* Peter Ujfalusi <peter.ujfal...@ti.com> [160322 01:54]: > Tony, > > On 03/21/16 22:21, Tony Lindgren wrote: > > Hi, > > > > * Peter Ujfalusi <peter.ujfal...@ti.com> [160321 01:39]: > >> > >> This is also interesting: > >> McBSP2 sidetone is in region 39 and 40 (module and L4 interconnect) which > >> is > >> unique in case of OMAP34xx and OMAP35xx, but it is overlapping with GPIO6 > >> on > >> OMAP36xx. Not sure what are the implications. > > > > Hmm GPIO6 is in a different L4 segment though? Maybe > > you did not account for the segments? > > > > 0x49000000 + 0x20000 + 0x2000/0x4000/0x6000 for McBSP > > 0x49000000 + 0x50000 + 0x8000 for GPIO6 > > > > Or maybe I don't understand at which physical address the > > overlap is? :) > > The addresses are not overlapping, but the Region Numbers for GPIO6 and McBSP2 > sidetone in Table 9-114 "Region Allocation for > L4-Per Interconnect". But only in case of OMAP36xx
Oh OK, yeah the TRM region numbers are wrong, they skip unused entries compared to the hardware AP table :) Basically the TRM AP region numbering is useless and wrong. Regards, Tony
