Merge branch 'for-next/sve' into for-next/core
authorWill Deacon <will@kernel.org>
Thu, 24 Jun 2021 13:07:04 +0000 (14:07 +0100)
committerWill Deacon <will@kernel.org>
Thu, 24 Jun 2021 13:07:04 +0000 (14:07 +0100)
commit3d1bf78c7b13a0ad88e77dce94df88400f47e0dd
tree8175a9c64a9dbfb47f73dd059a9f8233a1ce14ff
parenta4a49140ae84e74f5d831c2416fe4860c8ff9d34
parentad4711f962e08eff8d6e9b03f9670b1af6ea9395
Merge branch 'for-next/sve' into for-next/core

Optimise SVE switching for CPUs with 128-bit implementations.

* for-next/sve:
  arm64/sve: Skip flushing Z registers with 128 bit vectors
  arm64/sve: Use the sve_flush macros in sve_load_from_fpsimd_state()
  arm64/sve: Split _sve_flush macro into separate Z and predicate flushes