Fix gas testsuite unexpected failure for arm-aout

Nick Clifton nickc@redhat.com
Thu Mar 3 14:40:00 GMT 2005


Hi Guys,

  I am applying the patch below to fix an unexpected failure in the
  gas testsuite for the arm-aout target.  The test file arm/arch6zk.s
  did not contain enough instructions for the generated .text section
  to end on a 32-byte boundary and this is required by the aout
  format.  The result was that the assembler was inserting extra nop
  instructions at the end of the .text section which were then not
  matched by the arm/arch6zk.d file.

  I added the necessary nops and verified that both the arm-aout and
  arm-elf ports now have zero gas testsuite failures.

Cheers
  Nick

gas/testsuite/ChangeLog
2005-03-03  Nick Clifton  <nickc@redhat.com>

	* gas/arm/arch6zk.s: Add three nop instructions in order to pad
	the .text section out to a 32-byte boundary as will automatically
	be done by the arm-aout target.
	* gas/arm/arch6zk.d: Add expected nop disassemblies.

Index: gas/testsuite/gas/arm/arch6zk.s
===================================================================
RCS file: /cvs/src/src/gas/testsuite/gas/arm/arch6zk.s,v
retrieving revision 1.1
diff -c -3 -p -r1.1 arch6zk.s
*** gas/testsuite/gas/arm/arch6zk.s	30 Sep 2004 16:21:46 -0000	1.1
--- gas/testsuite/gas/arm/arch6zk.s	3 Mar 2005 14:37:46 -0000
*************** label:
*** 25,27 ****
--- 25,33 ----
  	# ARMV6Z instructions
  	smi 0xec31
  	smine 0x13ce
+ 
+ 	# Add three nop instructions to ensure that the 
+ 	# output is 32-byte aligned as required for arm-aout.
+ 	nop
+ 	nop
+ 	nop
  Index: gas/testsuite/gas/arm/arch6zk.d
===================================================================
RCS file: /cvs/src/src/gas/testsuite/gas/arm/arch6zk.d,v
retrieving revision 1.1
diff -c -3 -p -r1.1 arch6zk.d
*** gas/testsuite/gas/arm/arch6zk.d	30 Sep 2004 16:21:46 -0000	1.1
--- gas/testsuite/gas/arm/arch6zk.d	3 Mar 2005 14:37:46 -0000
*************** Disassembly of section .text:
*** 26,28 ****
--- 26,32 ----
  0+048 <[^>]*> e320f001 ?	yield
  0+04c <[^>]*> e16ec371 ?	smi	60465
  0+050 <[^>]*> 11613c7e ?	smine	5070
+ 0+054 <[^>]*> e1a00000 ?	nop[ 	]+\(mov r0,r0\)
+ 0+058 <[^>]*> e1a00000 ?	nop[ 	]+\(mov r0,r0\)
+ 0+05c <[^>]*> e1a00000 ?	nop[ 	]+\(mov r0,r0\)
+ 



More information about the Binutils mailing list