RE: [PATCH] Remove omap_sram_idle()

[Igor Stoppa <igor.stoppa@nokia.com>]

On Sun, 2007-12-16 at 08:44 -0600, ext Woodruff, Richard wrote:
> > 	This patch removes omap_sram_idle().
> 
> Why?
> 
> > Signed-off-by: Vivek Kutal <vivek.kutal@celunite.com>
> 
> Execution in SRAM on some OMAPs is needed and in others it provides a more likely path to hitting deep sleep states.  There is sensitivity of running in SDRAM when trying to shut this down.
> 
> On 1509 there was not alternative as sleeping when the SDR was not in self-refresh was impossible.  On 1510 an automated self-refresh was added based on an internal signal.
> 
> I've not tried an OMAP1 in a while so I'm not sure what issues there might be today.  There are tree's which do hit big and deep sleep and I think they still use something close to this.

Yup. Nokia 770 is one of them.
Or was, at least, I have to admit that I haven't checked an OMAP1 kernel
for a long time. But it used to work.

Vivek: did you make any measurement on your patch? It would be good if
you could share your findings.

-- 
Cheers, Igor

Igor Stoppa <igor.stoppa@nokia.com>
(Nokia Multimedia - CP - OSSO / Helsinki, Finland)