gallivm/rgtc: fix the truncation to 8-bit
authorDave Airlie <airlied@redhat.com>
Fri, 3 Apr 2020 03:54:17 +0000 (13:54 +1000)
committerDave Airlie <airlied@redhat.com>
Sat, 4 Apr 2020 02:33:49 +0000 (12:33 +1000)
commit03204dadbc1829128f3e0a5e74f4f85851f6e708
treeccc8630e30a7a42312704815e6eef9c1cd18c04e
parent0b06adb75054842294e4dbbe2e5af294470862fb
gallivm/rgtc: fix the truncation to 8-bit

The 8 bit type wasn't 8-bit so when doing signed work we lost
the sign bit. This fixes it to use a proper vector type,
even if we just end up in here with the 1-wide path for now.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4425>
src/gallium/auxiliary/gallivm/lp_bld_format_s3tc.c