Re: [PATCH] Add support for mkostemp, mkstemps and mkostemps

[Rich Felker <dalias@aerifal.cx>]

On Wed, Jan 30, 2013 at 08:12:58PM +0100, Szabolcs Nagy wrote:
> * Rich Felker <dalias@aerifal.cx> [2013-01-30 11:51:27 -0500]:
> > current time. Better use of the stack address in generating the
> > filenames could prevent knowing the set of output filenames for a
> > range of times without knowing the stack address in the program being
> > attacked. In fact, I'm a little bit worried that the current approach
> > discloses too much information about the stack address to an attacker.
> > If nothing else, I think some shuffling should be done so that the
> > (typically more valuable) high bits of the stack address are matched
> > with the low (least predictable) bits of the clock.
> 
> void __randname(char *p)
> {
> 	struct timespec ts;
> 	unsigned long r;
> 	int i;
> 
> 	clock_gettime(CLOCK_REALTIME, &ts);
> 	r = ts.tv_nsec*65537 ^ (uintptr_t)&ts / 16 + (uintptr_t)p;
> 	for (i=0; i<6; i++, r>>=5)
> 		p[i] = 'A'+(r&15)+(r&16)*2;
> }
> 
> this uses 30bits of r and mixes the random low bits of nsec
> into the high bits

Keep in mind it might be bits 8-15 that are most valuable with ASLR
(assuming the randomization only adjusts by small amounts and not so
much to waste lots of address space). I think this needs a little bit
more consideration.

> > > more significant improvement can be done by larger
> > > set of names and better entropy source
> > 
> > Other implementations probably use 36 bits or slightly less (base64
> > perhaps modified base64).
> > 
> > I could see it being feasible to increase this slightly and maybe even
> 
> <= 36bits is probably ok

You mean >=36? Or..?

Rich