| --- Log opened Wed Sep 09 00:00:36 2015 | ||
| andrzejr | Is this intentional: https://github.com/olofk/or1k_bootloaders/blob/master/spi_uimage_loader.S#L94 ? | 01:00 |
|---|---|---|
| olofk | andrzejr: That is supposed to be a startup-delay to give the Flash some time to initialize | 06:53 |
| dalias | is or1k stuff ever going upstream in gcc? | 06:55 |
| andrzejr | olofk, I meant the "1"<->"1b" label mismatch. Or is it a feature I am not aware of? | 07:34 |
| dalias | it's a feature you're unaware of | 07:37 |
| dalias | numeric labels are not unique and can't be used as targets themselves | 07:37 |
| dalias | 1b means "1-back" and refers to the previous (or current-location) 1: | 07:38 |
| dalias | 1f means "1-forward" and refers to the next 1: | 07:38 |
| andrzejr | dalias, thanks. One more question, what is the benefit of numeric labels in this context? I thought they would normally be used for annotating absolute addresses in de-assembled code. | 07:47 |
| dalias | the biggest advantage is that they're usable in inline asm | 07:48 |
| dalias | where the asm block mody may be duplicated many times | 07:48 |
| andrzejr | no idea about upstreaming gcc mods, I also would like to know the answer to that. | 07:48 |
| dalias | they also avoid producing label symbol spam without the need for target-specific knowledge of what form to use for local labels | 07:49 |
| andrzejr | I see, so it is a kind of a "local" label, this may be useful indeed. Have you seen any documentation of or1k assembler (going beyond the instruction set)? | 07:52 |
| dalias | no | 07:53 |
| olofk | dalias: blueCmd_ asked for an update from the last contributor yesterday | 09:10 |
| olofk | regarding gcc | 09:10 |
| latif | hi all... I want to disable the data cache before compling the linux kernel?? there examples for arm but not for openrisc.. Anyone knows?? | 12:54 |
| latif | How can I disable the data cache or instruction cache?? | 12:55 |
| _franck__ | latif: https://github.com/openrisc/orpsoc-cores/blob/master/systems/atlys/rtl/verilog/orpsoc_top.v#L322 | 13:00 |
| _franck__ | "...before compiling the linux kernel" ?? | 13:01 |
| latif | franck: as you know, I still can not run the kernel on atlys.. I just want to try disabling the data cache... that is the story. | 13:05 |
| latif | franck: I am using orpsoc-v2.. where can I do this in v2.?. the top module again?? | 13:07 |
| _franck__ | I think it is in some or1200-defines.v file or something | 13:13 |
| _franck__ | I can't recall | 13:13 |
| _franck__ | latif: https://github.com/rkrajnc/orpsocv2/blob/master/boards/xilinx/atlys/rtl/verilog/include/or1200_defines.v#L134 | 13:14 |
| latif | franck: thanks. I have just found it. | 13:31 |
| latif | hi all.. as some of you know, I have been working on booting linux kernel on atlys board.. As I said before the kernel was stopping by the time it is loading.. I have tried many things to undestand this but I could not... and rigth now I have managed to boot it.. I have just disable the data cache on or1200defines.v and then I tried again with this regenerated bit file..and it is done..working!!! | 16:10 |
| latif | but I dont understand... Why?? Why is it related with the data cache?? Should be a logical explanation.. Any idea?? | 16:12 |
| andrzejr | latif, does firmware invalidate the cache correctly before turning it on? The procedure is not as straightforward as it seems | 18:42 |
| --- Log closed Thu Sep 10 00:00:01 2015 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!