Menu
▾
▴
Re: [Nasm-devel] [RFC 0/2] Early draft for instruction flags generator
|
From: Cyrill G. <gor...@gm...> - 2013-10-06 21:39:38
|
On Sun, Oct 06, 2013 at 02:34:10PM -0700, H. Peter Anvin wrote:
> I would like to suggest that we pick out the flags that are actually used by the
> assembler for other than CPU feature determination, like the Sx flags,
> and move them to a separate word and have the rest indirect.
> On the other hand, perhaps that is a pretty useless distinction?
I think, yes, it makes sence. Note that in first place I intentionally
moved Sx, ARG and etc into first dword
my %insns_flag_bit = (
#
# dword bound, index 0 - specific flags
#
"SM" => [ 0, "size match"],
"SM2" => [ 1, "size match first two operands"],
"SB" => [ 2, "unsized operands can't be non-byte"],
"SW" => [ 3, "unsized operands can't be non-word"],
"SD" => [ 4, "unsized operands can't be non-dword"],
"SQ" => [ 5, "unsized operands can't be non-qword"],
"SO" => [ 6, "unsized operands can't be non-oword"],
"SY" => [ 7, "unsized operands can't be non-yword"],
"SZ" => [ 8, "unsized operands can't be non-zword"],
"SIZE" => [ 9, "unsized operands must match the bitsize"],
"SX" => [ 10, "unsized operands not allowed"],
"AR0" => [ 11, "SB, SW, SD applies to argument 0"],
"AR1" => [ 12, "SB, SW, SD applies to argument 1"],
"AR2" => [ 13, "SB, SW, SD applies to argument 2"],
"AR3" => [ 14, "SB, SW, SD applies to argument 3"],
"AR4" => [ 15, "SB, SW, SD applies to argument 4"],
"OPT" => [ 16, "optimizing assembly only"],
>
> Either way, good work on this!
Lets do a deal, I'll try to finish all this on the week, and then
once everything work as expected (ie at least all tests get passed),
we re-think about moving these "specific" flags into separate entry,
ok?
|
