logo       

Re: Re: IXP_NPE Driver Inclusion proposal: msg#00053

misc.nslu2.devel

Subject: Re: Re: IXP_NPE Driver Inclusion proposal

We are very unlikely to ever need NPE-A, as this is specialised for ATM data transfer, which no existing device similar to the NSLU2 supports.

Don't worry :)

Mike

On 14 Sep 2006, at 14:34, Rod Whitby wrote:

So does that mean that the absolute maximum possible microcode size for
any IXP4[26]X device is 3 x 24k = 132Kb uncompressed?

We happen to have just under 124Kb free in the last block, so it seems
we can support all the microcode we desire, and only really need to
compress it in an absolute worst case (definitely not for the IXP42X for
instance).

I would prefer to store it uncompressed if we have the unused space, so
that mc_grab can be easily used to get it.

-- Rod

Christian Hohnstaedt wrote:
On Thu, Sep 14, 2006 at 10:22:41PM +0930, Rod Whitby wrote:
Anyone got a preference for where we put the microcode in flash?

I propose that it go in the last block of flash, after the FIS
directory, but before the Sercomm trailer.

Slugimage denotes this as the "Payload" area (see
http://www.nslu2-linux.org/wiki/Info/LastBlockMemoryMap for details).

I propose the start of the currently defined Payload area be used for
NPE microcode.

What is the maximum size that we need? The unused payload area is

A microcode image is ~12-16k depending on the features.
It can be compressed very well about 1/3 of its original size.
And you can use "zcat" to download it.

The maximum data and code size for NPE-B and NPE-C on IXP425 is 16K.
For NPE-A on IXP425 it is 24k.
On ixp465 CPUs it is 24K for every NPE.

Hope that helps

Please forward this message to your
nslu2-developers-hHKSG33TihhbjbujkaE4pw@xxxxxxxxxxxxxxxx

Christian Hohnstaedt




Yahoo! Groups Links










Attachment: smime.p7s
Description: S/MIME cryptographic signature

<Prev in Thread] Current Thread [Next in Thread>
Google Custom Search

News | FAQ | advertise