[Maverick] [ti-omap4] SRU: A workaround for highmem issue on OMAP4 platform

[Nicolas Pitre]

On Fri, 1 Oct 2010, Ricardo Salveti de Araujo wrote:

> From the talk I had with Rob Clark:
> 
> <robclark> ndec / rsalveti: which errata?
> <rsalveti> robclark: PL310_ERRATA_588369
> <robclark> ok.. let me check
> <robclark> rsalveti: ok.. this one seems to apply to the cortex-a9 IP, prior to r2p0
> <robclark> (just need to check which rev on cortex is in es2.0)
> <robclark> r1p2.. so keep the errata enabled
> 
> So it seems we're OK with it.
> 
> And during my tests I was finally able to run for more than 28 hours (14 builds)
> without getting any errors, so it seems that it's finally better.
> 
> But unfortunatelly I can easily get errors when using highmem, no matter what
> setup I'm using, so probably a different issue.
> 
> Good thing is that it seems we finally got a workaround to use 1GB at
> Maverick.

Beware.  While the 2G:2G is a good workaround for the time being, it may 
cause problems to user space applications expecting a larger vm space to 
their disposal.  Good testing with large programs is recommended.  Or at 
least add this caveat to the release notes.


Nicolas