JFFS3 & performance

[David Woodhouse]

On Wed, 2005-01-12 at 09:22 -0800, Jared Hulbert wrote:
> Think of the flow of data to start executing an application from
> JFFS2.  We copy the data 3 times in RAM? decompress it, CRC it. 

We ought to be decompressing directly from flash to RAM. It takes 0.5x
flash, and 1x RAM. And the RAM usage is dynamic -- you can throw the
page away if memory is tight, and read it back again later.

>  Compare that to mmap() in XIP cramfs.  It just points to the flash
> address.

And takes twice the amount of flash. Yes, it uses less RAM, but I don't
see how it can reduce the ROM cost. 

I haven't seen a decent analysis of the real-life cost of startup -- the
CELF numbers at OLS included decompression of the kernel in their
startup timings.

-- 
dwmw2