i965/vec4: use byte_offset() instead of offset()
authorIago Toral Quiroga <itoral@igalia.com>
Mon, 3 Oct 2016 11:26:54 +0000 (13:26 +0200)
committerIago Toral Quiroga <itoral@igalia.com>
Thu, 27 Oct 2016 08:59:31 +0000 (10:59 +0200)
commitba63db1f2ecf78ab894fe411b4f27d8523cf952d
treed00af93100647a95eea19985903f22d91b07ecb0
parent5a4ce9f9a79934e4cfc2034c7d181a671e2563bf
i965/vec4: use byte_offset() instead of offset()

In a later patch we want to change the semantics of offset() to be in terms
of SIMD width and scalar channels so it is consistent with the definition
of the same helper in the scalar backend. However, some uses of offset()
in the vec4 backend do not operate naturally in terms of these
semantics. In these cases it is more natural to use the byte_offset() helper
instead.

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
src/mesa/drivers/dri/i965/brw_vec4_cse.cpp
src/mesa/drivers/dri/i965/brw_vec4_dead_code_eliminate.cpp
src/mesa/drivers/dri/i965/brw_vec4_live_variables.cpp
src/mesa/drivers/dri/i965/brw_vec4_tcs.cpp
src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp