Re: the SoC GPU driver interview
On Fri, Apr 26, 2013 at 12:47 PM, Riku Voipio <riku.voipio@iki.fi> wrote:
> Anyways, rather than push for more reverse engineering projects to get
> started, I think we should concentrate in finishing the current projects.
that assumes that there is a limited number of individuals in the
world with reverse-engineering skills: there isn't. i didn't have
reverse-engineering skills, but that didn't stop me from trying.
three years later i had full NT Domains interoperability built into
samba, including management of /etc/init.d scripts via the standard nt
gui management tools. i didn't have ARM hardware reverse-engineering
skills, but that didn't stop me from trying. nine wince smartphones
later i'd got 98% of drivers done on the 9th phone (the ipaq hw6915)
in about six weeks flat.
additionally, the guy who started the etnaviv project simply... loved
the way that libv was working on it, and was inspired to just... get
on with it. due to the simplicity of the vivante hardware he's
*overtaken* what libv has achieved, in a very short amount of time,
and has a *functional* but completely non-optimised gallium3d llvm
driver that's *faster* than vivante's own proprietary code.
so.... yeah. talking it down just makes people give up. that's not
good: we need people who *don't* know that it's "difficult". i didn't
know that MSRPC reverse-engineering was "difficult", which is why i
started doing it... and once started and having worked with paul to
get the "welcome to the samba domain" announcement, i sure as s**t
wasn't going to stop.
does that sound like a reasonable thing to do, riku - to assume that
there are people with more skills, enthusiasm and time than we have?
l.
Reply to: