public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* [PATCH 00/10] x86: disassembler table entry folding plus some cleanup
@ 2023-07-04 15:17 Jan Beulich
  2023-07-04 15:18 ` [PATCH 01/10] x86: fold certain legacy/VEX table entries Jan Beulich
                   ` (9 more replies)
  0 siblings, 10 replies; 12+ messages in thread
From: Jan Beulich @ 2023-07-04 15:17 UTC (permalink / raw)
  To: Binutils; +Cc: H.J. Lu

This continues the transformation of the tables we use to re-use
entries where possible and to shrink entry count / table size by
picking an appropriate decode order when multiple decode steps
involving table lookup are used for a given encoding.

The reduction of mod_table[] in particular suggests that going
forward we should strive to add entries there only if both of
the branches are actually populated. Single-branch cases
(register-only or memory-only forms) would instead be expressed
by (re)using suitable operand handlers.

Finally there's a little bit of tidying of things noticed while
doing the main part of the work.

01: fold certain legacy/VEX table entries
02: fold legacy/VEX {,V}MOV{H,L}* entries
03: {,V}MOVNT* don't need to go through mod_table[]
04: misc further memory-only insns don't need to go through mod_table[]
05: SIMD shift-by-immediate don't need to go through mod_table[]
06: slightly rework handling of some register-only insns
07: various operations on mask registers can avoid going through mod_table[]
08: misc further register-only insns don't need to go through mod_table[]
09: convert 0FXOP to just XOP in enumerator names
10: simplify table-referencing macros

Jan

^ permalink raw reply	[flat|nested] 12+ messages in thread

end of thread, other threads:[~2023-07-07 12:28 UTC | newest]

Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-07-04 15:17 [PATCH 00/10] x86: disassembler table entry folding plus some cleanup Jan Beulich
2023-07-04 15:18 ` [PATCH 01/10] x86: fold certain legacy/VEX table entries Jan Beulich
2023-07-04 15:19 ` [PATCH 02/10] x86: fold legacy/VEX {,V}MOV{H,L}* entries Jan Beulich
2023-07-04 15:20 ` [PATCH 03/10] x86: {,V}MOVNT* don't need to go through mod_table[] Jan Beulich
2023-07-04 15:21 ` [PATCH 04/10] x86: misc further memory-only insns " Jan Beulich
2023-07-04 15:21 ` [PATCH 05/10] x86: SIMD shift-by-immediate " Jan Beulich
2023-07-04 15:22 ` [PATCH 06/10] x86: slightly rework handling of some register-only insns Jan Beulich
2023-07-04 15:22 ` [PATCH 07/10] x86: various operations on mask registers can avoid going through mod_table[] Jan Beulich
2023-07-04 15:22 ` [PATCH 08/10] x86: misc further register-only insns don't need to go " Jan Beulich
2023-07-07 12:28   ` Jan Beulich
2023-07-04 15:23 ` [PATCH 09/10] x86: convert 0FXOP to just XOP in enumerator names Jan Beulich
2023-07-04 15:24 ` [PATCH 10/10] x86: simplify table-referencing macros Jan Beulich

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).