diff options
| author | Joe Ramsay <Joe.Ramsay@arm.com> | 2023-12-19 16:44:01 +0000 |
|---|---|---|
| committer | Szabolcs Nagy <szabolcs.nagy@arm.com> | 2023-12-20 08:41:25 +0000 |
| commit | cc0d77ba944cd4ce46c5f0e6d426af3057962ca5 (patch) | |
| tree | 840c09b10bcb0ad4f733e8cb4bce2acbd92e5945 /include | |
| parent | 3150cc0c9019bf9da841419f86dda8e7f26d676d (diff) | |
| download | glibc-cc0d77ba944cd4ce46c5f0e6d426af3057962ca5.tar.gz glibc-cc0d77ba944cd4ce46c5f0e6d426af3057962ca5.tar.xz glibc-cc0d77ba944cd4ce46c5f0e6d426af3057962ca5.zip | |
aarch64: Add half-width versions of AdvSIMD f32 libmvec routines
Compilers may emit calls to 'half-width' routines (two-lane single-precision variants). These have been added in the form of wrappers around the full-width versions, where the low half of the vector is simply duplicated. This will perform poorly when one lane triggers the special-case handler, as there will be a redundant call to the scalar version, however this is expected to be rare at Ofast. Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
Diffstat (limited to 'include')
| -rw-r--r-- | include/libc-symbols.h | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/include/libc-symbols.h b/include/libc-symbols.h index 5794614488..a226119295 100644 --- a/include/libc-symbols.h +++ b/include/libc-symbols.h @@ -600,8 +600,10 @@ for linking") #endif #if IS_IN (libmvec) +# define libmvec_hidden_proto(name, attrs...) hidden_proto (name, ##attrs) # define libmvec_hidden_def(name) hidden_def (name) #else +# define libmvec_hidden_proto(name, attrs...) # define libmvec_hidden_def(name) #endif |