[RFC2 RESEND PATCH] [AARCH64] ILP32: support stat syscall family

[Joseph Myers]

On Mon, 1 Aug 2016, Yury Norov wrote:

> 32-bit ports should have 64-bit time types and structures to be ready
> to Y2038, and for now we have no such structures.

If you are interested in Y2038 support, you need to engage in the threads 
discussing successive revisions of 
<https://sourceware.org/glibc/wiki/Y2038ProofnessDesign>.  I imagine that 
Albert ARIBAUD will be working on patches once the design is established.

> OK, I'll prepare the patch w/o unions. Anyway, I think we'd have
> 64-bit timespec for 32-bit ports. Are you OK if I will introduce
> struct timespec64 in separated installed header, and use it under
> SUPPORT_64_BIT_TIME_TYPES option?

SUPPORT_64_BIT_TIME_TYPES is in the user namespace; it has no business in 
installed headers.  The proposed design does not involve any APIs for 
explicit 64-bit time use (APIs like lseek64 with off64_t in the 
large-files case), only use of _TIME_BITS=64 to redirect existing APIs to 
new ABIs (those ABIs being in the implementation namespace, of course, and 
new ABIs only being present in those cases where the default interface 
uses 32-bit time_t).  Thus there should be no such type as struct 
timespec64 in installed headers, only different definitions of struct 
timespec depending on _TIME_BITS.

-- 
Joseph S. Myers
joseph@codesourcery.com