[PATCH/ARM] To fix vcvt bug.

[Sameera Deshpande]

Hi!

This patch adds check for range of the immediate operand in instruction
`vcvt' between floating point and fixed point.
If the type size of fixed point number is 16, the immediate value should be
in range [ 0, 16].
For 32 bit size, the immediate value should be in range [1, 32].

Tested using make check without regression.

Changelog:

2011-06-21  Sameera Deshpande  <sameera.deshpande@arm.com>


gas/


       * config/tc-arm.c (vfp_conv): Add check on range of immediate operand

       in vcvt instruction between floating-point and fixed-point.

       (operand_parse_code): Add "OP_oI32z".

       (parse_operands): OP_oI32z case added.


gas/testsuite/

       * gas/arm/vcvt-bad.d: New test.

       * gas/arm/vcvt-bad.s: New test.

       * gas/arm/vcvt-bad.l: New test.

       * gas/arm/vcvt.d: New test.

       * gas/arm/vcvt.s: New test. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: arm-vcvt-range.patch
Type: application/octet-stream
Size: 9854 bytes
Desc: not available
URL: <https://sourceware.org/pipermail/binutils/attachments/20110621/a54c17ee/attachment.obj>