[llvm-bugs] [Bug 44467] New: [ARM][load / store optimization pass] Missed optimization: coalesce pre-increment

via llvm-bugs llvm-bugs at lists.llvm.org
Mon Jan 6 02:05:14 PST 2020


https://bugs.llvm.org/show_bug.cgi?id=44467

            Bug ID: 44467
           Summary: [ARM][load / store optimization pass] Missed
                    optimization: coalesce pre-increment
           Product: new-bugs
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Keywords: missing-feature
          Severity: enhancement
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: diogo.sampaio at arm.com
                CC: htmldeveloper at gmail.com, llvm-bugs at lists.llvm.org

Created attachment 22991
  --> https://bugs.llvm.org/attachment.cgi?id=22991&action=edit
Minimal MIR case

Missed optimization:
When running

llc --print-before-all --start-before=arm-ldst-opt missed.mir -o -


we obtain:
        push.w  {r4, r5, r6, r7, r8, r10, r11, lr}
        sub     sp, #8
        strd    r0, r1, [sp]            @ 8-byte Folded Spill
        @ InlineAsm Start
        @ InlineAsm End
        ldrd    r0, r1, [sp], #8        @ 8-byte Folded Reload
        pop.w   {r4, r5, r6, r7, r8, r10, r11, pc}

I would expect the strd to be pre-incremented, as:
        push.w  {r4, r5, r6, r7, r8, r10, r11, lr}
        strd    r0, r1, [sp, #-8]!
        @ InlineAsm Start
        @ InlineAsm End
        ldrd    r0, r1, [sp], #8
        pop.w   {r4, r5, r6, r7, r8, r10, r11, pc}

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20200106/f87875d5/attachment-0001.html>


More information about the llvm-bugs mailing list