Re: Duration of mdelay and udelay depends on MPU frequency in SMP

Re: Duration of mdelay and udelay depends on MPU frequency in SMP

[Menon, Nishanth]

On Thu, May 12, 2011 at 01:42, Santosh Shilimkar
<santosh.shilimkar@ti.com> wrote:
> On 5/5/2011 12:23 PM, Saquib wrote:
>> Can anyone please help me to understand and fix this issue?
>>
>> When CPU frequency is "1008MHz", duration of delay was close to
>> theoretical value, however, when "300MHz", it was close to about 4 times
>> the theoretical value.
>>
>> I tested it on OMAP4(SMP) blaze board.
>
>>
> It's a well known problem if the udelay() is based of global lpj.
> You can read more here [1]
>
> Regards
> Santosh
> [1] http://eeek.borgchat.net/lists/arm-kernel/msg120702.html
I am curious about this topic now.
Searching Russel's patchworks for this:
http://www.arm.linux.org.uk/developer/patches/search.php?summary=udelay
I see nothing queued. What is the recommendation for udelay?

Regards,
Nishanth Menon

Re: Duration of mdelay and udelay depends on MPU frequency in SMP

[Stephen Boyd]

On 06/10/2011 03:19 PM, Menon, Nishanth wrote:
> On Thu, May 12, 2011 at 01:42, Santosh Shilimkar
> <santosh.shilimkar@ti.com> wrote:
>>
>> It's a well known problem if the udelay() is based of global lpj.
>> You can read more here [1]
>>
>> Regards
>> Santosh
>> [1] http://eeek.borgchat.net/lists/arm-kernel/msg120702.html
> I am curious about this topic now.
> Searching Russel's patchworks for this:
> http://www.arm.linux.org.uk/developer/patches/search.php?summary=udelay
> I see nothing queued. What is the recommendation for udelay?

The patches are sitting in Russell's patch tracker, waiting for him to
accept/reject them.

http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=6874/1
http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=6875/1
http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=6873/1

-- 
Sent by an employee of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.

Re: Duration of mdelay and udelay depends on MPU frequency in SMP

[Menon, Nishanth]

On Mon, Jun 13, 2011 at 22:24, Stephen Boyd <sboyd@codeaurora.org> wrote:
> On 06/10/2011 03:19 PM, Menon, Nishanth wrote:
>> On Thu, May 12, 2011 at 01:42, Santosh Shilimkar
>> <santosh.shilimkar@ti.com> wrote:
>>>
>>> It's a well known problem if the udelay() is based of global lpj.
>>> You can read more here [1]
>>>
>>> Regards
>>> Santosh
>>> [1] http://eeek.borgchat.net/lists/arm-kernel/msg120702.html
>> I am curious about this topic now.
>> Searching Russel's patchworks for this:
>> http://www.arm.linux.org.uk/developer/patches/search.php?summary=udelay
>> I see nothing queued. What is the recommendation for udelay?
>
> The patches are sitting in Russell's patch tracker, waiting for him to
> accept/reject them.
>
> http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=6874/1
> http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=6875/1
> http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=6873/1

thanks for the same and hope these do get through :)

Regards,
Nishanth Menon