LSB1.1: /proc/cpuinfo

[Alan Cox]

> > Which is down to MMX, MMX2, SSE, SSE2. Take a look at real existing apps
> > like mplayer, mp1e and xine.
> 
> Are you saying that these are all that is needed, and that this problem has
> already been solved?

I am saying that the existing applications don't show any evidence that
we need to document these and use cpuid quite successfully themselves

> I've seen comments that indicate that depending on SIGILL and SIGFPE to
> determine that something isn't present may not be fool proof. Also, it takes

For x86 SIGILL is safe barring some obscure errata which are not going to
bite anyone in normal use. FPU is always present but may be emulated. You
can again check this with cpuid.

> In setup.c there are several place where feature bits are turned on or off
> based model and errata info. Why should this kind of handlig be duplicated?

It wouldn't be. Applications cannot control those feature flags. 

If the code was complex I would be worry, but it is less code to use
cpuid for things like MMX checking than to go regexping through the /proc
file. It is also very well documented (once you realise Intel try and make
you run the competitors chips as a 486 and the others likewise about
chips that they didnt make)

Alan