On 2 Apr, Jeremy Jackson wrote:
> [EMAIL PROTECTED] wrote:
>
>> FYI:
>>
>> We are at a stage in development (again) where I am not keeping up to
>> date with the CVS repository. Once through this phase, I will get
>> re-synced.
>>
>> I have one bit of code that others might be interested in. I have code
>> for implementing a linuxbios serial console command line. This has
>> proven to be extremely helpful to me.
>>
>> It is mostly getline() and some string functions. If anyone needs this
>> code and is interested in merging into a clean patch against the
>> current CVS, please let me know and I'll send the relevant files.
>
> I would like to see the code. Attachments are fine .tar.bz2 etc...
> I'm wondering if it is possible (or has been done) to do this
> without interrupts... useing UART's in polling mode should be
> possible. If so combining Eric's L1 d-cache stack/data
> seg method couldn't we build a stripped down debugger
> like gdb/kgdb that will run only from rom? (a big one - 512k and
> up) and without setting up PIC's etc.
>
> I'm remembering that Alpha's SRM console allows reading/writing
> registers.... like maybe we could do interactive SDRAM setup?
> someone mentioned putting memtest86 into linuxbios...
>
> so the scenerio could be, system comes up with serial console
> right away, and can see register contents, change them, and
> test sdram for correct operation (timing, address wrap,etc)
> , interactively, without crashing
> from improper/no ram setup. Wouldn't this eliminate the
> code-burn-test-crash-recode cycle for pesky SDRAM code
> for each chipset? With a three second reboot in the worst case
> (hardware lockup), wouldn't this accelerate SDRAM and other
> development?
>
> I realize this may not be completely coherent. Comments?
Several people have said that they are interested in this code,
including Ron. To help avoid ending up with 37 different versions of
the same basic capability, I have sent the files to Ron who said we
would work it into the main linuxbios cvs respository.
I will let Ron let people know what he is doing with what I sent him.
The code uses no interrupts and is really quite simple.
> Regards,
>
> Jeremy
>
> PS - can someone tell me where the latest BX code lives?
> Someone's private workspace, or CVS? I have a BX board now
> and a DOC...
The BX code is in CVS. The only thing missing is SDRAM init. I use a
hard coded SDRAM init for my embedded system. Since SDRAM init is a
major PITA, I am using the term 'only' rather loosely here.
I keep hearing rumors of someone working on SDRAM init for the BX but
haven't heard much more than that.
Ty
--
Tyson D Sawyer iRobot Corporation
Senior Systems Engineer Real World Interface Div.
[EMAIL PROTECTED] Robots for the Real World
603-532-6900 ext 206 http://www.irobot.com
