<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <br>
    <br>
    <div class="moz-cite-prefix">On 8/6/19 3:02 AM, Hiroshi Yamauchi via
      llvm-dev wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAASM7N+gvmKZuTzx0LZG6h8z8y_ruUZ3ZqHAb0DVz=nF86+2iw@mail.gmail.com">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div dir="ltr">
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif">I had a chance
          to try -print-after-all with NPM.</div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif"><br>
        </div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif">It seems
          like there's still no output for the passes
          before objc-arc-contract (which is basically what I saw
          before.) Does anyone else see this?</div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif"><br>
        </div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif">Are we talking
          about the same thing?</div>
      </div>
    </blockquote>
    Apparently both yes and no.<br>
    How do you run this?<br>
    <br>
    IR printing is implemented through pass instrumentation and is
    enabled in e.g. opt through the use of StandardInstrumentation<br>
    in llvm::runPassPipeline.<br>
    If you set up NewPM by yourself w/o the use llvm::runPassPipeline
    then most likely you just do not have StandardInstrumentation
    installed.<br>
    <br>
    regards,<br>
      Fedor.
    <blockquote type="cite"
cite="mid:CAASM7N+gvmKZuTzx0LZG6h8z8y_ruUZ3ZqHAb0DVz=nF86+2iw@mail.gmail.com">
      <div dir="ltr">
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif"><br>
        </div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif">*** IR Dump
          After ObjC ARC contraction ***<br>
          *** IR Dump After Pre-ISel Intrinsic Lowering ***<br>
          *** IR Dump After Expand Atomic instructions ***<br>
          *** IR Dump After Canonicalize natural loops ***<br>
          *** IR Dump After Loop Strength Reduction ***<br>
          *** IR Dump After Merge contiguous icmps into a memcmp ***<br>
          *** IR Dump After Expand memcmp() to load/stores ***<br>
          *** IR Dump After Lower Garbage Collection Instructions ***<br>
          *** IR Dump After Shadow Stack GC Lowering ***<br>
          *** IR Dump After Remove unreachable blocks from the CFG ***<br>
          *** IR Dump After Constant Hoisting ***<br>
          *** IR Dump After Partially inline calls to library functions
          ***<br>
          *** IR Dump After Instrument function entry/exit with calls to
          e.g. mcount() (post inlining) ***<br>
          *** IR Dump After Scalarize Masked Memory Intrinsics ***<br>
          *** IR Dump After Expand reduction intrinsics ***<br>
          *** IR Dump After Interleaved Access Pass ***<br>
          *** IR Dump After Expand indirectbr instructions ***<br>
          *** IR Dump After CodeGen Prepare ***<br>
        </div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif">*** IR Dump
          After Rewrite Symbols ***<br>
          *** IR Dump After Exception handling preparation ***<br>
          *** IR Dump After Safe Stack instrumentation pass ***<br>
        </div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif">.....</div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif">(dump from the
          machine passes)</div>
        <div class="gmail_default"
          style="font-family:arial,helvetica,sans-serif"><br>
        </div>
      </div>
      <br>
      <div class="gmail_quote">
        <div dir="ltr" class="gmail_attr">On Sun, Jul 21, 2019 at 7:37
          AM Fedor Sergeev <<a
            href="mailto:fedor.v.sergeev@gmail.com"
            moz-do-not-send="true">fedor.v.sergeev@gmail.com</a>>
          wrote:<br>
        </div>
        <blockquote class="gmail_quote" style="margin:0px 0px 0px
          0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
          <div dir="auto">FWIW, print-*-all should work under NPM just
            fine and the only problem with print-* is (absent) uniform
            pass name processing for cl::opt.
            <div dir="auto">It is easy to introduce yet another option
              that takes NPM pass names (and that's what we actually did
              downstream).</div>
            <div dir="auto">Any suggestions on how to resolve this
              nuisance are welcome.</div>
            <div dir="auto"><br>
            </div>
            <div dir="auto">regards, </div>
            <div dir="auto">  Fedor. </div>
            <div dir="auto"><br>
            </div>
          </div>
          <br>
          <div class="gmail_quote">
            <div dir="ltr" class="gmail_attr">чт, 11 июл. 2019 г., 18:53
              Hiroshi Yamauchi via llvm-dev <<a
                href="mailto:llvm-dev@lists.llvm.org" target="_blank"
                moz-do-not-send="true">llvm-dev@lists.llvm.org</a>>:<br>
            </div>
            <blockquote class="gmail_quote" style="margin:0px 0px 0px
              0.8ex;border-left:1px solid
              rgb(204,204,204);padding-left:1ex">I don't exactly
              remember when I last tried it and I didn't realize<br>
              there was r342896. I'll check it out. Thanks.<br>
              <br>
              On Wed, Jul 10, 2019 at 1:14 PM Philip Pfaffe <<a
                href="mailto:philip.pfaffe@gmail.com" rel="noreferrer"
                target="_blank" moz-do-not-send="true">philip.pfaffe@gmail.com</a>>
              wrote:<br>
              ><br>
              > Printing was implemented in r342896.<br>
              > @Hiroshi: Are there specific issues or limitations
              you encountered with it?<br>
              ><br>
              > Cheers,<br>
              > Philip<br>
              ><br>
              > On Wed, Jul 10, 2019 at 8:48 PM Troy Johnson via
              llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org"
                rel="noreferrer" target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a>>
              wrote:<br>
              >><br>
              >> -print-after-all is very useful for debugging and
              learning about LLVM.  I would hope that would be
              implemented for the new PM before removing the old PM. 
              I'd personally consider it a blocker.<br>
              >><br>
              >> -Troy<br>
              >><br>
              >> > -----Original Message-----<br>
              >> > From: llvm-dev <<a
                href="mailto:llvm-dev-bounces@lists.llvm.org"
                rel="noreferrer" target="_blank" moz-do-not-send="true">llvm-dev-bounces@lists.llvm.org</a>>
              On Behalf Of Eric Christopher<br>
              >> > via llvm-dev<br>
              >> > Sent: Tuesday, July 09, 2019 7:40 PM<br>
              >> > To: Hiroshi Yamauchi <<a
                href="mailto:yamauchi@google.com" rel="noreferrer"
                target="_blank" moz-do-not-send="true">yamauchi@google.com</a>><br>
              >> > Cc: llvm-dev <<a
                href="mailto:llvm-dev@lists.llvm.org" rel="noreferrer"
                target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a>>;
              Yi Kong <<a href="mailto:yikong@google.com"
                rel="noreferrer" target="_blank" moz-do-not-send="true">yikong@google.com</a>><br>
              >> > Subject: Re: [llvm-dev] Status of the New
              Pass Manager<br>
              >> ><br>
              >> > They don't, but this isn't considered a
              blocker to removing the old one as far as I<br>
              >> > know.<br>
              >> ><br>
              >> > -eric<br>
              >> ><br>
              >> > On Tue, Jul 9, 2019 at 11:09 AM Hiroshi
              Yamauchi via llvm-dev <llvm-<br>
              >> > <a href="mailto:dev@lists.llvm.org"
                rel="noreferrer" target="_blank" moz-do-not-send="true">dev@lists.llvm.org</a>>
              wrote:<br>
              >> > ><br>
              >> > > FWIW, the flags like -print-after,
              -printer-after-all don't work well<br>
              >> > > with the new pass manager last time I
              checked.<br>
              >> > ><br>
              >> > > On Mon, Jul 8, 2019 at 12:20 PM Stephen
              Hines via llvm-dev<br>
              >> > > <<a
                href="mailto:llvm-dev@lists.llvm.org" rel="noreferrer"
                target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a>>
              wrote:<br>
              >> > > ><br>
              >> > > > The Android platform build (AOSP)
              has also switched to the new pass<br>
              >> > manager recently. We do have a few bugs that
              we are chasing (hence opt-outs),<br>
              >> > but it is working quite well otherwise.<br>
              >> > > ><br>
              >> > > > Our current list of issues:<br>
              >> > > > 1) Libsqlite still has a
              mysterious failure that we haven't been able to reduce<br>
              >> > well.<br>
              >> > > > 2) <a
                href="https://bugs.llvm.org/show_bug.cgi?id=42124"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://bugs.llvm.org/show_bug.cgi?id=42124</a>
              shows that inlining costs<br>
              >> > are a bit different under NPM. <a
                href="https://reviews.llvm.org/D63034" rel="noreferrer
                noreferrer" target="_blank" moz-do-not-send="true">https://reviews.llvm.org/D63034</a>
              is one proposed<br>
              >> > patch for addressing this.<br>
              >> > > > 3) libpdfium exposed a
              non-determinism issue with NPM where having the<br>
              >> > linux-libc-dev system package installed
              changes execution. We are still looking<br>
              >> > at why this happens.<br>
              >> > > > 4) Sanitizer coverage information
              isn't supported by the NPM yet<br>
              >> > (<a href="https://reviews.llvm.org/D62888"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://reviews.llvm.org/D62888</a>).<br>
              >> > > ><br>
              >> > > > Thanks,<br>
              >> > > > Steve<br>
              >> > > ><br>
              >> > > > On Mon, Jul 1, 2019 at 11:07 AM
              Alex Bradbury via llvm-dev <llvm-<br>
              >> > <a href="mailto:dev@lists.llvm.org"
                rel="noreferrer" target="_blank" moz-do-not-send="true">dev@lists.llvm.org</a>>
              wrote:<br>
              >> > > >><br>
              >> > > >> On Thu, 27 Jun 2019 at 17:46,
              Philip Reames via llvm-dev<br>
              >> > > >> <<a
                href="mailto:llvm-dev@lists.llvm.org" rel="noreferrer"
                target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a>>
              wrote:<br>
              >> > > >> ><br>
              >> > > >> > For our downstream usage,
              we've switched entirely to the new pass<br>
              >> > manager.  We made the switch a couple of
              months ago.  All of our testing is<br>
              >> > being done with the NPM, and we're about to
              start deleting (downstream) code<br>
              >> > which was only needed by the legacy pass
              manager.<br>
              >> > > >> ><br>
              >> > > >> > I believe several other
              major contributors are in the same state.  We<br>
              >> > really need to get upstream switched over so
              that all of the community's testing<br>
              >> > efforts are aligned again.<br>
              >> > > >><br>
              >> > > >> I hadn't realised it was so
              close to being ready. Do you see this<br>
              >> > > >> as a switch that could be made
              before 9.0, or after it?<br>
              >> > > >><br>
              >> > > >> Best,<br>
              >> > > >><br>
              >> > > >> Alex<br>
              >> > > >>
              _______________________________________________<br>
              >> > > >> LLVM Developers mailing list<br>
              >> > > >> <a
                href="mailto:llvm-dev@lists.llvm.org" rel="noreferrer"
                target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a><br>
              >> > > >> <a
                href="https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
              >> > > ><br>
              >> > > >
              _______________________________________________<br>
              >> > > > LLVM Developers mailing list<br>
              >> > > > <a
                href="mailto:llvm-dev@lists.llvm.org" rel="noreferrer"
                target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a><br>
              >> > > > <a
                href="https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
              >> > >
              _______________________________________________<br>
              >> > > LLVM Developers mailing list<br>
              >> > > <a
                href="mailto:llvm-dev@lists.llvm.org" rel="noreferrer"
                target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a><br>
              >> > > <a
                href="https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
              >> >
              _______________________________________________<br>
              >> > LLVM Developers mailing list<br>
              >> > <a href="mailto:llvm-dev@lists.llvm.org"
                rel="noreferrer" target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a><br>
              >> > <a
                href="https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
              >> _______________________________________________<br>
              >> LLVM Developers mailing list<br>
              >> <a href="mailto:llvm-dev@lists.llvm.org"
                rel="noreferrer" target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a><br>
              >> <a
                href="https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
              _______________________________________________<br>
              LLVM Developers mailing list<br>
              <a href="mailto:llvm-dev@lists.llvm.org" rel="noreferrer"
                target="_blank" moz-do-not-send="true">llvm-dev@lists.llvm.org</a><br>
              <a
                href="https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"
                rel="noreferrer noreferrer" target="_blank"
                moz-do-not-send="true">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
            </blockquote>
          </div>
        </blockquote>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <pre class="moz-quote-pre" wrap="">_______________________________________________
LLVM Developers mailing list
<a class="moz-txt-link-abbreviated" href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>
<a class="moz-txt-link-freetext" href="https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a>
</pre>
    </blockquote>
    <br>
  </body>
</html>