> From: linux-omap-owner@xxxxxxxxxxxxxxx [mailto:linux-omap- > owner@xxxxxxxxxxxxxxx] On Behalf Of Felipe Contreras > Sent: Saturday, November 29, 2008 6:52 AM > On Thu, Nov 27, 2008 at 2:29 AM, Kevin Hilman > <khilman@xxxxxxxxxxxxxxxxxxx> wrote: > > This series enables UART clock disabling after an inactivity period. > > It is based on top of the 2 8250 patches recently sent to linux-serial > > and CC'd to linux-omap. > > > > To enable: > > > > # echo 1 > /sys/power/clocks_off_while_idle > > # echo 1 > /sys/power/sleep_while_idle > > > > NOTE: it is expected that the first character typed is lost > > when coming out of idle. The first char serves as the wakeup > > event but is lost. > > Agh, can that be avoided somehow? Not unless you use some kind of flow control or protocol retransmits. When it’s a raw serial port your start bit is your wakeup character. From OFF mode this is waking the system up at the IO Pad level. The system can't repower, respin up DPLLs and restore context fast enough to avoid losing the start bit. You could try running at some really slow baud rate. More practically if you use some flow control signals they may allow it to work with out loss. If you're bluetooth or the like your protocol retransmits so you're ok. If all is just so you might see this most pronounced when screen is blanked and you're in CORE OFF. At that point it may not be a big deal. Even on a PC do you expect your wake up character to do something when breaking from a blank screen or screen saver? It might be some automated expect scripts or the like need tweaking. We have been using a similar strategy for a couple years with out huge impact. Regards, Richard W. ��.n��������+%������w��{.n�����{�������ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
