<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class=""><br class=""><div><br class=""><blockquote type="cite" class=""><div class="">On Feb 2, 2018, at 3:53 PM, Nemanja Ivanovic via llvm-commits <<a href="mailto:llvm-commits@lists.llvm.org" class="">llvm-commits@lists.llvm.org</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div dir="ltr" class=""><div class=""><div class=""><div class=""><div class=""><div class=""><div class=""><div class=""><div class=""><div class=""><div class=""><div class="">Sorry, but it seems to me that we are working around the problem with the hasExtraRegAllocReq flag and this workaround is introducing new problems that then need to be solved.<br class=""></div>Going back to the original problem that Escha reported, it seems to be that the back end ends up in the following state which arguably is probably not a reasonable state to be in (if register renaming is to be allowed):<br class=""><br class=""></div>- There are constraints added to the virtual registers that restrict which register class the assigned physical register must come from<br class=""></div>- The register allocator respects those constraints and selects a register from the correct class<br class=""></div>- The selected physical register is marked as renamable<br class=""></div>- Calling MachineInstr::getRegClassConstraint() on the machine instruction returns a register class that does not take into account the additional constraints<br class=""><br class=""></div>Ultimately, anything that does register renaming with the back end in that state can potentially break the code since it doesn't know what actual class the register rename is allowed to come from.<br class=""></div><br class=""></div>Wouldn't this mean that a target that can be in this state has at least these two reasonable options:<br class=""></div>1. Set some property that will ensure that the renamable flag is never set on any register - essentially what Geoff is suggesting in the latter part of his last email. I would just make it a property of TargetRegisterInfo (perhaps `virtual bool TargetRegisterInfo::allowRegisterRenaming()`).<br class=""></div></div></div></div></blockquote><div><br class=""></div><div>Again, yes I think we need that.</div><div><br class=""></div><div>More generally, I believe we want to step back with the renamable flag, think about its semantic and what in the infrastructure we would need to fix, if any, to make sure that semantic is preserved/preservable along the way.</div><br class=""><blockquote type="cite" class=""><div class=""><div dir="ltr" class=""><div class="">2. We make TargetInstrInfo::getRegClass() virtual and allow the target to override it so that the extra constraints can be taken into account if it wants to allow renaming of registers<br class=""><br class=""></div>Not to suggest that the two solutions are mutually exclusive.<br class=""></div><div class="gmail_extra"><br class=""><div class="gmail_quote">On Fri, Feb 2, 2018 at 5:57 PM, <span dir="ltr" class=""><<a href="mailto:gberry@codeaurora.org" target="_blank" class="">gberry@codeaurora.org</a>></span> wrote:<br class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div link="blue" vlink="purple" lang="EN-US" class=""><div class="m_4139097589071889215WordSection1"><p class="MsoNormal">I think the fix for this the way things are currently setup would be for the code that changes the COPY to the target opcode to clear the renamable bits on the operands.<u class=""></u></p></div></div></blockquote></div></div></div></blockquote><div><br class=""></div><div>Although it would work, I think it hints that we didn’t think the handle of the renamable flag all the way. I feel like we are reproducing the same problems that the kill flag had whereas we are trying to kill the kill flag.</div><div>Let us no repeat history :).</div><br class=""><blockquote type="cite" class=""><div class=""><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div link="blue" vlink="purple" lang="EN-US" class=""><div class="m_4139097589071889215WordSection1"><p class="MsoNormal"><u class=""></u></p><p class="MsoNormal">If we add a target global “all opcodes are hasExtraRegAllocReq” setting, we could avoid setting any COPY operands as renamable in the first place.<u class=""></u><u class=""></u></p><span class=""><p class="MsoNormal"><u class=""></u> <u class=""></u></p><div class=""><p class="MsoNormal">-- <u class=""></u><u class=""></u></p><p class="MsoNormal">Geoff Berry<u class=""></u><u class=""></u></p><p class="MsoNormal">Employee of Qualcomm Datacenter Technologies, Inc.<u class=""></u><u class=""></u></p><p class="MsoNormal"> Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc. Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.<u class=""></u><u class=""></u></p></div><p class="MsoNormal"><u class=""></u> <u class=""></u></p></span><div class=""><div style="border:none;border-top:solid #e1e1e1 1.0pt;padding:3.0pt 0in 0in 0in" class=""><p class="MsoNormal"><span class=""><b class="">From:</b> <a href="mailto:fglaser@apple.com" target="_blank" class="">fglaser@apple.com</a> [mailto:<a href="mailto:fglaser@apple.com" target="_blank" class="">fglaser@apple.com</a>] <b class="">On Behalf Of </b><a href="mailto:escha@apple.com" target="_blank" class="">escha@apple.com</a><br class=""></span><b class="">Sent:</b> Friday, February 2, 2018 5:30 PM<br class=""><b class="">To:</b> <a href="mailto:escha@apple.com" target="_blank" class="">escha@apple.com</a><br class=""><b class="">Cc:</b> <a href="mailto:gberry@codeaurora.org" target="_blank" class="">gberry@codeaurora.org</a>; <a href="mailto:junbuml@codeaurora.org" target="_blank" class="">junbuml@codeaurora.org</a>; <a href="mailto:marina.yatsina@intel.com" target="_blank" class="">marina.yatsina@intel.com</a>; <a href="mailto:kannan.narayanan@amd.com" target="_blank" class="">kannan.narayanan@amd.com</a>; <a href="mailto:nhaehnle@gmail.com" target="_blank" class="">nhaehnle@gmail.com</a>; <a href="mailto:wei.ding2@amd.com" target="_blank" class="">wei.ding2@amd.com</a>; Matthias Braun <<a href="mailto:matze@braunis.de" target="_blank" class="">matze@braunis.de</a>>; Nemanja Ivanovic <<a href="mailto:nemanja.i.ibm@gmail.com" target="_blank" class="">nemanja.i.ibm@gmail.com</a>>; llvm-commits <<a href="mailto:llvm-commits@lists.llvm.org" target="_blank" class="">llvm-commits@lists.llvm.org</a>>; <a href="mailto:reviews%2BD41835%2Bpublic%2B9c1dec7fb6e75ce0@reviews.llvm.org" target="_blank" class="">reviews+D41835+public+<wbr class="">9c1dec7fb6e75ce0@reviews.llvm.<wbr class="">org</a>; <a href="mailto:tpr.llvm@botech.co.uk" target="_blank" class="">tpr.llvm@botech.co.uk</a>; <a href="mailto:javed.absar@arm.com" target="_blank" class="">javed.absar@arm.com</a></p><div class=""><div class="h5"><br class=""><b class="">Subject:</b> Re: [PATCH] D41835: [MachineCopyPropagation] Extend pass to do COPY source forwarding<u class=""></u><u class=""></u></div></div><div class=""><br class="webkit-block-placeholder"></div></div></div><div class=""><div class="h5"><p class="MsoNormal"><u class=""></u> <u class=""></u></p><p class="MsoNormal">update: it gets worse. i can’t get the tests to pass because a lot of tests fail with this when the verifier is on:<u class=""></u><u class=""></u></p><div class=""><pre style="box-sizing:border-box;white-space:pre-wrap;word-wrap:break-word;font-variant-ligatures:normal" class=""><span style="color:#333333" class=""><u class=""></u> <u class=""></u></span></pre><pre class=""><span style="color:#333333" class="">*** Bad machine code: Illegal isRenamable setting for opcode with extra regalloc requirements ***<u class=""></u><u class=""></u></span></pre><div class=""><p class="MsoNormal"><u class=""></u> <u class=""></u></p></div></div><div class=""><p class="MsoNormal">This occurs on our copy instructions after we expand them into actual machine opcodes, because the machine opcode for the copy instruction isn’t Renamable (none of our instructions are).<u class=""></u><u class=""></u></p></div><div class=""><p class="MsoNormal"><u class=""></u> <u class=""></u></p></div><div class=""><p class="MsoNormal">—escha<u class=""></u><u class=""></u></p></div><div class=""><p class="MsoNormal"><u class=""></u> <u class=""></u></p><div class=""><p class="MsoNormal"><br class=""><br class=""><u class=""></u><u class=""></u></p><blockquote style="margin-top:5.0pt;margin-bottom:5.0pt" class=""><div class=""><p class="MsoNormal">On Feb 2, 2018, at 1:37 PM, via llvm-commits <<a href="mailto:llvm-commits@lists.llvm.org" target="_blank" class="">llvm-commits@lists.llvm.org</a>> wrote:<u class=""></u><u class=""></u></p></div><p class="MsoNormal"><u class=""></u> <u class=""></u></p><div class=""><div class=""><p class="MsoNormal">this is probably a reasonable solution. mostly, at this point, i’m worried about the fact that upstream now apparently has a different understanding of register classes than we do (in terms of what is “safe”), which could lead to other future issues for us.<u class=""></u><u class=""></u></p><div class=""><p class="MsoNormal"><u class=""></u> <u class=""></u></p></div><div class=""><p class="MsoNormal">—escha<u class=""></u><u class=""></u></p><div class=""><p class="MsoNormal"><br class=""><br class=""><u class=""></u><u class=""></u></p><blockquote style="margin-top:5.0pt;margin-bottom:5.0pt" class=""><div class=""><p class="MsoNormal">On Feb 2, 2018, at 1:36 PM, <a href="mailto:gberry@codeaurora.org" target="_blank" class="">gberry@codeaurora.org</a> wrote:<u class=""></u><u class=""></u></p></div><p class="MsoNormal"><u class=""></u> <u class=""></u></p><div class=""><div class=""><p class="MsoNormal">Ah, okay. So I assume you’re okay with disabling the post-RA run of this pass for this particular target to fix this particular issue?<u class=""></u><u class=""></u></p></div><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div><div class=""><div class=""><p class="MsoNormal">--<span class="m_4139097589071889215apple-converted-space"> </span><u class=""></u><u class=""></u></p></div><div class=""><p class="MsoNormal">Geoff Berry<u class=""></u><u class=""></u></p></div><div class=""><p class="MsoNormal">Employee of Qualcomm Datacenter Technologies, Inc.<u class=""></u><u class=""></u></p></div><div class=""><p class="MsoNormal">Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc. Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.<u class=""></u><u class=""></u></p></div></div><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div><div class=""><div style="border:none;border-top:solid #e1e1e1 1.0pt;padding:3.0pt 0in 0in 0in" class=""><div class=""><p class="MsoNormal"><b class="">From:</b><span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:fglaser@apple.com" target="_blank" class="">fglaser@apple.com</a> [<a href="mailto:fglaser@apple.com" target="_blank" class="">mailto:fglaser@apple.com</a>]<span class="m_4139097589071889215apple-converted-space"> </span><b class="">On Behalf Of<span class="m_4139097589071889215apple-converted-space"> </span></b><a href="mailto:escha@apple.com" target="_blank" class="">escha@apple.com</a><br class=""><b class="">Sent:</b><span class="m_4139097589071889215apple-converted-space"> </span>Friday, February 2, 2018 4:32 PM<br class=""><b class="">To:</b><span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:gberry@codeaurora.org" target="_blank" class="">gberry@codeaurora.org</a><br class=""><b class="">Cc:</b><span class="m_4139097589071889215apple-converted-space"> </span>Quentin Colombet <<a href="mailto:qcolombet@apple.com" target="_blank" class="">qcolombet@apple.com</a>>; <a href="mailto:reviews+D41835+public+9c1dec7fb6e75ce0@reviews.llvm.org" target="_blank" class="">reviews+D41835+public+<wbr class="">9c1dec7fb6e75ce0@reviews.llvm.<wbr class="">org</a>; Geoff Berry via Phabricator <<a href="mailto:reviews@reviews.llvm.org" target="_blank" class="">reviews@reviews.llvm.org</a>>; <a href="mailto:javed.absar@arm.com" target="_blank" class="">javed.absar@arm.com</a>; Matthias Braun <<a href="mailto:matze@braunis.de" target="_blank" class="">matze@braunis.de</a>>; Jonas Paulsson <<a href="mailto:paulsson@linux.vnet.ibm.com" target="_blank" class="">paulsson@linux.vnet.ibm.com</a>>; <a href="mailto:tstellar@redhat.com" target="_blank" class="">tstellar@redhat.com</a>; Matt Arsenault <<a href="mailto:Matthew.Arsenault@amd.com" target="_blank" class="">Matthew.Arsenault@amd.com</a>>; <a href="mailto:junbuml@codeaurora.org" target="_blank" class="">junbuml@codeaurora.org</a>; <a href="mailto:marina.yatsina@intel.com" target="_blank" class="">marina.yatsina@intel.com</a>; <a href="mailto:wei.ding2@amd.com" target="_blank" class="">wei.ding2@amd.com</a>; <a href="mailto:kannan.narayanan@amd.com" target="_blank" class="">kannan.narayanan@amd.com</a>; <a href="mailto:nhaehnle@gmail.com" target="_blank" class="">nhaehnle@gmail.com</a>; Nemanja Ivanovic <<a href="mailto:nemanja.i.ibm@gmail.com" target="_blank" class="">nemanja.i.ibm@gmail.com</a>>; llvm-commits <<a href="mailto:llvm-commits@lists.llvm.org" target="_blank" class="">llvm-commits@lists.llvm.org</a>>; <a href="mailto:tpr.llvm@botech.co.uk" target="_blank" class="">tpr.llvm@botech.co.uk</a><br class=""><b class="">Subject:</b><span class="m_4139097589071889215apple-converted-space"> </span>Re: [PATCH] D41835: [MachineCopyPropagation] Extend pass to do COPY source forwarding<u class=""></u><u class=""></u></p></div></div></div><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div><div class=""><p class="MsoNormal">looking at it again, it’s making more sense: your patch enabled MachineCopyPropagation in the RA pipeline. the pass might have *already* been broken, but it wasn’t running for us, and now it is, so it broke things in that way too.<u class=""></u><u class=""></u></p></div><div class=""><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal">—escha<u class=""></u><u class=""></u></p></div><div class=""><div class=""><p class="MsoNormal"><br class=""><br class=""><br class=""><u class=""></u><u class=""></u></p></div><blockquote style="margin-top:5.0pt;margin-bottom:5.0pt" class=""><div class=""><div class=""><p class="MsoNormal">On Feb 2, 2018, at 1:26 PM,<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:gberry@codeaurora.org" target="_blank" class=""><span style="color:purple" class="">gberry@codeaurora.org</span></a><span class="m_4139097589071889215apple-converted-space"> </span>wrot<wbr class="">e:<u class=""></u><u class=""></u></p></div></div><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div><div class=""><div class=""><div class=""><p class="MsoNormal">escha,<u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal">I’m confused by your comment about bundles. It was known that this patch doesn’t handle forwarding into bundles, but it sounds like to me you are seeing problems with the dead COPY removal part of this pass, which was not intentionally changed by this patch (other than adding a DEBUG statement when it happens).<u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">--<span class="m_4139097589071889215apple-converted-space"> </span><u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal">Geoff Berry<u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal">Employee of Qualcomm Datacenter Technologies, Inc.<u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal">Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc. Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div style="border:none;border-top:solid #e1e1e1 1.0pt;padding:3.0pt 0in 0in 0in" class=""><div class=""><div class=""><p class="MsoNormal"><b class="">From:</b><span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:fglaser@apple.com" target="_blank" class=""><span style="color:purple" class="">fglaser@apple.com</span></a><span class="m_4139097589071889215apple-converted-space"> </span>[<a href="mailto:fglaser@apple.com" target="_blank" class=""><span style="color:purple" class="">mailt<wbr class="">o:fglaser@apple.com</span></a>]<span class="m_4139097589071889215apple-converted-space"> </span><b class="">On Behalf Of<span class="m_4139097589071889215apple-converted-space"> </span></b><a href="mailto:escha@apple.com" target="_blank" class=""><span style="color:purple" class="">escha@apple.com</span></a><br class=""><b class="">Sent:</b><span class="m_4139097589071889215apple-converted-space"> </span>Friday, February 2, 2018 4:10 PM<br class=""><b class="">To:</b><span class="m_4139097589071889215apple-converted-space"> </span>Quentin Colombet <<a href="mailto:qcolombet@apple.com" target="_blank" class=""><span style="color:purple" class="">qcolombet@apple.com</span></a>><br class=""><b class="">Cc:</b><span class="m_4139097589071889215apple-converted-space"> </span>Geoff Berry <<a href="mailto:gberry@codeaurora.org" target="_blank" class=""><span style="color:purple" class="">gberry@codeaurora.org</span></a>>;<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:reviews+D41835+public+9c1dec7fb6e75ce0@reviews.llvm.org" target="_blank" class=""><span style="color:purple" class="">revie<wbr class="">ws+D41835+public+<wbr class="">9c1dec7fb6e75ce0@reviews.llvm.<wbr class="">org</span></a>; Geoff Berry via Phabricator <<a href="mailto:reviews@reviews.llvm.org" target="_blank" class=""><span style="color:purple" class="">reviews@reviews.llvm.org</span></a>>;<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:javed.absar@arm.com" target="_blank" class=""><span style="color:purple" class="">ja<wbr class="">ved.absar@arm.com</span></a>; Matthias Braun <<a href="mailto:matze@braunis.de" target="_blank" class=""><span style="color:purple" class="">matze@braunis.de</span></a>>; Jonas Paulsson <<a href="mailto:paulsson@linux.vnet.ibm.com" target="_blank" class=""><span style="color:purple" class="">paulsson@linux.vnet.ibm.com</span></a>>;<span class="m_4139097589071889215apple-converted-space"><wbr class=""> </span><a href="mailto:tstellar@redhat.com" target="_blank" class=""><span style="color:purple" class="">tstellar@redhat.com</span></a>; Matt Arsenault <<a href="mailto:Matthew.Arsenault@amd.com" target="_blank" class=""><span style="color:purple" class="">Matthew.Arsenault@amd.com</span></a>>;<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:junbuml@codeaurora.org" target="_blank" class=""><span style="color:purple" class="">j<wbr class="">unbuml@codeaurora.org</span></a>;<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:marina.yatsina@intel.com" target="_blank" class=""><span style="color:purple" class="">marina.<wbr class="">yatsina@intel.com</span></a>;<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:wei.ding2@amd.com" target="_blank" class=""><span style="color:purple" class="">wei.ding2@<wbr class="">amd.com</span></a>;<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:kannan.narayanan@amd.com" target="_blank" class=""><span style="color:purple" class="">kannan.narayanan@amd.<wbr class="">com</span></a>;<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:nhaehnle@gmail.com" target="_blank" class=""><span style="color:purple" class="">nhaehnle@gmail.com</span></a>; Nemanja Ivanovic <<a href="mailto:nemanja.i.ibm@gmail.com" target="_blank" class=""><span style="color:purple" class="">nemanja.i.ibm@gmail.com</span></a>>; llvm-commits <<a href="mailto:llvm-commits@lists.llvm.org" target="_blank" class=""><span style="color:purple" class="">llvm-commits@lists.llvm.org</span></a>>;<span class="m_4139097589071889215apple-converted-space"><wbr class=""> </span><a href="mailto:tpr.llvm@botech.co.uk" target="_blank" class=""><span style="color:purple" class="">tpr.llvm@botech.co.uk</span></a><br class=""><b class="">Subject:</b><span class="m_4139097589071889215apple-converted-space"> </span>Re: [PATCH] D41835: [MachineCopyPropagation] Extend pass to do COPY source forwarding<u class=""></u><u class=""></u></p></div></div></div></div><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><div class=""><p class="MsoNormal"><br class=""><br class=""><br class=""><br class=""><u class=""></u><u class=""></u></p></div></div><blockquote style="margin-top:5.0pt;margin-bottom:5.0pt" class=""><div class=""><div class=""><div class=""><p class="MsoNormal">On Feb 2, 2018, at 12:55 PM, Quentin Colombet <<a href="mailto:qcolombet@apple.com" target="_blank" class=""><span style="color:purple" class="">qcolombet@apple.com</span></a>> wrote:<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><div class=""><p class="MsoNormal"><br class=""><br class=""><br class=""><br class=""><u class=""></u><u class=""></u></p></div></div><blockquote style="margin-top:5.0pt;margin-bottom:5.0pt" class=""><div class=""><div class=""><div class=""><p class="MsoNormal">On Feb 2, 2018, at 12:21 PM,<span class="m_4139097589071889215apple-converted-space"> </span><a href="mailto:escha@apple.com" target="_blank" class=""><span style="color:purple" class="">escha@apple.com</span></a><span class="m_4139097589071889215apple-converted-space"> </span>wrote:<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><div class=""><div class=""><p class="MsoNormal">I mean, in that case we are likely to have to mark every single opcode (all 12,000 or so) with this requirement. At that point we might as well just opt out of the pass, I think? At least, it feels like a gross hack that papers over the fact that LLVM has changed how register classes work such that our entire approach is no longer valid.<u class=""></u><u class=""></u></p></div></div><div class=""><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">Also, it seems very weird to make this constraint-violating behavior *opt-out*. Maybe it should be opt-in, i.e. put doesNotHaveExtraSrcRegAllocReq on all instructions it’s okay for?<u class=""></u><u class=""></u></p></div></div></div></div></div></blockquote><div class=""><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">That sounds like a better approach to me.<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">After talking with escha, I agree that TableGen is not necessarily expressive enough to model all the constraints that need to be met and I would err on the safe side of being opt-in instead of opt-out.<u class=""></u><u class=""></u></p></div></div></div></div></div></div></blockquote><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">this is actually worse than i initially realized. one of our targets is VLIW and is completely broken by this patch, *even if we opt out as described*, because it doesn’t iterate over the operands of a bundle. so for example:<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">r3 = COPY foo<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">BUNDLE_TYPE_FOO<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal"><thing that uses r3><u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">END<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">it believes that r3 is dead because it doesn’t iterate over the bundle operands, so it never sees the use of r3.<u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal"> <u class=""></u><u class=""></u></p></div></div></div><div class=""><div class=""><div class=""><p class="MsoNormal">—escha<u class=""></u><u class=""></u></p></div></div></div></div></blockquote></div></div></div></div></blockquote></div><p class="MsoNormal"><u class=""></u> <u class=""></u></p></div></div><p class="MsoNormal">______________________________<wbr class="">_________________<br class="">llvm-commits mailing list<br class=""><a href="mailto:llvm-commits@lists.llvm.org" target="_blank" class="">llvm-commits@lists.llvm.org</a><br class=""><a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits" target="_blank" class="">http://lists.llvm.org/cgi-bin/<wbr class="">mailman/listinfo/llvm-commits</a><u class=""></u><u class=""></u></p></div></blockquote></div><p class="MsoNormal"><u class=""></u> <u class=""></u></p></div></div></div></div></div></blockquote></div><br class=""></div>
_______________________________________________<br class="">llvm-commits mailing list<br class=""><a href="mailto:llvm-commits@lists.llvm.org" class="">llvm-commits@lists.llvm.org</a><br class="">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits<br class=""></div></blockquote></div><br class=""></body></html>