<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Thanks for the comments, will address them tomorrow.</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
I thought potential performance benefits mainly comes from data not being placed in a common section, which is placed in a copy-on-write memory page.</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
And saying this from memory, I believe potential code-size savings comes from enabling the linker to do a better job in eliminating unused function sections, but I'd need to double check this tomorrow.</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
But perhaps these arguments are actually less relevant than it being a more sane and language conforming default (and it affects only C and not C++).</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div><span style="font-family: calibri, arial, helvetica, sans-serif; font-size: 12pt; line-height: normal; color: rgb(0, 0, 0); background-color: rgba(0, 0, 0, 0);">And because GCC has already flipped the switch with GCC10, (some) porting has been done, and
some porting guides exists, for example:</span></div>
<div><a href="https://gcc.gnu.org/gcc-10/porting_to.html" id="LPNoLP687108">https://gcc.gnu.org/gcc-10/porting_to.html</a></div>
<div><span style="font-family: calibri, arial, helvetica, sans-serif; font-size: 12pt; line-height: normal; color: rgb(0, 0, 0); background-color: rgba(0, 0, 0, 0);">and</span><br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<a href="https://wiki.gentoo.org/wiki/Gcc_10_porting_notes/fno_common" id="LPNoLP537035">https://wiki.gentoo.org/wiki/Gcc_10_porting_notes/fno_common</a><br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div id="appendonsend"></div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>From:</b> cfe-dev <cfe-dev-bounces@lists.llvm.org> on behalf of Fangrui Song via cfe-dev <cfe-dev@lists.llvm.org><br>
<b>Sent:</b> 25 February 2020 17:47<br>
<b>To:</b> cfe-dev <cfe-dev@lists.llvm.org><br>
<b>Subject:</b> Re: [cfe-dev] [ARM][AArch64] Default to -fno-common</font>
<div> </div>
</div>
<div class="BodyFragment"><font size="2"><span style="font-size:11pt;">
<div class="PlainText">I am in favor of this change (-fno-common is the sane default; GCC HEAD (future 11) will default to -fno-common),<br>
and left some comments on <a href="https://reviews.llvm.org/D75056">https://reviews.llvm.org/D75056</a><br>
<br>
But I have questions about the two claims.<br>
<br>
For performance benefits:<br>
Is it because<br>
<br>
1. we missed some optimization for CommonLinkage<br>
2. CommonLinkage is considered interposable.<br>
<br>
For code-size reasons:<br>
I think it may actually be a very minor regression. Why does a common<br>
symbol cost more bytes?<br>
<br>
On 2020-02-25, Sjoerd Meijer via cfe-dev wrote:<br>
>Thanks for your comments and suggestions. I think we can continue the discussion in D75056, and I will prepare a revision that enables this for all targets.<br>
><br>
>Cheers,<br>
>Sjoerd.<br>
>________________________________<br>
>From: James Y Knight <jyknight@google.com><br>
>Sent: 24 February 2020 21:49<br>
>To: Sjoerd Meijer <Sjoerd.Meijer@arm.com><br>
>Cc: Tom Stellard <tstellar@redhat.com>; cfe-dev@lists.llvm.org Developers <cfe-dev@lists.llvm.org><br>
>Subject: Re: [cfe-dev] [ARM][AArch64] Default to -fno-common<br>
><br>
>IMO, it would be less confusing to just change the default to false across the board.<br>
><br>
>If anything should still use -fcommon by default, I'd want that to be based on vendor/OS, not CPU.<br>
><br>
>On Mon, Feb 24, 2020 at 11:38 AM Sjoerd Meijer via cfe-dev <cfe-dev@lists.llvm.org<mailto:cfe-dev@lists.llvm.org>> wrote:<br>
>> Is there any reason to limit this to just the ARM targets?<br>
><br>
>No good reason other than that I don't know how beneficial this is for other targets.<br>
>But I definitely don't mind giving other targets the same treatment though while I am at it.<br>
><br>
>Cheers,<br>
>Sjoerd.<br>
><br>
>________________________________<br>
>From: Tom Stellard <tstellar@redhat.com<mailto:tstellar@redhat.com>><br>
>Sent: 24 February 2020 16:31<br>
>To: Sjoerd Meijer <Sjoerd.Meijer@arm.com<mailto:Sjoerd.Meijer@arm.com>>; cfe-dev@lists.llvm.org<mailto:cfe-dev@lists.llvm.org> Developers <cfe-dev@lists.llvm.org<mailto:cfe-dev@lists.llvm.org>><br>
>Subject: Re: [cfe-dev] [ARM][AArch64] Default to -fno-common<br>
><br>
>On 02/24/2020 08:28 AM, Sjoerd Meijer via cfe-dev wrote:<br>
>> Hello,<br>
>><br>
>> Subject says most of it: I would like to propose that the Arm targets default to -fno-common for performance and code-size reasons. In addition, GCC has also flipped the switch recently, see [1], so we would get the same behaviour as GCC which would also
be nice. I've learned that the switch for GCC was relatively straightforward (for users), and there is some GNU documentation with Q&As to help toolchain users to transition if they find problems. I have put a draft patch up for review here:
<a href="https://reviews.llvm.org/D75056">https://reviews.llvm.org/D75056</a>. This definitely needs some doc changes (and I need to look at the tests), but first I would like to see how we feel about flipping the switch: any ideas, comments, objections?<br>
>><br>
><br>
>Is there any reason to limit this to just the ARM targets?<br>
><br>
>-Tom<br>
><br>
>> Cheers,<br>
>> Sjoerd.<br>
>><br>
>><br>
>> [1] The GCC commit:<br>
>><br>
>> commit 6271dd984d7f920d4fb17ad37af6a1f8e6b796dc<br>
>> Author: Wilco Dijkstra <wdijkstr@arm.com<mailto:wdijkstr@arm.com>><br>
>> Date: Wed Nov 20 16:29:23 2019 +0000<br>
>><br>
>> PR85678: Change default to -fno-common<br>
>> GCC currently defaults to -fcommon. As discussed in the PR, this is an ancient<br>
>> C feature which is not conforming with the latest C standards. On many targets<br>
>> this means global variable accesses have a codesize and performance penalty.<br>
>> This applies to C code only, C++ code is not affected by -fcommon. It is about<br>
>> time to change the default.<br>
_______________________________________________<br>
cfe-dev mailing list<br>
cfe-dev@lists.llvm.org<br>
<a href="https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev">https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev</a><br>
</div>
</span></font></div>
</body>
</html>