|
Message
From: Jeremy Bennett<jeremy.bennett@e...>
Date: Sun Aug 24 23:41:58 CEST 2008
Subject: [openrisc] linux 2.6.19 problem with or1ksim according to
embecosm or1k_setup.pdf
On Sun, 2008-08-24 at 15:07 +0800, Ding Xiaoliang wrote:> please execuse me for posting this to the big list.
>
> I've tried to compile the or32 toolchain, simulator and linux kernel
> according to the document shows.
> The problem I met:
> * When try to compile linux 2.6.19 kernel, it seems some "kallsyms"
> symbols are missed, so that the compilation failed.
> * The linux configuration (.config) and compilation error log has been
> attached (linux_compile.log)
>
> I then deselected the CONFIG_KALLSYMS and CONFIG_MODULES, the
> compilation can continue with some warnings.
>
> * when try "make check" in or1ksim/testbench, the fb_test failed , while
> other tests passed
>
> Then, I try to boot linux with the or1ksim, but failed with a
> "Segmentation Fault" message.
> * The or1ksim outputlog (sim.log) has been
> attached
Hi Xiaoliang,
Please do post to this list - the more questions that are asked, the
more we will know.
What platform are you running on (Linux version, 32 or 64 bit)?
Please would you post a transcript of each command you used when
building the tool chain in the sequence you used them (include all the
commands to unzip tar files, patch sources etc). This will help work out
where the problem is occurring.
I think it is very unlikely that Linux will work if you deselect these
modules, but you should not get a segmentation fault in Or1ksim.
You can ignore the failure with "make check" in or1ksim/testbench. The
test bench does not have a very reliable check for success. It sometimes
checks for success before the test has finished, and reports a failure.
This often occurs with the fb test.
Hope this helps,
Jeremy
--
Tel: +44 (1202) 416955
Cell: +44 (7970) 676050
SkypeID: jeremybennett
Email: jeremy.bennett@e...
Web: www.embecosm.com
|
 |
Browse
