Re: [PATCH][ARM] Add movv4hf/v8hf expanders & later insns; disable VnHF immediates.

2016-01-18 Thread Christophe Lyon
On 18 January 2016 at 14:12, Kyrill Tkachov wrote: > Hi Alan, > > > On 18/01/16 12:14, Alan Lawrence wrote: >> >> This fixes ICEs on armeb for float16x[48]_t vectors, e.g. in >> check_effective_target_arm_neon_fp_16_ok. >> >> At present, without the expander, moving

Re: [PATCH][ARM] Add movv4hf/v8hf expanders & later insns; disable VnHF immediates.

2016-01-18 Thread Kyrill Tkachov
Hi Alan, On 18/01/16 12:14, Alan Lawrence wrote: This fixes ICEs on armeb for float16x[48]_t vectors, e.g. in check_effective_target_arm_neon_fp_16_ok. At present, without the expander, moving v4hf/v8hf values around is done via subregs. On armeb, this ICEs because REG_CANNOT_CHANGE_MODE_P.

[PATCH][ARM] Add movv4hf/v8hf expanders & later insns; disable VnHF immediates.

2016-01-18 Thread Alan Lawrence
This fixes ICEs on armeb for float16x[48]_t vectors, e.g. in check_effective_target_arm_neon_fp_16_ok. At present, without the expander, moving v4hf/v8hf values around is done via subregs. On armeb, this ICEs because REG_CANNOT_CHANGE_MODE_P. (On arm-*, moving via two subregs is less efficient