<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<meta name="Generator" content="Microsoft Word 15 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
{font-family:"Cambria Math";
panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
{font-family:Consolas;
panose-1:2 11 6 9 2 2 4 3 2 4;}
@font-face
{font-family:"Segoe UI";
panose-1:2 11 5 2 4 2 4 2 2 3;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:11.0pt;
font-family:"Calibri",sans-serif;}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
span.EmailStyle18
{mso-style-type:personal-reply;
font-family:"Calibri",sans-serif;
color:windowtext;}
.MsoChpDefault
{mso-style-type:export-only;
font-family:"Calibri",sans-serif;}
@page WordSection1
{size:8.5in 11.0in;
margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang="EN-US" link="blue" vlink="purple">
<div class="WordSection1">
<p class="MsoNormal">Thanks Craig!<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal"><b>From:</b> Craig Topper <craig.topper@gmail.com> <br>
<b>Sent:</b> Tuesday, January 26, 2021 12:06 PM<br>
<b>To:</b> Riyaz Puthiyapurayil <riyaz@synopsys.com><br>
<b>Cc:</b> Roman Lebedev <lebedev.ri@gmail.com>; llvm-dev@lists.llvm.org<br>
<b>Subject:</b> Re: [llvm-dev] LLVM 10 is miscompiling this code on x86_64 with an invalid shift count for `shrd`<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal">Which was flagged in post-commit review about a year later. <a href="https://urldefense.com/v3/__https:/reviews.llvm.org/D58475__;!!A4F2R9G_pg!LZwRdXlbAwQyD5-XlUqg2SEhRa_HhVybtA8bwSpe8ds1VQXwsCwI5QR-jHCTCqL9_U6F8ks7vCRy$">
https://reviews.llvm.org/D58475</a><o:p></o:p></p>
<div>
<p class="MsoNormal"><br clear="all">
<o:p></o:p></p>
<div>
<div>
<p class="MsoNormal">~Craig<o:p></o:p></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<div>
<p class="MsoNormal">On Tue, Jan 26, 2021 at 12:04 PM Craig Topper <<a href="mailto:craig.topper@gmail.com">craig.topper@gmail.com</a>> wrote:<o:p></o:p></p>
</div>
<blockquote style="border:none;border-left:solid #CCCCCC 1.0pt;padding:0in 0in 0in 6.0pt;margin-left:4.8pt;margin-right:0in">
<div>
<p class="MsoNormal">There's a typo in this line of X86InstrCompiler.td in llvm 10. It should be shiftMask64 not shiftMask32.<o:p></o:p></p>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<table class="MsoNormalTable" border="0" cellspacing="0" cellpadding="0" style="background:#0D1117;border-collapse:collapse;border-spacing:0px">
<tbody>
<tr style="box-sizing:border-box">
<td valign="top" style="background:transparent;padding:0in 7.5pt 0in 7.5pt;box-sizing:border-box;white-space:pre-wrap;overflow:visible" id="gmail-m_3301326694749536810gmail-LC1826">
<p class="MsoNormal" style="line-height:15.0pt"><span style="font-size:9.0pt;font-family:Consolas;color:#C9D1D9">// (shift x (and y, 63)) ==> (shift x, y)<o:p></o:p></span></p>
</td>
</tr>
<tr style="box-sizing:border-box">
<td width="50" nowrap="" valign="top" style="width:37.5pt;padding:0in 7.5pt 0in 7.5pt;box-sizing:border-box;min-width:50px" id="gmail-m_3301326694749536810gmail-L1827">
</td>
<td valign="top" style="padding:0in 7.5pt 0in 7.5pt;box-sizing:border-box;white-space:pre-wrap;overflow:visible" id="gmail-m_3301326694749536810gmail-LC1827">
<p class="MsoNormal" style="line-height:15.0pt"><span style="font-size:9.0pt;font-family:Consolas;color:#C9D1D9">def : Pat<(frag GR64:$src1, GR64:$src2, (shiftMask32 CL)),<o:p></o:p></span></p>
</td>
</tr>
<tr style="box-sizing:border-box">
<td width="50" nowrap="" valign="top" style="width:37.5pt;background:transparent;padding:0in 7.5pt 0in 7.5pt;box-sizing:border-box;min-width:50px" id="gmail-m_3301326694749536810gmail-L1828">
</td>
<td valign="top" style="background:transparent;padding:0in 7.5pt 0in 7.5pt;box-sizing:border-box;white-space:pre-wrap;overflow:visible" id="gmail-m_3301326694749536810gmail-LC1828">
<p class="MsoNormal" style="line-height:15.0pt"><span style="font-size:9.0pt;font-family:Consolas;color:#C9D1D9">(!cast<Instruction>(name # "64rrCL") GR64:$src1, GR64:$src2)>;<o:p></o:p></span></p>
</td>
</tr>
</tbody>
</table>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal"><br clear="all">
<o:p></o:p></p>
<div>
<div>
<p class="MsoNormal">~Craig<o:p></o:p></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<div>
<p class="MsoNormal">On Tue, Jan 26, 2021 at 11:57 AM Riyaz Puthiyapurayil via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>> wrote:<o:p></o:p></p>
</div>
<blockquote style="border:none;border-left:solid #CCCCCC 1.0pt;padding:0in 0in 0in 6.0pt;margin-left:4.8pt;margin-right:0in">
<p class="MsoNormal">Curiously, for the following test, with -fsanitize=undefined -O3, clang-10 reports no UB errors and produces the correct run-time result. But without -fsanitize=undefined, it produces wrong output. Clang (trunk) produces correct results
either way. I think there is definitely a bug that was fixed recently (in the last year or so). Could it be this one:<br>
<br>
<a href="https://urldefense.com/v3/__https:/reviews.llvm.org/D75748?id=__;!!A4F2R9G_pg!LZwRdXlbAwQyD5-XlUqg2SEhRa_HhVybtA8bwSpe8ds1VQXwsCwI5QR-jHCTCqL9_U6F8nTqMJUS$" target="_blank">https://reviews.llvm.org/D75748?id=</a><br>
<br>
Roman, you were a reviewer for this.<br>
<br>
-----Original Message-----<br>
From: Riyaz Puthiyapurayil <br>
Sent: Tuesday, January 26, 2021 11:48 AM<br>
To: 'Roman Lebedev' <<a href="mailto:lebedev.ri@gmail.com" target="_blank">lebedev.ri@gmail.com</a>><br>
Cc: <a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
Subject: RE: [llvm-dev] LLVM 10 is miscompiling this code on x86_64 with an invalid shift count for `shrd`<br>
<br>
My bad! There was a typo that caused the UBSAN error. Here is the corrected test. This time UBSAN reports no errors. But clang-10 still produces unexpected result whereas trunk does not.<br>
<br>
#include <stdio.h><br>
#include <stdint.h><br>
#include <string.h><br>
<br>
__attribute__ ((noinline)) uint64_t foo4(__int128_t* arr1) {<br>
unsigned chunk = 34 >> 5;<br>
unsigned bitIndex = 34 - chunk * 32;<br>
<br>
unsigned* arr = (unsigned *)arr1 + chunk;<br>
__int128_t v1 = *arr1;<br>
v1 = v1 >> bitIndex;<br>
<br>
return (uint64_t) v1;<br>
}<br>
<br>
__attribute__ ((noinline)) uint64_t foo3(__int128_t* arr1, unsigned index) {<br>
unsigned chunk = index >> 5;<br>
unsigned bitIndex = index - chunk * 32;<br>
<br>
unsigned *arr = (unsigned *)arr1 + chunk;<br>
__int128_t v1 = *arr1;<br>
v1 = v1 >> bitIndex;<br>
<br>
return (uint64_t) v1;<br>
}<br>
<br>
int main() {<br>
<br>
unsigned arr[4];<br>
<br>
arr[0]= 0x6f7cfd6f;<br>
arr[1]= 0xd96c9806;<br>
arr[2]= 0x89420144;<br>
arr[3]= 0x8a20f548;<br>
<br>
__int128_t arr1;<br>
<br>
memcpy(&arr1, arr, sizeof(arr1));<br>
<br>
printf("foo3 %lx\n",foo3(&arr1,34) & 0x3ffffffff); <br>
printf("foo4 %lx\n",foo4(&arr1) & 0x3ffffffff); }<br>
<br>
-----Original Message-----<br>
From: Roman Lebedev <<a href="mailto:lebedev.ri@gmail.com" target="_blank">lebedev.ri@gmail.com</a>><br>
Sent: Tuesday, January 26, 2021 11:31 AM<br>
To: Riyaz Puthiyapurayil <<a href="mailto:riyaz@synopsys.com" target="_blank">riyaz@synopsys.com</a>><br>
Cc: <a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
Subject: Re: [llvm-dev] LLVM 10 is miscompiling this code on x86_64 with an invalid shift count for `shrd`<br>
<br>
<a href="https://urldefense.com/v3/__https:/godbolt.org/z/5bcsj9__;!!A4F2R9G_pg!P3wQMqAfQM_jzlQy2F-1LD2x1ekeAKGQckArJiV5UsQIR2NhtjHbH5ivGPqtfl8bZLzgtjq5KIG7$" target="_blank">https://urldefense.com/v3/__https://godbolt.org/z/5bcsj9__;!!A4F2R9G_pg!P3wQMqAfQM_jzlQy2F-1LD2x1ekeAKGQckArJiV5UsQIR2NhtjHbH5ivGPqtfl8bZLzgtjq5KIG7$</a>
<br>
<br>
example.c:43:37: runtime error: subtraction of unsigned offset from<br>
0x000000000022 overflowed to 0x7fffd01217c8<br>
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior example.c:43:37 in<br>
example.c:19:37: runtime error: applying non-zero offset<br>
140576899311424 to null pointer<br>
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior example.c:19:37 in<br>
foo3 19bdf3f5b<br>
foo4 19bdf3f5b<br>
<br>
On Tue, Jan 26, 2021 at 10:29 PM Riyaz Puthiyapurayil <<a href="mailto:Riyaz.Puthiyapurayil@synopsys.com" target="_blank">Riyaz.Puthiyapurayil@synopsys.com</a>> wrote:<br>
><br>
> Apologies, outlook messed up the formatting in the previous email. Here is it again.<br>
><br>
><br>
><br>
> Do you see any UB in the following modified test? Here also, I see a difference between foo3 and foo4. Again, trunk and gcc match. But not Clang-10.<br>
><br>
><br>
><br>
> #include <stdio.h><br>
><br>
> #include <stdint.h><br>
><br>
> #include <string.h><br>
><br>
><br>
><br>
> __attribute__ ((noinline)) uint64_t foo4(__int128_t* arr1)<br>
><br>
> {<br>
><br>
> unsigned chunk = 34 >> 5;<br>
><br>
> unsigned bitIndex = 34 - chunk * 32;<br>
><br>
><br>
><br>
> unsigned* arr = (unsigned *)arr + chunk;<br>
><br>
> __int128_t v1 = *arr1;<br>
><br>
> v1 = v1 >> bitIndex;<br>
><br>
><br>
><br>
> return (uint64_t) v1;<br>
><br>
> }<br>
><br>
><br>
><br>
> __attribute__ ((noinline)) uint64_t foo3(__int128_t* arr1, unsigned<br>
> index)<br>
><br>
> {<br>
><br>
> unsigned chunk = index >> 5;<br>
><br>
> unsigned bitIndex = index - chunk * 32;<br>
><br>
><br>
><br>
> unsigned *arr = (unsigned *)arr + chunk;<br>
><br>
> __int128_t v1 = *arr1;<br>
><br>
> v1 = v1 >> bitIndex;<br>
><br>
><br>
><br>
> return (uint64_t) v1;<br>
><br>
> }<br>
><br>
><br>
><br>
> int main() {<br>
><br>
><br>
><br>
> unsigned arr[4];<br>
><br>
><br>
><br>
> arr[0]= 0x6f7cfd6f;<br>
><br>
> arr[1]= 0xd96c9806;<br>
><br>
> arr[2]= 0x89420144;<br>
><br>
> arr[3]= 0x8a20f548;<br>
><br>
><br>
><br>
> __int128_t arr1;<br>
><br>
><br>
><br>
> memcpy(&arr1, arr, sizeof(arr1));<br>
><br>
><br>
><br>
> printf("foo3 %lx\n",foo3(&arr1,34) & 0x3ffffffff);<br>
><br>
> printf("foo4 %lx\n",foo4(&arr1) & 0x3ffffffff);<br>
><br>
> }<br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
> From: Riyaz Puthiyapurayil<br>
> Sent: Tuesday, January 26, 2021 11:25 AM<br>
> To: 'Roman Lebedev' <<a href="mailto:lebedev.ri@gmail.com" target="_blank">lebedev.ri@gmail.com</a>><br>
> Cc: <a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
> Subject: RE: [llvm-dev] LLVM 10 is miscompiling this code on x86_64 <br>
> with an invalid shift count for `shrd`<br>
><br>
><br>
><br>
> Do you see any UB in the following modified test? Here also, I see a <br>
> difference between foo3 and foo4. Again, trunk and gcc match. But not<br>
> Clang-10<br>
><br>
><br>
><br>
> #include <stdio.h><br>
><br>
> #include <stdint.h><br>
><br>
> #include <string.h><br>
><br>
><br>
><br>
> __attribute__ ((noinline)) uint64_t foo4(__int128_t* arr1)<br>
><br>
> {<br>
><br>
> unsigned chunk = 34 >> 5;<br>
><br>
> unsigned bitIndex = 34 - chunk * 32;<br>
><br>
><br>
><br>
> unsigned* arr = (unsigned *)arr + chunk;<br>
><br>
> __int128_t v1 = *arr1;<br>
><br>
> v1 = v1 >> bitIndex;<br>
><br>
><br>
><br>
> return (uint64_t) v1;<br>
><br>
> }<br>
><br>
><br>
><br>
> __attribute__ ((noinline)) uint64_t foo3(__int128_t* arr1, unsigned<br>
> index)<br>
><br>
> {<br>
><br>
> unsigned chunk = index >> 5;<br>
><br>
> unsigned bitIndex = index - chunk * 32;<br>
><br>
><br>
><br>
> unsigned *arr = (unsigned *)arr + chunk;<br>
><br>
> __int128_t v1 = *arr1;<br>
><br>
> v1 = v1 >> bitIndex;<br>
><br>
><br>
><br>
> return (uint64_t) v1;<br>
><br>
> }<br>
><br>
><br>
><br>
> int main() {<br>
><br>
><br>
><br>
> unsigned arr[4];<br>
><br>
><br>
><br>
> arr[0]= 0x6f7cfd6f;<br>
><br>
> arr[1]= 0xd96c9806;<br>
><br>
> arr[2]= 0x89420144;<br>
><br>
> arr[3]= 0x8a20f548;<br>
><br>
><br>
><br>
> __int128_t arr1;<br>
><br>
><br>
><br>
> memcpy(&arr1, arr, sizeof(arr1));<br>
><br>
><br>
><br>
> printf("foo3 %lx\n",foo3(&arr1,34) & 0x3ffffffff); // prints<br>
> 1365b2601 (clang-10) 19bdf3f5b (gcc, clang-12)<br>
><br>
> printf("foo4 %lx\n",foo4(&arr1) & 0x3ffffffff); // prints 19bdf3f5b<br>
><br>
> }<br>
><br>
><br>
><br>
> -----Original Message-----<br>
> From: Riyaz Puthiyapurayil<br>
> Sent: Tuesday, January 26, 2021 11:10 AM<br>
> To: 'Roman Lebedev' <<a href="mailto:lebedev.ri@gmail.com" target="_blank">lebedev.ri@gmail.com</a>><br>
> Subject: RE: [llvm-dev] LLVM 10 is miscompiling this code on x86_64 <br>
> with an invalid shift count for `shrd`<br>
><br>
><br>
><br>
> You may be right about that the cast from unsigned* to __int128_t* being UB. But I think that has nothing to do with the bug. That code can be fixed to use a __int128_t array and change all the casts from __int128_t* to unsigned* (instead of the other way
around) and we would still have the issue.<br>
><br>
><br>
><br>
> -----Original Message-----<br>
><br>
> From: Roman Lebedev <<a href="mailto:lebedev.ri@gmail.com" target="_blank">lebedev.ri@gmail.com</a>><br>
><br>
> Sent: Tuesday, January 26, 2021 10:34 AM<br>
><br>
> To: Riyaz Puthiyapurayil <<a href="mailto:riyaz@synopsys.com" target="_blank">riyaz@synopsys.com</a>><br>
><br>
> Cc: <a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
><br>
> Subject: Re: [llvm-dev] LLVM 10 is miscompiling this code on x86_64 <br>
> with an invalid shift count for `shrd`<br>
><br>
><br>
><br>
> On Tue, Jan 26, 2021 at 9:28 PM Riyaz Puthiyapurayil via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>> wrote:<br>
><br>
> ><br>
><br>
> > It appears that LLVM 10 is miscompiling the following test on x86_64 but the trunk version seems to work fine. Before I spend any time debugging this, it would be helpful if someone can point me to a bug fix if there was one.<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > #include <stdio.h><br>
><br>
> ><br>
><br>
> > #include <stdint.h><br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > __attribute__ ((noinline)) uint64_t foo4(unsigned* arr)<br>
><br>
> ><br>
><br>
> > {<br>
><br>
> ><br>
><br>
> > unsigned chunk = 34 >> 5;<br>
><br>
> ><br>
><br>
> > unsigned bitIndex = 34 - chunk * 32;<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > arr = arr + chunk;<br>
><br>
> ><br>
><br>
> > __int128_t* arr1 = (__int128_t*) arr;<br>
><br>
> I'm pretty sure this is UB, because the alignment of __int128_t is bigger than that of unsigned.<br>
><br>
><br>
><br>
> ><br>
><br>
> > __int128_t v1 = *arr1;<br>
><br>
> ><br>
><br>
> > v1 = v1 >> bitIndex;<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > return (uint64_t) v1;<br>
><br>
> ><br>
><br>
> > }<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > __attribute__ ((noinline)) uint64_t foo3(unsigned* arr, unsigned<br>
><br>
> > index)<br>
><br>
> ><br>
><br>
> > {<br>
><br>
> ><br>
><br>
> > unsigned chunk = index >> 5;<br>
><br>
> ><br>
><br>
> > unsigned bitIndex = index - chunk * 32;<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > arr = arr + chunk;<br>
><br>
> ><br>
><br>
> > __int128_t* arr1 = (__int128_t*) arr;<br>
><br>
> ><br>
><br>
> > __int128_t v1 = *arr1;<br>
><br>
> ><br>
><br>
> > v1 = v1 >> bitIndex;<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > return (uint64_t) v1;<br>
><br>
> ><br>
><br>
> > }<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > int main() {<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > unsigned arr[5];<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > arr[0]= 0x6f7cfd6f;<br>
><br>
> ><br>
><br>
> > arr[1]= 0xd96c9806;<br>
><br>
> ><br>
><br>
> > arr[2]= 0x89420144;<br>
><br>
> ><br>
><br>
> > arr[3]= 0x8a20f548;<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > printf("foo3 %lx\n",foo3(arr,34) & 0x3ffffffff); // prints<br>
><br>
> > 222508051 instead of 1365b2601<br>
><br>
> ><br>
><br>
> > printf("foo4 %lx\n",foo4(arr) & 0x3ffffffff); // prints<br>
> > 1365b2601<br>
><br>
> > as expected<br>
><br>
> ><br>
><br>
> > }<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > If you look at the assembly, there is an ‘and cl, 31’ missing in foo3 which is needed to make sure that the shift count is in the range [0...31]. This is missing in the generated code. Note that if the count in CL register is greater than 31, x86_64 manual
states that the behavior of shrd is undefined.<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > # clang 10<br>
><br>
> ><br>
><br>
> > foo3(unsigned int*, unsigned int): # @foo3(unsigned int*, unsigned int)<br>
><br>
> ><br>
><br>
> > mov ecx, esi<br>
><br>
> ><br>
><br>
> > mov edx, esi<br>
><br>
> ><br>
><br>
> > shr edx, 5<br>
><br>
> ><br>
><br>
> > mov rax, qword ptr [rdi + 4*rdx]<br>
><br>
> ><br>
><br>
> > mov rdx, qword ptr [rdi + 4*rdx + 8]<br>
><br>
> ><br>
><br>
> > shrd rax, rdx, cl<br>
><br>
> ><br>
><br>
> > ret<br>
><br>
> ><br>
><br>
> ><br>
><br>
> ><br>
><br>
> > # Trunk<br>
><br>
> ><br>
><br>
> > foo3(unsigned int*, unsigned int): # @foo3(unsigned int*, unsigned int)<br>
><br>
> ><br>
><br>
> > mov ecx, esi<br>
><br>
> ><br>
><br>
> > mov edx, esi<br>
><br>
> ><br>
><br>
> > shr edx, 5<br>
><br>
> ><br>
><br>
> > mov rax, qword ptr [rdi + 4*rdx]<br>
><br>
> ><br>
><br>
> > mov rdx, qword ptr [rdi + 4*rdx + 8]<br>
><br>
> ><br>
><br>
> > and cl, 31<br>
><br>
> ><br>
><br>
> > shrd rax, rdx, cl<br>
><br>
> ><br>
><br>
> > ret<br>
><br>
> ><br>
><br>
><br>
><br>
> Roman.<br>
><br>
><br>
><br>
> > _______________________________________________<br>
><br>
> > LLVM Developers mailing list<br>
><br>
> > <a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
><br>
> > <a href="https://urldefense.com/v3/__https:/lists.llvm.org/cgi-bin/mailman/l" target="_blank">
https://urldefense.com/v3/__https://lists.llvm.org/cgi-bin/mailman/l</a><br>
> > is<br>
><br>
> > tinfo/llvm-dev__;!!A4F2R9G_pg!Mq1ZgXCY7kwDMwNwdn4Tcz-KgkoXMlP_rjq7_U<br>
> > 9M<br>
><br>
> > rh1gbrtqvDoSi6cMuSLZaXIPgtijPW55JJ_8$<br>
_______________________________________________<br>
LLVM Developers mailing list<br>
<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
<a href="https://urldefense.com/v3/__https:/lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev__;!!A4F2R9G_pg!LZwRdXlbAwQyD5-XlUqg2SEhRa_HhVybtA8bwSpe8ds1VQXwsCwI5QR-jHCTCqL9_U6F8kv5Y8Tw$" target="_blank">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><o:p></o:p></p>
</blockquote>
</div>
</blockquote>
</div>
</div>
</body>
</html>