Re: GCC in-line assembly and the removal of -mcx16

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, 30 May 2017, Alexander Monakov wrote:
> I believe implementing ldaxp/stlxp in the backend and using them to expand
> 128-bit CAS (and other 128-bit atomics, except for atomic load) ...

Sorry, that was unclear.  To clarify, ldaxp/stlxp, if implemented, could be
used to implement 128-bit __sync built-ins, if AArch64 were to follow the
approach currently taken on x86.  However, they could not be used to implement
128-bit __atomic built-ins, like 128-bit cas is today not used on x86 for that
purpose either, because __atomic builtins provide a 'load' operation, and an
implementation using ldaxp-stlxp to implement an atomic load would fault on
read-only memory.

Alexander



[Index of Archives]     [Linux C Programming]     [Linux Kernel]     [eCos]     [Fedora Development]     [Fedora Announce]     [Autoconf]     [The DWARVES Debugging Tools]     [Yosemite Campsites]     [Yosemite News]     [Linux GCC]

  Powered by Linux