File: bug_28748.s

package info (click to toggle)
llvm-toolchain-14 1%3A14.0.6-16
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 1,496,368 kB
  • sloc: cpp: 5,593,980; ansic: 986,873; asm: 585,869; python: 184,223; objc: 72,530; lisp: 31,119; f90: 27,793; javascript: 9,780; pascal: 9,762; sh: 9,482; perl: 7,468; ml: 5,432; awk: 3,523; makefile: 2,547; xml: 953; cs: 573; fortran: 567
file content (19 lines) | stat: -rw-r--r-- 667 bytes parent folder | download | duplicates (4) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
 
# RUN: llvm-mc -arch=hexagon -mcpu=hexagonv65 -filetype=obj %s | llvm-objdump --mcpu=hexagonv65 -d - | FileCheck --implicit-check-not='{' %s

# This case requires compounding only some of the instructions which are
# possible compounds.  Compounding all possible opcodes is ideal for code size
# but does not always result in a packet with a valid shuffle, whereas the
# non-compounded instructions may be a valid shuffle.

foo:
{ r0=c0
  p0=cmp.eq(r0,#0); if (p0.new) jump:nt foo
  jump foo
  r1=r0 }

# CHECK-LABEL:  <foo>:
# CHECK-NEXT:  { r0 = sa0
# CHECK-NEXT:    p0 = cmp.eq(r0,#0); if (p0.new) jump:nt 0x0
# CHECK-NEXT:    jump 0x0
# CHECK-NEXT:    r1 = r0 }