Re: [PATCH] kallsyms data size reduction / lookup speedup

[Paulo Marques <pmarques@grupopie.com>]

Matt Mackall wrote:
> On Wed, Aug 25, 2004 at 05:04:46AM +0100, pmarques@grupopie.com wrote:
> 
>>As always, comments, suggestions, flames will be greatly appreciated :)
> 
> 
> Please post patches inline so they're easier to comment on.
> Attachments are a nuisance.

Sorry about that. I've had problems in the past with my email client 
word wrapping patches, so to be sure the patch goes untouched I sent it 
this way.

Since I've changed email client since then, next time I'll try inlining 
again.

> Am I correct that this is completely replacing stem compression with
> your substring dictionary approach?

Yes, you are correct.

Right now I'm working on making the proc interface more eficient by 
removing all the seq_file stuff, that was needed because of the O(n) 
lookup time we had previously.

Not using seq_file with stem decompression would make a simple "cat 
/proc/kallsyms" to be O(n^2).

Bets regards,

-- 
Paulo Marques - www.grupopie.com