On 07/06/2012 08:30 AM, Richard Earnshaw wrote: > Could you check that this does the correct thing for > > ADR r0, .+4 > > Which should also generate encoding A2 Yes, it does. The above disassembles as 00000000 <foo> e24f0004 sub r0, pc, #4 which matches the A2 pattern: 1110 00 1 0010 0 1111 0000 000000000100 -Sandra