1gorka writes: 2--- 3I have userspace on the gumstix [xscale, not omap]. The problem that 4got me in trouble was that in lexception.s (or l.s), 5 6 MOVM.DB.W [R0-R14], (R13) 7 8works differently for this architecture (and probably for others, as 9it is unclear how it should behave by reading the arm specs). This 10happens only for kernel faults as the others (syscall, user faults) 11use MOVM.DB.W.S which uses the banked user registers. 12 13The problem is that in this arch the value of R13 saved is the value 14after R13 itself has been modified, whereas in the others (bitsy, 15pico...), it was the value before. Adding 4*15 to the stack before 16the RFE solves the problem. 17--- 18 19In fact, the 2005 ARM arch. ref. man. (ARM DDI 0100I) says, under STM (1), 20that if Rn appears in the set of registers (and isn't the first one) 21and .W is specified, the stored value of Rn is unpredictable. 22The arm v7-ar arch. ref. man. says such usage is obsolete. 23