On Mon, Nov 27, 2017 at 9:41 PM, Arnd Bergmann email@example.com wrote:
One argument for using c) plus the emulation in glibc is that glibc has to do emulation anyway, to allow running user space with 64-bit time_t on older kernels that don't have the new getrusage system call.
To clarify my point here, if we stay with approach c), I think it should work directly with the Albert's proposed patch "Y2038: add function __getrusage_t64" , and we can remove the " // TODO: use 64-bit-time syscall if available" there. If we pick approach b), we still need the same glibc patch, but would also implement the interface to the new system call. As Eric said, this would be slightly cleaner, but not really help us since 32-bit fields in rusage are sufficient on the kernel interface side.