cfi causes vm_fault on IXP435

Alex Vinogradovs avinogradovs at clearpathnet.com
Thu Feb 5 11:04:27 PST 2009


No worries, I am just testing the things out. This board in particular 
has been customized for my company, so
I was just wondering if it would work without hacking the kernel ;)

Alex.


Sam Leffler wrote:
> Well I didn't enable it for Cambria because it wasn't right :-)
>
> I know that at the very least the flash config on the 2358 is 32M but 
> I only map 16M so if everything else worked you couldn't access all 
> the memory.  But otherwise I hit this fault and haven't had time to 
> diagnose it.  Not sure when I'll get to it; would love to have some 
> help (once you get it mapped correctly you can add support to the cfi 
> driver to dynamically map 1M blocks so we don't have to map all of 
> flash).
>
> BTW since you tried this w/o asking you probably noticed I also just 
> committed support for the SrataFlash protection register.  It looks 
> like it might be working and I was just confused about the ability to 
> write the user segment multiple times.  The doc is a bit confusing and 
> seems to say this 64-bit segment is OTP (write once).  I think I need 
> to move the code that lets you write it under the CFI_AMEDANDDANGEROUS 
> option...
>
>    Sam
>



More information about the freebsd-arm mailing list