tree a6df902bc12ced4445e6447263459d034c18b1d8
parent 9060ca01913776cbed78b03e72d7700adc2201b7
author John Brawn <john.brawn@arm.com> 1746186918 +0100
committer GitHub <noreply@github.com> 1746186918 +0100
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsFcBAABCAAQBQJoFLKmCRC1aQ7uu5UhlAAAedEQAJCJpLPqiAqtgQGxNUsbgQyF
 cVptPHmVRoehu398PYXpfZbpyeTgxDoU8ClHgkIRQtSIzGUtVChVo1zykeQmpd0F
 rqsXKr8KfDFdLaA7GBZQOrw16QjuVTGGQnxalvPJvyVRtmqBmqT7ZM5opYCzSzE/
 gjp4VoWqbL9rY6vEBQIYMyAlDuuIu81U+B/EAwG5/HiAYSzpLUyZgfZlOwJJmG0Y
 ox6Ecw1kpBlQWd7t2gzB+vi+Bd7zUCGM3IkloNlSccOGA0O5T+2nY8dc6J2e/wXg
 Vl0xE1eT7liw/g+acQg6ufjT01CsB1wTyqKpGga1PVA+UosZi6y1td3Wlp8rgrEp
 oNujj4Ph2d1j2aQZYwekmCPg+o/jhaUc9JQV6kiu6NorcLpWH3M1YnJ+Fs1yW/JO
 AcOPyCk/1jaC0vHdyB8grNg4MLytilhChxVEEdMxRhRECBjW2OdyGv66wjpBMsW+
 IFS7lgRBPT/URLqXeS6k9PeNuIzuPzDoxT272hceXa/Q9z+37tnRDHI8izjCpFmN
 BUJRR7Dft7GnJ9Ts/qykihcLnZwWvflhiCyUOhYGD6aJBQ6m9wNWteri6pi0T5FA
 uw5Vyp1iCGz2CqTOWrVcV+uw6+bbv9WJLUnfLrxa6Jgdgxix1uM/NCXqRjiotHI2
 VkZIc3vXglcc0ZkjYX+s
 =R/7T
 -----END PGP SIGNATURE-----
 

[AArch64] Use pattern to select bf16 fpextend (#137212)

Currently bf16 fpextend is lowered to a vector shift. Instead leave it
as fpextend and have an instruction selection pattern which selects to a
shift later. Doing this means that DAGCombiner patterns for fpextend
will be applied, leading to better codegen. It also means that in some
situations we use a mov instruction where we previously have a dup
instruction, but I don't think this makes any difference.