BogoMIPS change in 3.6+

BogoMIPS change in 3.6+

[Rob Herring]

The BogoMIPS value has changed on highbank since the commit below in
3.6. It is now half the value it used to be. Does that matter other than
perception (I've checked that I'm not running at half the freq)?

Rob

commit d0a533b18235d36206b9b422efadb7cee444dfdb
Author: Will Deacon <will.deacon@arm.com>
Date:   Fri Jul 6 15:47:17 2012 +0100

    ARM: 7452/1: delay: allow timer-based delay implementation to be
selected

    This patch allows a timer-based delay implementation to be selected by
    switching the delay routines over to use get_cycles, which is
    implemented in terms of read_current_timer. This further allows us to
    skip the loop calibration and have a consistent delay function in the
    face of core frequency scaling.

    To avoid the pain of dealing with memory-mapped counters, this
    implementation uses the co-processor interface to the architected timers
    when they are available. The previous loop-based implementation is
    kept around for CPUs without the architected timers and we retain both
    the maximum delay (2ms) and the corresponding conversion factors for
    determining the number of loops required for a given interval. Since the
    indirection of the timer routines will only work when called from C,
    the sa1100 sleep routines are modified to branch to the loop-based delay
    functions directly.

BogoMIPS change in 3.6+

[Arnd Bergmann]

On Thursday 31 January 2013 10:15:11 Rob Herring wrote:
> The BogoMIPS value has changed on highbank since the commit below in
> 3.6. It is now half the value it used to be. Does that matter other than
> perception (I've checked that I'm not running at half the freq)?

I think the only danger is that your customers might think they
are running at half the clock rate. Nothing in the kernel should
be using the BogoMIPS value.

	Arnd

BogoMIPS change in 3.6+

[Will Deacon]

Hi Rob,

On Thu, Jan 31, 2013 at 04:15:11PM +0000, Rob Herring wrote:
> The BogoMIPS value has changed on highbank since the commit below in
> 3.6. It is now half the value it used to be. Does that matter other than
> perception (I've checked that I'm not running at half the freq)?

Are you registering a delay timer? If not, then the delay loop should be
identical to what it was before, so that's certainly a surprise...

Are your delays ballpark correct (try 2ms)?

Will

BogoMIPS change in 3.6+

[Santosh Shilimkar]

On Thursday 31 January 2013 09:45 PM, Rob Herring wrote:
> The BogoMIPS value has changed on highbank since the commit below in
> 3.6. It is now half the value it used to be. Does that matter other than
> perception (I've checked that I'm not running at half the freq)?
>
I also stumbled on this while testing Will's series.[1]
There should be no other functional issue apart from getting surprised.

Regards,
Santosh

[1] 
http://lists.infradead.org/pipermail/linux-arm-kernel/2012-July/109281.html

BogoMIPS change in 3.6+

[Will Deacon]

On Thu, Jan 31, 2013 at 04:24:56PM +0000, Santosh Shilimkar wrote:
> On Thursday 31 January 2013 09:45 PM, Rob Herring wrote:
> > The BogoMIPS value has changed on highbank since the commit below in
> > 3.6. It is now half the value it used to be. Does that matter other than
> > perception (I've checked that I'm not running at half the freq)?
> >
> I also stumbled on this while testing Will's series.[1]
> There should be no other functional issue apart from getting surprised.

Right, but you were using the arch-timers w/ Cortex-A15. I thought Highbank
was A9?

Will

BogoMIPS change in 3.6+

[Santosh Shilimkar]

On Thursday 31 January 2013 09:55 PM, Will Deacon wrote:
> On Thu, Jan 31, 2013 at 04:24:56PM +0000, Santosh Shilimkar wrote:
>> On Thursday 31 January 2013 09:45 PM, Rob Herring wrote:
>>> The BogoMIPS value has changed on highbank since the commit below in
>>> 3.6. It is now half the value it used to be. Does that matter other than
>>> perception (I've checked that I'm not running at half the freq)?
>>>
>> I also stumbled on this while testing Will's series.[1]
>> There should be no other functional issue apart from getting surprised.
>
> Right, but you were using the arch-timers w/ Cortex-A15. I thought Highbank
> was A9?
>
Ahh.. Sorry looking at description I just thought it could be the same
issue.

Regards,
Santosh

BogoMIPS change in 3.6+

[Rob Herring]

On 01/31/2013 10:23 AM, Will Deacon wrote:
> Hi Rob,
> 
> On Thu, Jan 31, 2013 at 04:15:11PM +0000, Rob Herring wrote:
>> The BogoMIPS value has changed on highbank since the commit below in
>> 3.6. It is now half the value it used to be. Does that matter other than
>> perception (I've checked that I'm not running at half the freq)?
> 
> Are you registering a delay timer? If not, then the delay loop should be
> identical to what it was before, so that's certainly a surprise...
> 
> Are your delays ballpark correct (try 2ms)?
> 

Yes, the delays are correct and this is an A9. So I guess this may be a
surprise, but seems it is a non-issue. I'm sure I'll be getting asked
about this...

Rob