x86: respect ".arch nonop" when selecting which NOPs to emit
authorJan Beulich <jbeulich@suse.com>
Mon, 23 Oct 2023 08:32:32 +0000 (10:32 +0200)
committerJan Beulich <jbeulich@suse.com>
Mon, 23 Oct 2023 08:32:32 +0000 (10:32 +0200)
commitcd75cd859e58ff7161871849c59f01e9df3cdbc3
treedbf7d2d071c1a19c38553e42449ed44a96d6bc22
parent5e0729b655a2594c7b6750fb53f2f99750744ca2
x86: respect ".arch nonop" when selecting which NOPs to emit

Making GENERIC64 a special case was never correct; prior to the
generalization of ".arch .no*" to cover all ISA extensions other
processor families supporting long NOPs should have been covered as
well. When introducing ".arch .nonops" (among others) it wasn't
apparent that a hidden implication of .cpunop not being possible to
separately turn off existed here. Seeing that the two large case label
blocks in the 2nd switch() already had identical behavior, simply
collapse all of the (useful) case labels into a single "default" one.
gas/config/tc-i386.c
gas/config/tc-i386.h
gas/testsuite/gas/i386/x86-64-nops-1a-core2.d [new file with mode: 0644]
gas/testsuite/gas/i386/x86-64-nops-1a-g64.d [new file with mode: 0644]
gas/testsuite/gas/i386/x86-64.exp