbort,
> >>>>>>>>>>>>> in both cases already wrapped in ipipe_fault_entry/exit, thus
> >>>>>>>>>>>>> it
> >>>>>>>>>>>>> shouldn't invoke them once again.
> >>>>>>>>>>>>
> >>>>>>>>>>>> Sorry, this was a misinterpretation - do_sect_fault is invoked
> >>>>>>>>>>>> before
> >>>>>>>>>>>> ipipe_fault_entry.
> >>>>>>>>>>>>
> >>>>>>>>>>>> What I need to add, though:
> >>>>>>>>>>>>
> >>>>>>>>>>>> - do_DataAbort and do_PrefetchAbort call __ipipe_report_trap
> >>>>>>>>>>>> after
> >>>>>>>>>>>> ipipe_fault_entry, thus with hard IRQs on.
> >>>>>>>>>>>
> >>>>>>>>>>> This would break LPAE with the Xenomai nucleus as a module on
> >>>>>>>>>>> 2.6.x, by
> >>>>>>>>>>> treading over a non-linear kernel mapping before the page table
> >>>>>>>>>>> could be
> >>>>>>>>>>> fixed up. do_translation_fault() must run via the fsr handler
> >>>>>>>>>>> indirection before any non-linear access.
> >>>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> Sorry, if you do that _after_ the fault entry notification, then
> >>>>>>>>>> it's ok
> >>>>>>>>>> in theory. However, I don't understand why we would need to notify
> >>>>>>>>>> when
> >>>>>>>>>> only a minor fixup is required, that does not entail a mode
> >>>>>>>>>> migration.
> >>>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> To be clearer, do you intend to report the minor fault upon
> >>>>>>>>> do_translation_fault() returning zero, or are you referring to a
> >>>>>>>>> different context?
> >>>>>>>>
> >>>>>>>> No, I'm just talking about this potential change:
> >>>>>>>>
> >>>>>>>> diff --git a/arch/arm/mm/fault.c b/arch/arm/mm/fault.c
> >>>>>>>> index 38834c6..b42632a 100644
> >>>>>>>> --- a/arch/arm/mm/fault.c
> >>>>>>>> +++ b/arch/arm/mm/fault.c
> >>>>>>>> @@ -629,10 +629,10 @@ do_DataAbort(unsigned long addr, unsigned int
> >>>>>>>> fsr, struct pt_regs *regs)
> >>>>>>>> if (!inf->fn(addr, fsr & ~FSR_LNX_PF, regs))
> >>>>>>>> return;
> >>>>>>>>
> >>>>>>>> - irqflags = ipipe_fault_entry();
> >>>>>>>> -
> >>>>>>>> if (__ipipe_report_trap(IPIPE_TRAP_UNKNOWN, regs))
> >>>>>>>> - goto out;
> >>>>>>>> + return;
> >>>>>>>> +
> >>>>>>>> + irqflags = ipipe_fault_entry();
> >>>>>>>>
> >>>>>>>> printk(KERN_ALERT "Unhandled fault: %s (0x%03x) at 0x%08lx\n",
> >>>>>>>> inf->name, fsr, addr);
> >>>>>>>> @@ -642,7 +642,7 @@ do_DataAbort(unsigned long addr, unsigned int
> >>>>>>>> fsr, struct pt_regs *regs)
> >>>>>>>> info.si_code = inf->code;
> >>>>>>>> info.si_addr = (void __user *)addr;
> >>>>>>>> arm_notify_die("", regs, &info, fsr, 0);
> >>>>>>>> -out:
> >>>>>>>> +
> >>>>>>>> ipipe_fault_exit(irqflags);
> >>>>>>>> }
> >>>>>>>>
> >>>>>>>> @@ -669,10 +669,10 @@ do_PrefetchAbort(unsigned long addr, unsigned
> >>>>>>>> int ifsr, struct pt_regs *regs)
> >>>>>>>> if (!inf->fn(addr, ifsr | FSR_LNX_PF, regs))
> >>>>>>>> return;
> >>>>>>>>
> >>>>>>>> - irqflags = ipipe_fault_entry();
> >>>>>>>> -
> >>>>>>>> if (__ipipe_report_trap(IPIPE_TRAP_UNKNOWN, regs))
> >>>>>>>> - goto out;
> >>>>>>>> + return;
> >>>>>>>> +
> >>>>>>>> + irqflags = ipipe_fault_entry();
> >>>>>>>>
> >>>>>>>> printk(KERN_ALERT "Unhandled prefetch abort: %s (0x%03x) at
> >>>>>>>> 0x%08lx\n",
> >>>>>>>> inf->name, ifsr, addr);
> >>>>>>>> @@ -682,7 +682,7 @@ do_PrefetchAbort(unsigned long addr, unsigned
> >>>>>>>> int ifsr, struct pt_regs *regs)
> >>>>>>>> info.si_code = inf->code;
> >>>>>>>> info.si_addr = (void __user *)addr;
> >>>>>>>> arm_notify_die("", regs, &info, ifsr, 0);
> >>>>>>>> -out:
> >>>>>>>> +
> >>>>>>>> ipipe_fault_exit(irqflags);
> >>>>>>>> }
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> This seems more consistent - if not more correct - as it now does the
> >>>>>>>> reporting with hard irqs off, like in the other cases.
> >>>>>>>>
> >>>>>>>
> >>>>>>> Ack, definitely. The pattern is to cause any migration first if need
> >>>>>>> be,
> >>>>>>> _then_ flip the virtual IRQ state, so that ipipe_fault_restore()
> >>>>>>> always
> >>>>>>> reinstates the interrupt state in effect after the caller has migrated
> >>>>>>> to the root domain.
> >>>>>>
> >>>>>> Is it even useful ? After a relax, the state of the root thread
> >>>>>> stall bit and irq flags are well known...
> >>>>>
> >>>>> We still need to disable IRQs for root. HW IRQs are likely already on,
> >>>>> right?
> >>>>>
> >>>>> And, again, we should refrain from restoring any root irq state on
> >>>>> return - it belongs to Linux (once we migrated and synchronized the
> >>>>> state).
> >>>>
> >>>> The ipipe_fault_exit in my tree is:
> >>>>
> >>>> static inline void ipipe_fault_exit(unsigned long x)
> >>>> {
> >>>> if (!arch_demangle_irq_bits(&x))
> >>>> local_irq_enable();
> >>>> else
> >>>> hard_local_irq_restore(x);
> >>>> }
> >>>>
> >>>> And I must say I am not sure I understand how it works. To me it
> >>>> seems:
> >>>
> >>> It mangles both the real and virtual states in one word.
> >>>
> >>>> hard_local_irq_disable() should always be called in case entry.S
> >>>> expects us to return as we entered: with hw irqs off
> >>>
> >>> Which is what ipipe_fault_exit() does by testing the mangled state. If
> >>> the fault entered with virtual IRQs on, then you must exit with both the
> >>> stall bit and CPSR_I bit cleared.
> >>
> >> Absolutely not. Imagine a Linux task, with root unstalled
> >> experiencing a fault. entry.S is entered root is still unstalled,
> >> with hardware irqs off. On fault entry, we must reflect this
> >> hardware irq state on the stall bit and enable hw irqs. Then when
> >> the fault is handled, undo that, unstall the root stage, disable hw
> >> irqs and return to entry.S, so that it may resume the execution of
> >> the Linux task. If it returns quickly to user-space, a stalled root
> >> at this point would be a disaster, because nothing, certainly not
> >> entry.S will unstall the root stage.
> >>
> >> Then there is the case where Linux re-enables irqs in the course of
> >> the handler, in that case, we should not return to entry.S with hw
> >> irqs off.
> >
> > To summarize:
> >
> > when the fault is entered over head domain. The root stage can be
> > left unstalled on fault exit. The hw irqs state should reflect the
> > root stage stall bit state at exit.
>
> On x86, we pick up the desired state of hw and root after migration to
> root completed.
>
> >
> > when the fault is entered over root domain. The root stage stall bit
> > must be restorer as it was on entry, and the hw irqs state should
> > reflect the root stage stall bit state at exit.
>
> Right. Actually, it should be restored to the state in the register set
> on return from the Linux handler.
No, the fault return path will take care of that by restoring the
saved cpsr. What is important is that if an irq return path assumes
that hw irqs are disabled to start pulling registers and playing
with sp, we must have disabled the hw irqs before returning to it,
to avoid register corruption if an irq happens at the bad time. On
the other hand, if Linux called local_irq_enable(), then entry.S
will not assume that hw irqs are off, and we can avoid the
hard_local_irq_disable() before returning from the fault, so as to
avoid creating a useless masking section (which is probably the
reason why Linux calls local_irq_enable()).
> But as Linux does not seem to
> manipulate that state, we are fine with restoring to the entry state.
Well, Linux seems to call local_irq_enable() in one of the fault
handlers.
--
Gilles.
_______________________________________________
Xenomai mailing list
[email protected]
http://www.xenomai.org/mailman/listinfo/xenomai
