<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=""><br class=""><div><blockquote type="cite" class=""><div class="">On Jul 12, 2016, at 12:20 PM, Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" class="">c_plawre@qca.qualcomm.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div class="WordSection1" style="page: WordSection1; font-family: Helvetica; font-size: 12px; font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Mehdi,<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> I am looking for an understanding of 1) IPRA in general, 2) IPRA in LLVM.<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Whether I want to use LTO or not is a separate issue.<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">1) I currently believe it is a true statement that:<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> If all external functions are known to not call back into the “whole-program”<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> Being compiled, then IPRA is free to do anything at all to the functions being<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> Compiled, not limited to only “upgrades” calling convention changes, but<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> Also allowing “downgrades” calling convention changes as well.<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Do you think my current belief #1 is correct ?</span></div></div></div></blockquote><div><br class=""></div><div>Yes, with some extra assumptions (you don’t use dlsym for instance, and you won’t link to another file with a global initializer that can call any of these).</div><div><br class=""></div><div>I expressed this earlier (which include the other issues I mentioned just before) as “we can turn the linkage of every function into local” (or private, or static, whatever denomination you prefer).</div><div><br class=""></div><br class=""><blockquote type="cite" class=""><div class="WordSection1" style="page: WordSection1; font-family: Helvetica; font-size: 12px; font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">2) it seems that LLVM currently limits itself to “upgrades” calling convention changes,<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">The reason being so that not all call sites are required to be changed,<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">therefore calls through function pointers can use the default calling convention<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">If for example there is insufficient analysis to know for sure what functions can be<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">called from that site.<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Is my understanding #2 of IPRA in LLVM correct ?</span></div></div></blockquote><div><br class=""></div><div><br class=""></div><div>I don’t believe this is correct, currently IPRA will limit itself to this for function that can be called from another module.</div><div>I will freely change the calling convention, including downgrades, when it knows that it can see all call sites (+ extra conditions, like no recursion being involved I think).</div><div><br class=""></div><div><br class=""></div><blockquote type="cite" class=""><div class="WordSection1" style="page: WordSection1; font-family: Helvetica; font-size: 12px; font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="color: rgb(31, 73, 125); font-family: Calibri, sans-serif; font-size: 11pt;" class=""> </span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">“whole-program” here is a misnomer since there are external functions, but I don’t<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Have a better term for this.</span></div></div></blockquote><div><br class=""></div><div>I believe you can talk about “main module”, i.e. the module defines the entry point for the program.</div><div>Note LLVM can’t make assumption about the lack of dlsym() or global initializer in other module for example, so the linkage type of functions is what tells us about the possibility to call back or not.</div><div><br class=""></div><div><br class=""></div><div>— </div><div>Mehdi</div><div><br class=""></div><div><br class=""></div><div><br class=""></div><blockquote type="cite" class=""><div class=""><div class="WordSection1" style="page: WordSection1; font-family: Helvetica; font-size: 12px; font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">“upgrades” means some scratch regs are converted to save<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; text-indent: 0.5in;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">(the callee either doesn’t touch them at all, or does do save/restore)<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">“downgrades” means some save regs are converted to scratch<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> (the callee no longer does save/restore to some registers, and does clobber them)<o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""><o:p class=""> </o:p></span></div><div class=""><div style="border-style: solid none none; border-top-color: rgb(225, 225, 225); border-top-width: 1pt; padding: 3pt 0in 0in;" class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><b class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">From:</span></b><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class=""><span class="Apple-converted-space"> </span><a href="mailto:mehdi.amini@apple.com" class="">mehdi.amini@apple.com</a> [<a href="mailto:mehdi.amini@apple.com" class="">mailto:mehdi.amini@apple.com</a>]<span class="Apple-converted-space"> </span><br class=""><b class="">Sent:</b><span class="Apple-converted-space"> </span>Monday, July 11, 2016 8:41 PM<br class=""><b class="">To:</b><span class="Apple-converted-space"> </span>Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" class="">c_plawre@qca.qualcomm.com</a>><br class=""><b class="">Cc:</b><span class="Apple-converted-space"> </span>vivek pandya <<a href="mailto:vivekvpandya@gmail.com" class="">vivekvpandya@gmail.com</a>>; llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" class="">llvm-dev@lists.llvm.org</a>>; <a href="mailto:llvm-dev-request@lists.llvm.org" class="">llvm-dev-request@lists.llvm.org</a>; Hal Finkel <<a href="mailto:hfinkel@anl.gov" class="">hfinkel@anl.gov</a>><br class=""><b class="">Subject:</b><span class="Apple-converted-space"> </span>Re: [llvm-dev] IPRA, interprocedural register allocation, question<o:p class=""></o:p></span></div></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><br class=""><br class="">Sent from my iPhone<o:p class=""></o:p></div></div><div class=""><p class="MsoNormal" style="margin: 0in 0in 12pt; font-size: 12pt; font-family: 'Times New Roman', serif;"><br class="">On Jul 11, 2016, at 7:48 PM, Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" style="color: purple; text-decoration: underline;" class="">c_plawre@qca.qualcomm.com</a>> wrote:<o:p class=""></o:p></p></div><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Mehdi,</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> I’m compiling embedded applications which are small enough to do</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">whole-program-compilation. There’s no advantage in breaking them up into</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">separate compilation pieces and linking them, even though in source form</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">they are composed of a couple of separate source files.</span><o:p class=""></o:p></div></div></blockquote><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Ok, so LTO case basically.<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><br class=""><br class=""><o:p class=""></o:p></div><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">So for me the compilation unit is always the entire program (and includes main())</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Except for some hand-coded-assembly-language support functions that are “external”</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">to the compilation unit and in my case never call back into the compilation unit,</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">IE they are always “leaf” functions from the point of view of the compilation unit’s call-graph.</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Hence I would like a clang function attribute that says this function is “leaf”</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">So that IPRA can know that none of the functions it is compiling is ever called</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">From outside this compilation unit.</span><o:p class=""></o:p></div></div></blockquote><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I believe the usual (and best way from the compiler point of view) way to address your particular scenario is to have a proper export list and use LTO.<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">For instance if you never call into the program from one of your hand-coded assembly routines, LTO should be able to turn every global functions/variables into local ones.<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">And I apologize to everyone for confusingly using the term “compilation unit”</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">When I meant “whole program”.</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Yes I am aware of the fact that if you change a function’s calling convention</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">By converting some scratch regs into save regs (for example because they aren’t even touched)</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Then you are safe to call it from either the default calling convention or the</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Optimized calling convention. This is the safe thing to do, and is why I will</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Only use “preserves_most” and “preserves_all” optimized calling conventions,</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">As those will have been implemented by a back-end writer who is aware of</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">All these compilations (as opposed to the “registermask=” calling convention</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Which is much less safe)</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">I do however feel that IPRA in the whole-program case should not be restricted to</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Only scratch-becoming-save changes, I don’t have any data to support the notion,</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">But it begs to be investigated, unless someone can somehow prove that it can’t help</span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Performance.</span><o:p class=""></o:p></div></blockquote><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Beside an attribute on declarations, what do you suggest exactly?<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">-- <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Mehdi<o:p class=""></o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><br class=""><br class=""><o:p class=""></o:p></div><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div><div class=""><div style="border-style: solid none none; border-top-color: rgb(225, 225, 225); border-top-width: 1pt; padding: 3pt 0in 0in;" class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><b class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">From:</span></b><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class=""><span class="Apple-converted-space"> </span><a href="mailto:mehdi.amini@apple.com" style="color: purple; text-decoration: underline;" class="">mehdi.amini@apple.com</a><span class="Apple-converted-space"> </span>[<a href="mailto:mehdi.amini@apple.com" style="color: purple; text-decoration: underline;" class="">mailto:mehdi.amini@apple.com</a>]<span class="Apple-converted-space"> </span><br class=""><b class="">Sent:</b><span class="Apple-converted-space"> </span>Monday, July 11, 2016 7:06 PM<br class=""><b class="">To:</b><span class="Apple-converted-space"> </span>Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" style="color: purple; text-decoration: underline;" class="">c_plawre@qca.qualcomm.com</a>><br class=""><b class="">Cc:</b><span class="Apple-converted-space"> </span>vivek pandya <<a href="mailto:vivekvpandya@gmail.com" style="color: purple; text-decoration: underline;" class="">vivekvpandya@gmail.com</a>>; llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" style="color: purple; text-decoration: underline;" class="">llvm-dev@lists.llvm.org</a>>;<span class="Apple-converted-space"> </span><a href="mailto:llvm-dev-request@lists.llvm.org" style="color: purple; text-decoration: underline;" class="">llvm-dev-request@lists.llvm.org</a>; Hal Finkel <<a href="mailto:hfinkel@anl.gov" style="color: purple; text-decoration: underline;" class="">hfinkel@anl.gov</a>><br class=""><b class="">Subject:</b><span class="Apple-converted-space"> </span>Re: [llvm-dev] IPRA, interprocedural register allocation, question</span><o:p class=""></o:p></div></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div><div class=""><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">On Jul 11, 2016, at 6:45 PM, Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" style="color: purple; text-decoration: underline;" class="">c_plawre@qca.qualcomm.com</a>> wrote:<o:p class=""></o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Vivek,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> Here’s the way I see it, let me know if you agree or disagree,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">You cannot optimize a function’s calling convention (register-usage) unless</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">You can see and change every caller,</span><o:p class=""></o:p></div></div></div></blockquote><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">That’s true only if you want to “downgrade” the guarantees, i.e. if you want to reduce the callee-saved registers.<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">You can freely provide more information to limit the amount of caller-saved registers to a partial list of call-sites, which is in practice changing the “local" calling convention while keeping it compatible with the public one.<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><br class=""><br class=""><br class=""><o:p class=""></o:p></div><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">and you only know this for non-static functions</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">if you know that all calls to external functions cannot call back into the current</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">compilation unit.</span><o:p class=""></o:p></div></div></div></blockquote><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I’m not sure why you consider calls to external functions and call back? If you don’t see main() (the common case) you don’t need a call to an external function to have a possible call to an externally visible function in the current module.<o:p class=""></o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><br class=""><br class=""><br class=""><o:p class=""></o:p></div><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">#1 gives you the info necessary to change the call-site to the external function</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">So you don’t need #2 to do RA around the call-site to the external function, instead</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">You need #2 before you can change any non-static function’s calling convention</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">within the current compilation unit, assuming you have this information for all</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">external functions.</span><o:p class=""></o:p></div></div></div></blockquote><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">If I understand the case you have in mind, it is only when you see the main() function in the current module and you’re trying to prove that an externally visible function could not be called from outside the module basically?<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">It seems to me that this is a bit orthogonal to IPRA: multiple optimizations (IPRA included) work best when functions are deduced local, non-recursive, are not tail called (for IPRA in particular), and don’t have their address taken. <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">The “infer-func-attr” and “globalopt” passes try to do their best to make this happen, especially during LTO.<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">The attribute case that Vivek is adding seems more murky though.<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">— <o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Mehdi<o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><br class=""><br class=""><br class=""><o:p class=""></o:p></div><blockquote style="margin-top: 5pt; margin-bottom: 5pt;" class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">To be more concrete, let foo() be a non-static function in the current compilation</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Unit, any calls to foo() from external functions will have to use the “default”</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Calling convention, so foo’s calling convention cannot be changed. We have to</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Know that none of the external functions can call-back to the compilation unit</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">(they are “leaf” functions relative to the compilation unit) before we can change</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Foo()’s calling convention.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Also, the issue of escaping-pointer-to-function is made clear by the example</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Of the atexit() and exit() library functions, IE even static functions can end up</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Being called by external functions. So exit() can never be declared “leaf”, and</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">To get the benefit of IPRA it needs to be within the compilation unit, either</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">By whole-program compilation or by LTO, if it is used.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">--Peter.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><b class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">From:</span></b><span class="apple-converted-space"><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class=""> </span></span><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">vivek pandya [<a href="mailto:vivekvpandya@gmail.com" style="color: purple; text-decoration: underline;" class="">mailto:vivekvpandya@gmail.com</a>]<span class="apple-converted-space"> </span><br class=""><b class="">Sent:</b><span class="apple-converted-space"> </span>Friday, July 08, 2016 9:26 PM<br class=""><b class="">To:</b><span class="apple-converted-space"> </span>Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" style="color: purple; text-decoration: underline;" class="">c_plawre@qca.qualcomm.com</a>><br class=""><b class="">Cc:</b><span class="apple-converted-space"> </span>llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" style="color: purple; text-decoration: underline;" class="">llvm-dev@lists.llvm.org</a>>;<span class="Apple-converted-space"> </span><a href="mailto:llvm-dev-request@lists.llvm.org" style="color: purple; text-decoration: underline;" class="">llvm-dev-request@lists.llvm.org</a>; Hal Finkel <<a href="mailto:hfinkel@anl.gov" style="color: purple; text-decoration: underline;" class="">hfinkel@anl.gov</a>>; Tim Amini Golling <<a href="mailto:mehdi.amini@apple.com" style="color: purple; text-decoration: underline;" class="">mehdi.amini@apple.com</a>><br class=""><b class="">Subject:</b><span class="apple-converted-space"> </span>Re: Re:[llvm-dev] IPRA, interprocedural register allocation, question</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">On Sat, Jul 9, 2016 at 8:15 AM, Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="color: purple;" class="">c_plawre@qca.qualcomm.com</span></a>> wrote:<o:p class=""></o:p></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Vivek,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> IIUC it seems that we need two pieces of information to do IPRA,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">1. what registers the callee clobbers</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">2. what the callee does to the call-graph</span><o:p class=""></o:p></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Yes I think this is enough, but in your case we don't require #2 <o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">And it is #2 that we are missing when we define an external function,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Even when we declare it with a preserves or a regmask attribute,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Because I think once we have effect of attribute at IR/MI level then we can just parse it and populate register usage information vector for declared function and then we can propagate reg mask on each call site encountered.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">But I am not user will it be easy to get new attribute working or we may need to hack clang for that too.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I would also like to have thoughts from my mentors (Mehdi Amini and Hal Finkel) about this.<o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">So what I / we need is another attribute that says this is a leaf function,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">At least in my case all I’m really concerned with are leaf functions</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I am stating with a simple function declaration which have a custom attribute.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">-Vivek<o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Thoughts ?</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">--Peter Lawrence.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><b class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">From:</span></b><span class="apple-converted-space"><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class=""> </span></span><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">vivek pandya [mailto:</span><a href="mailto:vivekvpandya@gmail.com" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">vivekvpandya@gmail.com</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">]<span class="apple-converted-space"> </span><br class=""><b class="">Sent:</b><span class="apple-converted-space"> </span>Friday, July 08, 2016 10:24 AM<br class=""><b class="">To:</b><span class="apple-converted-space"> </span>Lawrence, Peter <</span><a href="mailto:c_plawre@qca.qualcomm.com" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">c_plawre@qca.qualcomm.com</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">><br class=""><b class="">Cc:</b><span class="apple-converted-space"> </span>llvm-dev <</span><a href="mailto:llvm-dev@lists.llvm.org" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">llvm-dev@lists.llvm.org</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">>;<span class="apple-converted-space"> </span></span><a href="mailto:llvm-dev-request@lists.llvm.org" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">llvm-dev-request@lists.llvm.org</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class=""><br class=""><b class="">Subject:</b><span class="apple-converted-space"> </span>Re: Re:[llvm-dev] IPRA, interprocedural register allocation, question</span><o:p class=""></o:p></div></div><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">On Fri, Jul 8, 2016 at 1:42 PM, vivek pandya <<a href="mailto:vivekvpandya@gmail.com" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="color: purple;" class="">vivekvpandya@gmail.com</span></a>> wrote:<o:p class=""></o:p></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">On Fri, Jul 8, 2016 at 9:47 AM, Lawrence, Peter <<a href="mailto:c_plawre@qca.qualcomm.com" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="color: purple;" class="">c_plawre@qca.qualcomm.com</span></a>> wrote:<o:p class=""></o:p></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Vivek,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> I am looking into these function attributes in the clang docs</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> Preserve_most</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> Preserve_all</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">They are not available in the 3.6.2 that I am currently using, but I hope they exist in 3.8</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">These should provide enough info to solve my problem,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">at the MC level calls to functions with these attributes</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">with be code-gen’ed through different “calling conventions”,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">and CALL instructions to them should have different register USE and DEF info,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Yes I believe that preserve_most or preserve_all should help you even with out IPRA. But just to note IPRA can even help further for example on X86 preserve_most cc will not preserve R11 (this can be verified from X86CallingConv.td and X86RegisterInfo.cpp) how ever IPAR calculates regmask based on the actual register usage and if procedure with preserve_most cc does not use R11 and none callsite inside of function body then IPRA will mark R11 as preserved. Also IPRA produces RegMask which is super set of RegMask due to calling convention.<o:p class=""></o:p></div></div></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I believe that __attribute__ ((registermask = ....)) can provide more flexibility compare to preserve_all or preserve_most CC in some case. So believe that we should try it out.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">-Vivek<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">This CALL instruction register USE and DEF info should already be useful</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">to the intra-procedural register allocator (allowing values live across these</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">calls to be in what are otherwise caller-save registers),</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">at least that’s how I read the MC dumps, every call instruction seems to have</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">every caller-save register flagged as “imp-def”, IE implicitly-defined by the instruction,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">and hopefully what is considered a caller-save register at a call-site is defined by the callee.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">And this should be the information that IPRA takes advantage of in its bottom-up analysis.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Yes that is expected help from IPRA. <o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Which leads me to this question, when compiling an entire whole program at one time,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">so there is no linking and no LTO, will there ever be IPRA that works within LLC for this scenario,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">and is this an objective of your project, or are you focusing only on LTO ?</span><o:p class=""></o:p></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">The current IPRA infrastructure works at compile time so it's scope of optimization is restricted to a compilation unit. So IPRA can only construct correct register usage information if the procedure's code is generated by same compiler instance that means we can't optimize library calls or procedure defined in other module. This is because we can't keep register usage information data across two different compiler instance. <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Now if we consider LTO, it eliminates above limitation by making a large IR module from smaller modules before generating code and thus we can have register usage information (at lest) for procedure which was previously defined in other module, because now with LTO every thing is in one module. So that also clarifies that IPRA does not do anything at link time.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Now coming to LLC, it can use IPRA and optimize for functions defined in current module. So yes while compiling whole program ( a single huge .bc file) IPRA can be used with LLC. Also just note that if a software is written in separate files per module (which is very common) and still you want to maximize benefits of IPRA, then we can use llvm-link tool to combine several .bc files to produce a huge .bc file and use that with LLC to get maximum benefits. <o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div></blockquote><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">I know this is not the typical “linux” scenario (dynamic linking of not only standard libraries,</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">but also sometimes even application libraries, and lots of static linking because of program</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">size), but it is a typical “embedded” scenario, which is where I am currently.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I don't understand this use case but we can have further improvement in IPRA for example if you have several libraries which has already compiled and codegen, but you are able to provide information of register usage for the functions of that libraries than we can think about an approach were we can store register usage information into a file (which will obviously increase compile time) and use that information across different compiler instances so that we can provide register usage information with out having actual code while compiling.<o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">Other thoughts or comments ?</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I am looking for some ideas that can improve current IPRA. So if you feel anything relevant please let me know we can discuss and implement feasible ideas.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Thanks,<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Vivek <o:p class=""></o:p></div></div></div><div class=""><div class=""><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class="">--Peter Lawrence.</span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125);" class=""> </span><o:p class=""></o:p></div></div><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><b class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">From:</span></b><span class="apple-converted-space"><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class=""> </span></span><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">vivek pandya [mailto:</span><a href="mailto:vivekvpandya@gmail.com" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">vivekvpandya@gmail.com</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">]<span class="apple-converted-space"> </span><br class=""><b class="">Sent:</b><span class="apple-converted-space"> </span>Wednesday, July 06, 2016 2:09 PM<br class=""><b class="">To:</b><span class="apple-converted-space"> </span>llvm-dev <</span><a href="mailto:llvm-dev@lists.llvm.org" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">llvm-dev@lists.llvm.org</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">>;<span class="apple-converted-space"> </span></span><a href="mailto:llvm-dev-request@lists.llvm.org" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">llvm-dev-request@lists.llvm.org</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">; Lawrence, Peter <</span><a href="mailto:c_plawre@qca.qualcomm.com" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: purple;" class="">c_plawre@qca.qualcomm.com</span></a><span style="font-size: 11pt; font-family: Calibri, sans-serif;" class="">><br class=""><b class="">Subject:</b><span class="apple-converted-space"> </span>Re:[llvm-dev] IPRA, interprocedural register allocation, question</span><o:p class=""></o:p></div></div><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Hello Peter,<o:p class=""></o:p></div></div><div class=""><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Thanks to pointing out this interesting case. <o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><p class="MsoNormal" style="margin: 0in 0in 12pt; font-size: 12pt; font-family: 'Times New Roman', serif;">Vivek,<br class=""> I have an application where many of the leaf functions are<br class="">Hand-coded assembly language, because they use special IO instructions<br class="">That only the assembler knows about. These functions typically don't<br class="">Use any registers besides the incoming argument registers, IE they don't<br class="">Need to use any additional callee-save nor caller-save registers.<o:p class=""></o:p></p></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">If inline asm template has specified clobbered list properly than IPRA is able to use that information and it propagates correct register mask (and that also means that skipping clobbers list while IPRA enabled may broke executable)<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">For example in following code:<o:p class=""></o:p></div></div></div><div class=""><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">int gcd( int a, int b ) {<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> int result ;<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> /* Compute Greatest Common Divisor using Euclid's Algorithm */<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> __asm__ __volatile__ ( "movl %1, %%r15d;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "movl %2, %%ecx;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "CONTD: cmpl $0, %%ecx;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "je DONE;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "xorl %%r13d, %%r13d;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "idivl %%ecx;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "movl %%ecx, %%r15d;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "movl %%r13d, %%ecx;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "jmp CONTD;"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> "DONE: movl %%r15d, %0;" : "=g" (result) : "g" (a), "g" (b) : "ecx" ,"r13", "r15"<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> );<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> return result ;<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">}<o:p class=""></o:p></div></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">IPRA calculates and propagates correct regmask in which it marks CH, CL, ECX .. clobbered and R13, R15 is not marked clobbered as it is callee saved and LLVM code generators also insert spill/restores code for them.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Is there any way in your IPRA interprocedural register allocation project that<br class="">The user can supply this information for external functions ?<o:p class=""></o:p></div></div></blockquote><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">By external word do you here mean function defined in other module than being used? In that case as IPRA can operate on only one module at time register usage propagation is not possible. But there is a work around for this problem. You can use IPRA with link time optimization enabled because the way LLVM LTO works it creates a big IR modules out of source files and them optimize and codegen it so in that case IPRA can have actual register usage info (if function will be compiled in current module). <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">In case you want to experiment with IPRA please apply <a href="http://reviews.llvm.org/D21395" target="_blank" style="color: purple; text-decoration: underline;" class=""><span style="color: purple;" class="">http://reviews.llvm.org/D21395</span></a><span class="apple-converted-space"> </span>this patch before you begin.<o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></div><div class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">-Vivek<o:p class=""></o:p></div></div></div><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""> <o:p class=""></o:p></div></div></blockquote><blockquote style="border-style: none none none solid; border-left-color: rgb(204, 204, 204); border-left-width: 1pt; padding: 0in 0in 0in 6pt; margin: 5pt 0in 5pt 4.8pt;" class=""><div class=""><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">Perhaps using some form of __attribute__ ?<br class="">Maybe __attribute__ ((registermask = ....)) ?<br class=""><br class=""><br class="">--Peter Lawrence.</div></div></blockquote></div></div></div></div></div></div></div></blockquote></div></div></div></div></div></blockquote></div></div></div></div></div></div></div></blockquote></div></div></div></blockquote></div></blockquote></div></div></blockquote></div><br class=""></body></html>