[llvm-bugs] [Bug 44918] New: [AArch64] SLH is completely broken with GlobalISel

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=44918

            Bug ID: 44918
           Summary: [AArch64] SLH is completely broken with GlobalISel
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: AArch64
          Assignee: unassignedbugs at nondot.org
          Reporter: efriedma at quicinc.com
                CC: arnaud.degrandmaison at arm.com, kristof.beyls at arm.com,
                    llvm-bugs at lists.llvm.org, peter.smith at linaro.org,
                    Ties.Stuij at arm.com

Testcase:

target triple = "aarch64-unknown-linux-gnu"
define void @foo(i32 %inp, void ()* nocapture %bar) speculative_load_hardening
{
entry:
  %cmp = icmp slt i32 %inp, 10
  br i1 %cmp, label %if.then, label %if.end

if.then:
  tail call void %bar() #1
  br label %if.end

if.end:
  ret void
}

Build with llc -O0.  Produces:

        cmp     w0, #10                 // =10
        cset    w8, ge
        tst     w8, #0x3
        str     x1, [sp, #8]            // 8-byte Folded Spill
        b.eq    .LBB0_2

(LBB0_2 is if.end).

Note the condition is backwards: we branch to if.end if the input is less than
10.  Apparently nobody ever did any testing of speculative load hardening at
-O0?

-- 
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/20200214/9e441608/attachment.html>