[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [microblaze-uclinux] BRAM
I'm not sure I see the advantage of doing this versus using the BRAM as
cache???
Paul
Quoting Bret Ketchum <bret.ketchum@xxxxxxxxxxxxxxxxxxx>:
>
> There must be a file, say:
>
> ./linux-2.4.x/arch/microblaze/mbvanilla.ld
>
> that describes to the linking process the base address of
> interesting memory regions such as the area to use for stack. It just
> isn't
> intuitive and I haven't taken the time to look.
>
> Bret
>
> -----Original Message-----
> From: owner-microblaze-uclinux@xxxxxxxxxxxxxx
> [mailto:owner-microblaze-uclinux@xxxxxxxxxxxxxx] On Behalf Of
> alucero@xxxxxxxxx
> Sent: Wednesday, November 23, 2005 12:32 PM
> To: microblaze-uclinux@xxxxxxxxxxxxxx
> Subject: Re: [microblaze-uclinux] BRAM
>
> What you comment is hard to get, since the stack is per process. However,
> I'm
> working now in allocating the irq code at LMB (Block Ram). Theorically,
> the
> performace should increase.
>
> By now, the results are a bit strange. Under normal load the performance
> is
> fine, but under stress with ethernet interrupts the LMB code is not
> enough
> to
> obtain always an aceptable latency (data used at irq should be at LMB
> too).
> The problem, I guess, is at the Microblaze cache and/or how the opb bus
> does
>
> the arbitration. In the worst cases the latency should be higher, but
> till
> certain limits, which is not the case now.
>
> I'll do a more specific explanation of the work in short, when I do all
> the
> tests facing with the LMB.
>
> Coming back to your question, the work could be done but you should work
> creating a new memory zone to be managed by the uClinux kernel, and using
> a
> specfic flag to request space from this new zone. The kernel code should
> be
> modified to use this new memory zone for specific tasks. I'm working in a
> dirty way by now, modifying the create_module function to insert the data
> and
> code module at LMB. This is just to test my implementation latencies. If
> you
>
> are intested we could discuss how to resolve the implementation for this
> new
>
> memoty zone.
>
> Regards.
>
> On Wednesday 23 November 2005 17:41, Bret Ketchum wrote:
> > Has anybody considered or have actually run the uClinux kernel with
> > the stack placed in BRAM? Does anybody have the recipe? Thanks in
> > advance....
> >
> > Bret
> >
> > ___________________________
> > microblaze-uclinux mailing list
> > microblaze-uclinux@xxxxxxxxxxxxxx
> > Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
> > Mailing List Archive :
> > http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/
>
> --
> Alejandro Lucero
> OS3, OS Serveis i Solucions
> www.os3sl.com
> Ingeniería Informática
> +34 665687168
> Av.Benjamin Franklin
> CEEI. Parque Tecnológico de Paterna
> Valencia(Spain)
>
___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/