<html>
    <head>
      <base href="https://llvm.org/bugs/" />
    </head>
    <body><span class="vcard"><a class="email" href="mailto:spatel+llvm@rotateright.com" title="Sanjay Patel <spatel+llvm@rotateright.com>"> <span class="fn">Sanjay Patel</span></a>
</span> changed
              <a class="bz_bug_link 
          bz_status_RESOLVED  bz_closed"
   title="RESOLVED FIXED - Integer divide by constant close to UINT*_MAX can be compare rather than divide"
   href="https://llvm.org/bugs/show_bug.cgi?id=28672">bug 28672</a>
        <br>
             <table border="1" cellspacing="0" cellpadding="8">
          <tr>
            <th>What</th>
            <th>Removed</th>
            <th>Added</th>
          </tr>

         <tr>
           <td style="text-align:right;">Status</td>
           <td>NEW
           </td>
           <td>RESOLVED
           </td>
         </tr>

         <tr>
           <td style="text-align:right;">Resolution</td>
           <td>---
           </td>
           <td>FIXED
           </td>
         </tr></table>
      <p>
        <div>
            <b><a class="bz_bug_link 
          bz_status_RESOLVED  bz_closed"
   title="RESOLVED FIXED - Integer divide by constant close to UINT*_MAX can be compare rather than divide"
   href="https://llvm.org/bugs/show_bug.cgi?id=28672#c3">Comment # 3</a>
              on <a class="bz_bug_link 
          bz_status_RESOLVED  bz_closed"
   title="RESOLVED FIXED - Integer divide by constant close to UINT*_MAX can be compare rather than divide"
   href="https://llvm.org/bugs/show_bug.cgi?id=28672">bug 28672</a>
              from <span class="vcard"><a class="email" href="mailto:spatel+llvm@rotateright.com" title="Sanjay Patel <spatel+llvm@rotateright.com>"> <span class="fn">Sanjay Patel</span></a>
</span></b>
        <pre>I'm marking this as fixed with:
<a href="https://reviews.llvm.org/rL282209">https://reviews.llvm.org/rL282209</a>

...because that gets rid of the urem in InstCombine.

There are a couple of potential follow-up bugs:

1. As noted, we can extend this transform even for non-constant divisors if we
can prove that the MSB (the sign bit) is set.

2. The backend doesn't seem to recognize negated constants:

ARM:
@ BB#0:                                 @ %entry
    ldr    r2, .LCPI0_0
    ldr    r3, .LCPI0_1
    add    r1, r0, #1
    cmp    r1, r3
    addhs    r1, r0, r2
    mov    r0, r1
    bx    lr
    .p2align    2
@ BB#1:
.LCPI0_0:
    .long    294397890               @ 0x118c27c2
.LCPI0_1:
    .long    4000569407              @ 0xee73d83f

x86:
    leal    1(%rdi), %ecx
    leal    294397890(%rdi), %eax
    cmpl    $-294397889, %ecx       ## imm = 0xEE73D83F
    cmovbl    %ecx, %eax
    retq</pre>
        </div>
      </p>
      <hr>
      <span>You are receiving this mail because:</span>
      
      <ul>
          <li>You are on the CC list for the bug.</li>
      </ul>
    </body>
</html>