| --- Log opened Thu Apr 11 00:00:10 2013 | ||
| mor1kx | [mor1kx] skristiansson pushed 6 new commits to master: https://github.com/openrisc/mor1kx/compare/0674cac45545...9ed548e54621 | 05:41 |
|---|---|---|
| mor1kx | mor1kx/master 2a75b22 Stefan Kristiansson: cappuccino/ctrl: remove obsolete last_branch_target_pc calculation... | 05:41 |
| mor1kx | mor1kx/master 703b7de Stefan Kristiansson: cappuccino/ctrl: whitespace fix | 05:41 |
| mor1kx | mor1kx/master ad042d6 Stefan Kristiansson: cappuccino/ctrl: remove tt & pic exception prevention on branches... | 05:41 |
| olofk | stekern, juliusb: I have patches for mor1kx, but I'm not in a position to push stuff to a tree. Do you want them in an e-mail or posted to the list? | 10:57 |
| stekern | I reckon anything goes, but can't you push them to a 'throw-away-tree' (on your favorite git server) | 11:01 |
| stekern | ? | 11:01 |
| stekern | but if sent as e-mail patches, probably to the mailing lists are best | 11:02 |
| mor1kx | [mor1kx] skristiansson pushed 2 new commits to master: https://github.com/openrisc/mor1kx/compare/9ed548e54621...68121b19a6c0 | 11:14 |
| mor1kx | mor1kx/master 3d7ef8f Stefan Kristiansson: cappuccino/ctrl: remove unused 'decode_stage_exceptions' | 11:14 |
| mor1kx | mor1kx/master 68121b1 Stefan Kristiansson: cappuccino/execute_ctrl: refractor wb logic for load and mfspr... | 11:14 |
| stekern | bah, annoying typo in the commit msg | 11:15 |
| mor1kx | [mor1kx] skristiansson force-pushed master from 68121b1 to 86f0c99: https://github.com/openrisc/mor1kx/commits/master | 11:16 |
| mor1kx | mor1kx/master 86f0c99 Stefan Kristiansson: cappuccino/execute_ctrl: refactor wb logic for load and mfspr... | 11:16 |
| -!- Netsplit *.net <-> *.split quits: vdonnefort, _franck_, orsoc1, stekern | 14:17 | |
| -!- Netsplit over, joins: _franck_, orsoc1, stekern, vdonnefort | 14:19 | |
| stekern | hmm, I think I have an idea how to tame the fetcher, so we can become best buddies | 16:06 |
| stekern | 2 things make ir like a wild horse out of control, 1. it's expected that fetch_valid is kept high 2. fetch_valid can unstall execute stage without going through decode | 16:08 |
| stekern | s/ir/it | 16:09 |
| stekern | and 3. it pushes out the fetched instruction to decode even though padv_i isn't asserted | 16:10 |
| stekern | but 3. is a consequence of 1 and 2 | 16:10 |
| stekern | that kind of behaviour is probably the best way to go when you have no cache and very slow bus | 16:11 |
| stekern | but it creates a lot of corner cases | 16:12 |
| stekern | because I want to think about the stages as a registered bus, where the padv's are the read enable and the _valids are the acks | 16:13 |
| stekern | I take back what I said about unstalling execute, that it should be able to do | 19:28 |
| --- Log closed Fri Apr 12 00:00:11 2013 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!