| stekern | yay, mor1kx now boots linux without doing "false" bus errors | 18:04 |
|---|---|---|
| stekern | i.e. the fetcher doesn't try to falsely access the next instruction after an exception have turned off the immu | 18:05 |
| stekern | juliusb: this means bus accesses on cappuccino will be a bit slower, but I think it's worth it, since we can then register them and up fmax a bit and it will (hopefully) simplify the cache logic | 18:06 |
| stekern | cache refills from bus should still be fast (and pipelined) though | 18:07 |
| stekern | I'm thinking we should have a seperate "ROM-port" for people that wants to run straight from fast on-chip RAM/ROM | 18:08 |
| stekern | completely bypassing the wb-logic | 18:08 |
| stekern | wb as in wishbone in this case | 18:11 |
| _franck_ | ll | 18:48 |
| juliusb | stekern: (regarding or1k-timer tests) I don't quite understand why that patch works,but I'll take your word for it ;) I recall there's a few tests with assumptions about timing of execution, so not surprised if a few things must be tweaked to egt them to work across several implementations | 20:01 |
| juliusb | stekern: I see, you put in a registering stage somewhere in the fetch logic which ensures that we don't fetch beyond an exception-causing-instruction? sounds good to me! :) | 20:06 |
| juliusb | regarding ROM port - I completely agree! It's a commonly done thing | 20:06 |
| juliusb | so I've been very busy lately at work | 20:06 |
| juliusb | looking at an in-house processor and a Cortex M3, sort of comparing them in various ways, so I've gotten a bit more familiar with the characteristics and implementation of embedded processors in low-power ASICs | 20:07 |
| juliusb | been very interesting | 20:07 |
| juliusb | it's very uncommon that you have much between your program code and the processor | 20:07 |
| juliusb | they are usually very tightly coupled | 20:07 |
| juliusb | olofk: hej! | 20:09 |
| olofk | juliusb: Hi! It's been a while | 20:10 |
| juliusb | too long | 20:10 |
| olofk | Definitely. I hope things settle down a bit here soon | 20:11 |
| olofk | So, what's up? Was anyone at fosdem? | 20:11 |
| juliusb | yeah, i'm back in the saddle this weekend looking to push along the mor1kx | 20:11 |
| juliusb | I didn't go :( | 20:11 |
| olofk | Good stuff. How's it coming along? | 20:12 |
| olofk | Got to run now. I'll be back later this evening | 20:12 |
| juliusb | yeah just getting back into it | 20:13 |
| juliusb | nps, hej da | 20:13 |
| juliusb | there's a good looking OSHUG on this week in London if anyone is nearby: http://oshug.org/event/23 | 20:19 |
| _franck_ | blueCmd: tdesc stuff pushed | 23:40 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!