On omap processor specs http://focus.ti.com/lit/an/sprt495/sprt495.pdf I found that it can run deep to 125 MHz @ 0.95 V but with conky program i see cpu freq low only at 250 MHz.
Can be an improvement to battery life if cpu freq down to 125 MHz when stand by. This can be a software ( kernel ) limitation or is hardware related ?

stand by? Do you mean locked and screen off?

standby = no pulseaudio, no network traffic, no background programs no nothing.
even a static image on screen mean standby.

In standby the CPU switches completely off. In that state CPU frequency doesn't matter.

Should be moved to brainstorm. I would vote for this.

Edit:
I don't think the cpu shuts off completely in standby.

Yes. This is what it thought when I asked my previous question. Anyone can test this by running powertop while in standby.

Useless. As you can verify with powertop, the CPU spends most of the time (when the device is in "standby") in C4 state where the clock is not running.

See also bug 7116.

I vote for
Even if proc can reach halt state, I want also 125 MHz freq.
'standby' is not a happy word let say instead 'waiting input'

I think desktop proc state C4 don't fit on omap architecture

C4 on OMAP is not so deep as C6-C7 when they agree on base 0 or 1

Look closer on page 6 in my first post - 22 mW on audio decode !
and "More specifically, the OMAP3530 processor divides
its functional blocks into 18 power domains, each with its own switch. The PRCM can switch all the power
domains, but many of them can also be user-controlled."

and some definitions
C1 . MPU WFI + Core active
C2 . MPU WFI + Core inactive
C3 . MPU CSWR + Core inactive
C4 . MPU OFF + Core inactive
C5 . MPU CSWR + Core CSWR
C6 . MPU OFF + Core CSWR
C7 . MPU OFF + Core OFF

Powertop statistics don't fit omap power granularity

At some point it becomes more power efficient to complete the work faster, so the CPU can go back to sleep faster..