| --- Log opened Thu Nov 05 00:00:56 2015 | ||
| _franck_ | andrzejr: olofk did a pull request to add this: https://github.com/openrisc/linux/pull/5 | 06:20 |
|---|---|---|
| wallento | hi andrezejr | 07:57 |
| wallento | no, andrzejr | 07:58 |
| wallento | sorry | 07:58 |
| wallento | we have some debug stuff for the trace ports. sorry I did not follow the entire discussion | 07:58 |
| bandvig | Hello all! | 08:34 |
| bandvig | stekern: IMMU and DMMU contain HW-reload logic. However, there isn't support the feature in software, am I right? | 08:36 |
| stekern | no, software reload is the default | 08:36 |
| bandvig | So, HW-reload support could be activated. By the way, how? | 08:39 |
| stekern | there's some parameter, which I can't remember the name of ;) | 08:40 |
| stekern | ah, now I understand your first question, you didn't ask about sw reload of the tlbs | 08:40 |
| stekern | you asked for sw support of the hw reload | 08:41 |
| bandvig | exactly | 08:41 |
| stekern | yeah, you need some patches for linux, I have them in some branch somewhere | 08:41 |
| stekern | https://github.com/skristiansson/linux/tree/pte-wip | 08:42 |
| stekern | there (I hope) ;) | 08:42 |
| bandvig | ok, thanks | 08:43 |
| olofk | juliusb_: I think ghdl also uses LLVM internally | 10:14 |
| olofk | ghdl is more advanced. Supports most of VHDL2008 IIRC. One drawback is that it's written in ADA | 10:15 |
| olofk | So... do we have a cache problem in Linux? | 11:25 |
| olofk | https://github.com/openrisc/linux/pull/6 | 11:25 |
| stekern | olofk: yes, I've spoke to the guy in private mails | 11:51 |
| stekern | I don't quite get how we have not hit any bugs related to that before though | 11:51 |
| olofk | stekern: That's what I was thinking as well. Looked like a severe bug | 12:24 |
| olofk | But there has been some talks about cache problems here recently | 12:25 |
| olofk | Maybe some other linux changes that uncovered this? | 12:25 |
| -!- Netsplit *.net <-> *.split quits: alex-guo, rhythmx, sfa | 17:04 | |
| -!- alex-guo_ is now known as alex-guo | 17:04 | |
| -!- jn___ is now known as jn__ | 17:42 | |
| andrzejr | olofk, I have checked this patch with my hardware with memory bursts enabled - same symptoms as without the patch (kernel crashes or gets stuck during boot and there is no output to the console) | 20:19 |
| andrzejr | since my issue is solved by disabling bursting in the DDR2 WB slave it is most likely hw related. | 20:21 |
| --- Log closed Fri Nov 06 00:00:57 2015 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!