<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><div><div class=""><font color="#000000" class="">I had thought to use the memset intrinsic, </font>unfortunately I did not succeed to cross compiling compiler-rt on my Mac.</div><div class=""><br class=""></div><div class="">Regards,</div><div class=""><br class=""></div><div class="">Pierre Molinaro</div><div class=""><br class=""></div><div><br class=""></div><br class=""><blockquote type="cite" class=""><div class="">Le 28 août 2015 à 17:22, John Criswell <<a href="mailto:jtcriswel@gmail.com" class="">jtcriswel@gmail.com</a>> a écrit :</div><br class="Apple-interchange-newline"><div class="">On 8/28/15 11:20 AM, devh8h wrote:<br class=""><blockquote type="cite" class="">It is a very basic "blink-led" program, on a Teensy 3.1. There is no operating system. The BSS clear function is called at boot. I would write a general BSS clear function, that behaves correctly even if the BSS section is empty.<br class=""></blockquote><br class="">I thought it was something like that.<br class=""><br class="">Let me know if the memset intrinsic approach works.<br class=""><br class="">Regards,<br class=""><br class="">John Criswell<br class=""><br class=""><blockquote type="cite" class=""><br class="">Thank,<br class=""><br class="">Pierre Molinaro<br class=""><br class=""><blockquote type="cite" class="">Le 28 août 2015 à 17:00, John Criswell <<a href="mailto:jtcriswel@gmail.com" class="">jtcriswel@gmail.com</a>> a écrit :<br class=""><br class="">On 8/28/15 10:52 AM, devh8h via llvm-dev wrote:<br class=""><blockquote type="cite" class="">Hi,<br class=""><br class="">I am writing a function that clears the BSS section on an Cortex-M4 embedded system.<br class=""></blockquote>I assume that, for some reason, the operating system is not demand-paging in zeroed memory. Is that correct?<br class=""><br class=""><blockquote type="cite" class="">The LLVM (version 3.7.0rc3) code I had wrote is :<br class="">;------------<br class="">target datalayout = "e-m:e-p:32:32-i64:64-v128:64:128-a:0:32-n32-S64"<br class="">target triple = "thumbv7em-none--eabi"<br class=""><br class="">@__bss_start = external global i32<br class="">@__bss_end = external global i32<br class=""><br class="">define void @clearBSS () nounwind {<br class="">entry:<br class=""> br label %bssLoopTest<br class=""> bssLoopTest:<br class=""> %p = phi i32* [@__bss_start, %entry], [%p.next, %bssLoop]<br class=""> %completed = icmp eq i32* %p, @__bss_end<br class=""> br i1 %completed, label %clearCompleted, label %bssLoop<br class=""> bssLoop:<br class=""> store i32 0, i32* %p, align 4<br class=""> %p.next = getelementptr inbounds i32, i32* %p, i32 1<br class=""> br label %bssLoopTest<br class=""> clearCompleted:<br class=""> ret void<br class="">}<br class="">;------------<br class=""><br class="">This code runs. But when I optimize it with :<br class=""><span class="Apple-tab-span" style="white-space:pre"> </span>opt -disable-simplify-libcalls -Os -S source.ll -o optimized.ll<br class=""><br class="">I get the following code for the @clearBSS function :<br class="">;------------<br class="">define void @clearBSS() nounwind {<br class="">entry:<br class=""> br label %bssLoop<br class=""><br class="">bssLoop: ; preds = %entry, %bssLoop<br class=""> %p1 = phi i32* [ @__bss_start, %entry ], [ %p.next, %bssLoop ]<br class=""> store i32 0, i32* %p1, align 4<br class=""> %p.next = getelementptr inbounds i32, i32* %p1, i32 1<br class=""> %completed = icmp eq i32* %p.next, @__bss_end<br class=""> br i1 %completed, label %clearCompleted, label %bssLoop<br class=""><br class="">clearCompleted: ; preds = %bssLoop<br class=""> ret void<br class="">}<br class="">;------------<br class="">The optimizer has transformed the while loop into a repeat until.<br class=""><br class="">I think it assumes the two variables @__bss_start and @__bss_end are distinct. But they are solved at link time, and they are the same if the BSS section is empty : in this case, the optimized function fails.<br class=""><br class="">Is there a way to prevent the optimizer to assume the two variables are distinct ? Or what is the proper way to deal with link time values ?<br class=""></blockquote>Have you tried using the memset intrinsic? You could case bss_start and bss_end to integers, subtract them to find the length, and then use memset to zero the memory. I would think memset should work if the length is zero.<br class=""><br class="">Regards,<br class=""><br class="">John Criswell<br class=""><br class=""><blockquote type="cite" class="">Thanks,<br class=""><br class="">Pierre Molinaro<br class=""><br class="">_______________________________________________<br class="">LLVM Developers mailing list<br class=""><a href="mailto:llvm-dev@lists.llvm.org" class="">llvm-dev@lists.llvm.org</a><br class="">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev<br class=""></blockquote><br class="">-- <br class="">John Criswell<br class="">Assistant Professor<br class="">Department of Computer Science, University of Rochester<br class=""><a href="http://www.cs.rochester.edu/u/criswell" class="">http://www.cs.rochester.edu/u/criswell</a><br class=""><br class=""></blockquote></blockquote><br class=""><br class="">-- <br class="">John Criswell<br class="">Assistant Professor<br class="">Department of Computer Science, University of Rochester<br class=""><a href="http://www.cs.rochester.edu/u/criswell" class="">http://www.cs.rochester.edu/u/criswell</a><br class=""><br class=""></div></blockquote></div><br class=""></body></html>