<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">Reverted in r323841.<br>
<br>
-Eli<br>
<br>
On 1/30/2018 4:14 PM, Galina Kistanova wrote:<br>
</div>
<blockquote type="cite"
cite="mid:CAJ8eiNz8gGQctOfu=qkqJZLxQraBbyy1NiuwQ+B9WN=S32EXxQ@mail.gmail.com">
<div dir="ltr">Please do to revert it till the fix gets approved.<br>
It is not good to keep the bot red for too long. This hides new
problem which later hard to track down.<br>
<br>
Thanks<br>
<br>
Galina<br>
</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Tue, Jan 30, 2018 at 11:51 AM,
Friedman, Eli <span dir="ltr"><<a
href="mailto:efriedma@codeaurora.org" target="_blank"
moz-do-not-send="true">efriedma@codeaurora.org</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<div text="#000000" bgcolor="#FFFFFF">
<div class="m_-679428887874725126moz-cite-prefix"><a
class="m_-679428887874725126moz-txt-link-freetext"
href="https://reviews.llvm.org/D42655" target="_blank"
moz-do-not-send="true">https://reviews.llvm.org/<wbr>D42655</a>
is up for review.<br>
<br>
I was hoping the fix would be reviewed quickly, but I
can revert in the meantime if it's blocking you.<br>
<br>
-Eli
<div>
<div class="h5"><br>
<br>
On 1/30/2018 10:55 AM, Galina Kistanova wrote:<br>
</div>
</div>
</div>
<div>
<div class="h5">
<blockquote type="cite">
<div dir="ltr">Hello Eli,<br>
<br>
This commit broke the test LLVM ::
CodeGen/ARM/pr25838.ll on one of our builders:<br>
r323559<br>
<a
href="http://lab.llvm.org:8011/builders/llvm-clang-x86_64-expensive-checks-win/builds/7546"
target="_blank" moz-do-not-send="true">http://lab.llvm.org:8011/<wbr>builders/llvm-clang-x86_64-<wbr>expensive-checks-win/builds/<wbr>7546</a><br>
<br>
. . .<br>
Failing Tests (2):<br>
LLVM :: CodeGen/ARM/pr25838.ll<br>
LLVM :: DebugInfo/Generic/string-<wbr>offsets-multiple-cus.ll<br>
<br>
Previous revision:<br>
<a
href="http://lab.llvm.org:8011/builders/llvm-clang-x86_64-expensive-checks-win/builds/7610"
target="_blank" moz-do-not-send="true">http://lab.llvm.org:8011/<wbr>builders/llvm-clang-x86_64-<wbr>expensive-checks-win/builds/<wbr>7610</a><br>
<br>
Please have a look?<br>
<br>
Thanks<br>
<br>
Galina<br>
<br>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Fri, Jan 26, 2018 at
12:23 PM, Eli Friedman via llvm-commits <span
dir="ltr"><<a
href="mailto:llvm-commits@lists.llvm.org"
target="_blank" moz-do-not-send="true">llvm-commits@lists.llvm.org</a>></span>
wrote:<br>
<blockquote class="gmail_quote"
style="margin:0 0 0 .8ex;border-left:1px
#ccc solid;padding-left:1ex">Author:
efriedma<br>
Date: Fri Jan 26 12:23:00 2018<br>
New Revision: 323559<br>
<br>
URL: <a
href="http://llvm.org/viewvc/llvm-project?rev=323559&view=rev"
rel="noreferrer" target="_blank"
moz-do-not-send="true">http://llvm.org/viewvc/llvm-pr<wbr>oject?rev=323559&view=rev</a><br>
Log:<br>
[LivePhysRegs] Preserve pristine regs in
blocks with no successors.<br>
<br>
One common source of blocks with no
successors is calls to noreturn<br>
functions; we want to preserve pristine
registers in case they throw an<br>
exception.<br>
<br>
The whole pristine register thing is messy
(we should really prefer to<br>
explicitly model registers), but this fills
a hole in the model for now.<br>
<br>
Fixes <a
href="https://bugs.llvm.org/show_bug.cgi?id=36073"
rel="noreferrer" target="_blank"
moz-do-not-send="true">https://bugs.llvm.org/show_bug<wbr>.cgi?id=36073</a>.<br>
<br>
Differential Revision: <a
href="https://reviews.llvm.org/D42509"
rel="noreferrer" target="_blank"
moz-do-not-send="true">https://reviews.llvm.org/D4250<wbr>9</a><br>
<br>
<br>
Added:<br>
llvm/trunk/test/CodeGen/Thumb/<wbr>stm-scavenging.ll<br>
Modified:<br>
llvm/trunk/lib/CodeGen/LivePhy<wbr>sRegs.cpp<br>
<br>
Modified: llvm/trunk/lib/CodeGen/LivePhy<wbr>sRegs.cpp<br>
URL: <a
href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/LivePhysRegs.cpp?rev=323559&r1=323558&r2=323559&view=diff"
rel="noreferrer" target="_blank"
moz-do-not-send="true">http://llvm.org/viewvc/llvm-pr<wbr>oject/llvm/trunk/lib/CodeGen/<wbr>LivePhysRegs.cpp?rev=323559&<wbr>r1=323558&r2=323559&view=diff</a><br>
==============================<wbr>==============================<wbr>==================<br>
--- llvm/trunk/lib/CodeGen/LivePhy<wbr>sRegs.cpp
(original)<br>
+++ llvm/trunk/lib/CodeGen/LivePhy<wbr>sRegs.cpp
Fri Jan 26 12:23:00 2018<br>
@@ -225,10 +225,10 @@ void
LivePhysRegs::addLiveOutsNoPri<wbr>stine<br>
<br>
void LivePhysRegs::addLiveOuts(cons<wbr>t
MachineBasicBlock &MBB) {<br>
const MachineFunction &MF =
*MBB.getParent();<br>
- if (!MBB.succ_empty()) {<br>
+ if (!MBB.isReturnBlock()) {<br>
addPristines(MF);<br>
addLiveOutsNoPristines(MBB);<br>
- } else if (MBB.isReturnBlock()) {<br>
+ } else {<br>
// For the return block: Add all callee
saved registers.<br>
const MachineFrameInfo &MFI =
MF.getFrameInfo();<br>
if (MFI.isCalleeSavedInfoValid())<br>
<br>
Added: llvm/trunk/test/CodeGen/Thumb/<wbr>stm-scavenging.ll<br>
URL: <a
href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/Thumb/stm-scavenging.ll?rev=323559&view=auto"
rel="noreferrer" target="_blank"
moz-do-not-send="true">http://llvm.org/viewvc/llvm-pr<wbr>oject/llvm/trunk/test/CodeGen/<wbr>Thumb/stm-scavenging.ll?rev=<wbr>323559&view=auto</a><br>
==============================<wbr>==============================<wbr>==================<br>
--- llvm/trunk/test/CodeGen/Thumb/<wbr>stm-scavenging.ll
(added)<br>
+++ llvm/trunk/test/CodeGen/Thumb/<wbr>stm-scavenging.ll
Fri Jan 26 12:23:00 2018<br>
@@ -0,0 +1,47 @@<br>
+; RUN: llc < %s | FileCheck %s<br>
+target triple = "thumbv6---gnueabi"<br>
+<br>
+; Use STM to save the three registers<br>
+; CHECK-LABEL: use_stm:<br>
+; CHECK: .save {r7, lr}<br>
+; CHECK: .setfp r7, sp<br>
+; CHECK: stm r3!, {r0, r1, r2}<br>
+; CHECK: bl throws_1<br>
+define void @use_stm(i32 %a, i32 %b, i32
%c, i32* %d) local_unnamed_addr noreturn
"no-frame-pointer-elim"="true" {<br>
+entry:<br>
+ %arrayidx = getelementptr inbounds i32,
i32* %d, i32 2<br>
+ store i32 %a, i32* %arrayidx, align 4<br>
+ %arrayidx1 = getelementptr inbounds i32,
i32* %d, i32 3<br>
+ store i32 %b, i32* %arrayidx1, align 4<br>
+ %arrayidx2 = getelementptr inbounds i32,
i32* %d, i32 4<br>
+ store i32 %c, i32* %arrayidx2, align 4<br>
+ tail call void @throws_1(i32 %a, i32 %b,
i32 %c) noreturn<br>
+ unreachable<br>
+}<br>
+<br>
+; Don't use STM: there is no available
register to store<br>
+; the address. We could transform this with
some extra math, but<br>
+; that currently isn't implemented.<br>
+; CHECK-LABEL: no_stm:<br>
+; CHECK: .save {r7, lr}<br>
+; CHECK: .setfp r7, sp<br>
+; CHECK: str r0,<br>
+; CHECK: str r1,<br>
+; CHECK: str r2,<br>
+; CHECK: bl throws_2<br>
+define void @no_stm(i32 %a, i32 %b, i32 %c,
i32* %d) local_unnamed_addr noreturn
"no-frame-pointer-elim"="true" {<br>
+entry:<br>
+ %arrayidx = getelementptr inbounds i32,
i32* %d, i32 2<br>
+ store i32 %a, i32* %arrayidx, align 4<br>
+ %arrayidx1 = getelementptr inbounds i32,
i32* %d, i32 3<br>
+ store i32 %b, i32* %arrayidx1, align 4<br>
+ %arrayidx2 = getelementptr inbounds i32,
i32* %d, i32 4<br>
+ store i32 %c, i32* %arrayidx2, align 4<br>
+ tail call void @throws_2(i32 %a, i32 %b,
i32 %c, i32* %d) noreturn<br>
+ unreachable<br>
+}<br>
+<br>
+<br>
+declare void @throws_1(i32, i32, i32)
noreturn<br>
+declare void @throws_2(i32, i32, i32, i32*)
noreturn<br>
+<br>
<br>
<br>
______________________________<wbr>_________________<br>
llvm-commits mailing list<br>
<a href="mailto:llvm-commits@lists.llvm.org"
target="_blank" moz-do-not-send="true">llvm-commits@lists.llvm.org</a><br>
<a
href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits"
rel="noreferrer" target="_blank"
moz-do-not-send="true">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/llvm-commits</a><br>
</blockquote>
</div>
<br>
</div>
</div>
</blockquote>
<p><br>
</p>
</div>
</div>
<span class="HOEnZb"><font color="#888888">
<pre class="m_-679428887874725126moz-signature" cols="72">--
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project</pre>
</font></span></div>
</blockquote>
</div>
<br>
</div>
</blockquote>
<p><br>
</p>
<pre class="moz-signature" cols="72">--
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project</pre>
</body>
</html>