local_irq_save vs raw_local_irq_save

[anish singh]

On Jul 5, 2013 1:20 PM, "Raj Ravi" <mekaviraj at gmail.com> wrote:
>
> Hi,
> I have gone through the code.
>
> Please correct / provide your insights on my understanding.
> 1. There is no difference between local_irq_save and raw_local_irq_save
with ARM architecture.  ?
Yes,mostly there is no difference except in
local_irq_save we enable some tracing options.
>
> 2. As per the below code,
> <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
>  13 static inline unsigned long arch_local_irq_save(void)
>  14 {
>  15         unsigned long flags;
>  16
>  17         asm volatile(
>  18                 "       mrs     %0, cpsr        @
arch_local_irq_save\n"
>  19                 "       cpsid   i"
>  20                 : "=r" (flags) : : "memory", "cc");
>  21         return flags;
>  22 }
> <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
> Here, Flags contain the result of PRIMASK register. before interrupts are
disabled ??
Well that is assembly and my weakness
so my best answer would be that it just stores the previous state of
interrupt status register and we restore it back when we call
restore function.
>
> References:
> About CPSID instruction :
>
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0662b/BABHBAAB.html
> CPSID i ; Disable all interrupts except NMI (set PRIMASK.PM)
> Registers overview :
>
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0662b/CHDBIBGJ.html
> About PRIMASK register:
>
http://infocenter.arm.com/help/topic/com.arm.doc.dui0662b/CHDBIBGJ.html#BABBBGEA
> ARM GCC inline assembly:
> http://www.ethernut.de/en/documents/arm-inline-asm.html
>
> Please provide your insights.
>
> Thanks !
> Kavi
>
>
>
> On Thu, Jul 4, 2013 at 4:38 PM, anish singh <anish198519851985 at gmail.com>
wrote:
>>
>> On Thu, Jul 4, 2013 at 3:19 PM, Raj Ravi <mekaviraj at gmail.com> wrote:
>> > Thanks Anish.
>> >
>> > As for  local_irq_save  and raw_local_irq_save, it seems
>> > local_irq_save - called at many different places while
raw_local_irq_save is
>> > called in fewer places.
>> >
>> >                              local_irq_save - does not disable Hardware
>> > interrupts.
>> >     Whereas,raw_local_irq_save - does DISABLE hardware interrupts.
>> >
>> > As per code, raw_local_irq_save is related with timer interrupts.
>> > Please provide your insights on this.
>> http://lxr.free-electrons.com/source/include/linux/irqflags.h#L61
>> read this please.
>> >
>> > Also, in local_irq_save - how exactly disabling interrupts is achieved
( in
>> > spite of " NOT DISABLING " hardware interrupts) ?
>> >
>> >
>> >
>> >
>> >
>> > On Thu, Jul 4, 2013 at 2:45 PM, anish singh <
anish198519851985 at gmail.com>
>> > wrote:
>> >>
>> >> On Thu, Jul 4, 2013 at 1:55 PM, Raj Ravi <mekaviraj at gmail.com> wrote:
>> >> > Hi,
>> >> >
>> >> > Please let me know the difference between
>> >> > local_irq_save  and raw_local_irq_save ?
>> >>
>> >> did you check who calls this raw_local_irq_save?
>> >> similarly who calls local_irq_save?
>> >> finding out these will answer your questions
>> >> > At which situation one should use raw_local_irq_save?
>> >> >
>> >> >
>> >> > Also, I read "flags"  is where interrupt state is saved while
disabling
>> >> > interrupts in local cpu (for local_irq_save).
>> >> >
>> >> > What EXACTLY this  flags contains?  What do you mean by interrupt
state
>> >> > here?
>> >>
>> >> interrupt state meaning: which all interrupts are enabled or disabled
at
>> >> this
>> >> instant.If a particular interrupt was enabled at that instant then we
>> >> don't
>> >> do any thing with our locking mechanism which causes it be be
disabled.
>> >> What we are doing with those flags is nothing but maintaining the
state.
>> >> >
>> >> > Thanks!
>> >> > Kavi
>> >> >
>> >> > _______________________________________________
>> >> > Kernelnewbies mailing list
>> >> > Kernelnewbies at kernelnewbies.org
>> >> > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>> >> >
>> >
>> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20130705/45465d3e/attachment.html