https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37621
Gerald Pfeifer changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
CC|
--- Comment #3 from n dot pipenbrinck at cubic dot org 2008-09-24 22:41
---
ROL/ROR on the native integer size is not supported via intrinsics, but the
compiler will fold two shifts into a rotate.
If I want to manipulate only the lower 16 bit of an 32 bit integer (e.g. issue
a rolw) as
--- Comment #2 from dlenski at gmail dot com 2008-09-24 17:43 ---
Hi Andrew,
It seems to me that these modifiers are quite necessary for flexible x86
assembly. What is the point of the "q" and "Q" constraints if there's no way
to specifically refer to the 16-bit or 8-bit components of t
--- Comment #1 from pinskia at gcc dot gnu dot org 2008-09-24 00:35 ---
IIRC these are not documented because they can and will change at any time and
don't always produce correct code.
Also you can use most of the builtin functions to get the same as the
inline-asm ...
--
http://g
