This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: GNURX toolchain generates incorrect opcode for "mov.b #0xff, [r0]" instruction.


Hi Vinay,

> Below patch fixes the issue reported. This is verified using following testcase.

Great - please could you add this testcase to the patch as a new entry in 
gas/testsuite/gas/rx ?  I suggest that you call the source file pr19665.s and the
test control file pr19665.d.  Then just add a line "run_dump_test "pr19665" to the
end of rx.exp.


> The patch is also regression tested. Please review and commit.

Really ?  Because when I regression tested it I saw a new gas testsuite failure:

  regexp_diff match failure
  regexp "^ [0-9a-f]+:	3c 00 00                mov\.b	#0, 0\[r0\]$"
  line   " 180:	f8 04 00                      	mov.b	#0, [r0]"
  regexp_diff match failure
  regexp "^ [0-9a-f]+:	3c 00 ff                mov\.b	#255, 0\[r0\]$"
  line   " 198:	f8 04 ff                      	mov.b	#255, [r0]"
  regexp_diff match failure
  regexp "^ [0-9a-f]+:	3d 00 7f                mov\.w	#127, 0\[r0\]$"
  line   " 1c8:	f8 05 7f                      	mov.w	#127, [r0]"
  regexp_diff match failure
  regexp "^ [0-9a-f]+:	3d 00 00                mov\.w	#0, 0\[r0\]$"
  line   " 1e0:	f8 05 00                      	mov.w	#0, [r0]"
  regexp_diff match failure
  regexp "^ [0-9a-f]+:	3e 00 7f                mov.l	#127, 0\[r0\]$"
  line   " 228:	f8 06 7f                      	mov.l	#127, [r0]"
  FAIL: gas/rx/mov

Cheers
  Nick


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]