| --- Log opened Sun Jul 26 00:00:32 2015 | ||
| -!- Netsplit *.net <-> *.split quits: stekern, knz_, mithro | 11:05 | |
| andrzejr | I have run some simulations of byte and half-word access - there seem to be multiple bugs canceling each other. | 20:44 |
|---|---|---|
| andrzejr | stekern_, assuming l.lbz r7.0x1003(r3), what should wb_adr and wb_sel look like? I was expecting 0x????1000 and 0x8 respectively, but mor1kx issues 0x????1003 and 0x1. Are bus slaves expected to deal with unaligned access after all? | 20:52 |
| andrzejr | wb_interconn is (partially) dealing with that but the results is still imo incorrect: l.lwz r7,0x1000(r3) and l.lbz r7,0x1003(r3) result in different values in r7 | 21:06 |
| andrzejr | e.g. 0x12000000 and 0x00000012 respectively. Which one is correct? | 21:08 |
| --- Log closed Mon Jul 27 00:00:33 2015 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!