On Thu, Oct 22, 2015 at 01:30:42PM +0200, Jens Kuske wrote: > On 22/10/15 11:14, Maxime Ripard wrote: > > On Thu, Oct 22, 2015 at 10:57:45AM +0200, Jean-Francois Moine wrote: > >> On Thu, 22 Oct 2015 10:47:35 +0200 > >> Maxime Ripard wrote: > >> > >>> Not really. The uart0 reset is the bit 16, in the reset register 4. > >>> > >>> 4 * 32 + 16 = 44. > >>> > >>> Not 112, but still not 208 either. > >> > >> The registers are numbered 1..5, then > >> > >> (4 - 1) * 32 + 16 = 112 > > > > Not on my version, and even then, UARTs are on the last reset > > register, which would still make 144. > > > > Maxime > > > > There are holes between reg2 and reg3 and reg4 for some reason, but even > if we would correct that with some of_xlate() function they won't > completely line up with the gates. Indeed. Still, dealing with the holes and sticking to what the datasheet says seems like the right solution. Maxime -- Maxime Ripard, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com