Re: [aarch64] question about user.h, ucontext.h, procfs.h

[Mike Frysinger <vapier at gentoo dot org>]

On 23 Apr 2015 23:03, Szabolcs Nagy wrote:
> * Mike Frysinger <vapier@gentoo.org> [2015-04-23 16:13:10 -0400]:
> > On 23 Apr 2015 14:46, Szabolcs Nagy wrote:
> > > > 
> > > > aarch64 user_fpsimd_struct has
> > > > 
> > > >   __uint128_t vregs[32];
> > > > 
> > > > member to represent the 128bit floating point registers
> > > > 
> > > > is it ok to expose __uint128_t in public headers?
> > > 
> > > (i'd prefer 'long double' so compilers without __uint128
> > > can use signal.h)
> > 
> > vregs is supposed to hold the content of the registers.  it doesn't need to be 
> > able to represent the data directly (i.e. use doubles).  using uint128_t is the 
> > right thing to do here.
> 
> uint128_t is not in the reserved namespace of signal.h
> if _POSIX_C_SOURCE is defined before its inclusion
> 
> __uint128_t is namespace clean, just non-standard compiler
> extension, but probably i shouldn't worry about compiler
> support on a new arch

i don't mean exactly "uint128_t", i mean wrt double -- you don't want that
-mike

Attachment: signature.asc
Description: Digital signature