| _franck_ | juliusb: , stekern : as you know very weel the or1k, you should save me some time: | 00:03 |
|---|---|---|
| _franck_ | http://picpaste.com/pics/problem-i3gxqjCM.1348866148.png | 00:03 |
| _franck_ | could you tell me why I can't the instruction at address 0x01fc4f6c ? am I missing something ? | 00:04 |
| _franck_ | forgot "see" in my sentence... | 00:04 |
| stekern | at least one of the failures I see in the or1k-linux-gcc regressions doesn't seem to be our fault: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53476 | 06:40 |
| stekern | _franck_: from what I can tell from that waveform, icbiu is behaving badly (at least if it whould be considered wishboney). It is terminating the cycle without getting an ack. | 06:44 |
| stekern | IIRC that's not allowed. | 06:45 |
| stekern | hmm, no, you don't have the biu ack in the wave, do you? | 06:46 |
| stekern | I loocked at the icqmem_ack_o | 06:47 |
| stekern | -o | 06:48 |
| stekern | but the same logic can be applied to icqmem_cycstb_i, it goes low without an ack | 06:49 |
| stekern | so what is getting acked at "10" is the request that got initiated at "-5" | 06:51 |
| stekern | otoh, it seems to make use of the rty signal, I'm not sure how that is interpreted in this context though. | 06:57 |
| stekern | either case, that doesn't seem to work as it should at least | 06:58 |
| stekern | another gcc FAIL I can "write off", pr51106-2 fails here too: http://gcc.gnu.org/ml/gcc-testresults/2012-09/msg02896.html | 07:33 |
| stekern | I've put the testresults up here: | 09:16 |
| stekern | http://opencores.org/or1k/UClibc_tool_chain_test_results#GCC_4.8.0 | 09:16 |
| stekern | will update as progress is made | 09:17 |
| stekern | hmm, I wonder if it's safe to switch on DWARF2_ASM_LINE_DEBUG_INFO in or1k.h | 16:15 |
| stekern | there is a comment above it that our assembler doesn't support it, but that seems old | 16:16 |
| stekern | jeremybennett: any idea? you switched it off in 4.2.2 commit 251 2010-08-26 | 16:24 |
| stekern | I mean, was it not supported because the old as was too old or does it need some target specific implementation? | 16:43 |
| stekern | I can't at least find anything target specific in or1k-src, except in gdb | 16:43 |
| stekern | found this too: http://opencores.org/bug,view,1828 | 16:49 |
| chengcai | has or1200 a real product? | 18:27 |
| stekern | you will never know without patience | 19:29 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!
