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] |
On 08/04/13 16:46, Jan Beulich wrote:
On 08.04.13 at 17:41, Richard Earnshaw <rearnsha@arm.com> wrote:On 08/04/13 15:29, Jan Beulich wrote:Instructions shared between VFP and SIMD so far got rejected after using .arch_extension simd .arch_extension nofp because their table entries made them dependent on just VFP. Use a distinct feature mask for them instead. gas/ 2013-04-08 Jan Beulich <jbeulich@suse.com> * gas/config/tc-arm.c (fpu_vfp_v1xd_or_neon_ext): New. (insns) Use fpu_vfp_v1xd_or_neon_ext.This needs some tests as well, please. In particular, the tests need to ensure that the floating-point SIMD versions are correctly disabled for the case you cite above.Not sure I follow: The floating point variant of the SIMD encodings aren't tied to the presence of VFP, and I don't think I've seen a separate integer-only-SIMD feature flag in the gas sources. Jan
The class of common mnemonics you've changed includes instructions such as vmul. When you have Neon but no FP, both the scalar FP and vector FP variants of the instructions should be disabled, but the integer versions remain.
This needs some testing to ensure that the right things are disabled/enabled accordingly.
R.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |