<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<br>
<br>
<div class="moz-cite-prefix">On 01/04/2016 09:55 AM, Amaury SECHET
wrote:<br>
</div>
<blockquote
cite="mid:CANGV3T2CLd_cocdkhw5DgUPpO1qXuVc1teAqy66sPdzMvNXe7Q@mail.gmail.com"
type="cite">
<div dir="ltr"><br>
<div class="gmail_extra"><br>
<div class="gmail_quote">2016-01-04 18:21 GMT+01:00 Philip
Reames <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:listmail@philipreames.com" target="_blank">listmail@philipreames.com</a>></span>:<br>
<blockquote class="gmail_quote" style="margin:0px 0px 0px
0.8ex;border-left:1px solid
rgb(204,204,204);padding-left:1ex">
<div text="#000000" bgcolor="#FFFFFF"><span class="">
<div>On 01/04/2016 07:32 AM, Amaury SECHET wrote:<br>
</div>
<blockquote type="cite">
<div dir="ltr">
<div>
<div>After a bit more investigation, it turns
out that because %0 is stored into %1 (after
bitcast) and so %3 may have access to it and
clobber it.<br>
</div>
</div>
</div>
</blockquote>
</span> Can you give a bit more context? I'm not sure
which of the examples you're talking about.<br>
<br>
</div>
</blockquote>
<div><br>
</div>
<div>Sure. Let's look at <a moz-do-not-send="true"
href="http://pastebin.com/K0J9yGq1">http://pastebin.com/K0J9yGq1</a><br>
<br>
</div>
<div>Because of the store line 7, it is assumed that the
call line 8 may see %0 and even modify the memory it
points to. As a result, it is assumed that the load line
11 may not be eliminated.<br>
<br>
</div>
<div>Which seems actually correct in the general case.<br>
</div>
</div>
</div>
</div>
</blockquote>
This seems like a restatement of what I said in my original
response:<br>
"You have to teach the alias analysis that an unescaped noalias
pointer can't alias the global state allocmemory might access.
Slightly surprised we don't get this today, but oh well. "<br>
<br>
Or am I missing something?<br>
<blockquote
cite="mid:CANGV3T2CLd_cocdkhw5DgUPpO1qXuVc1teAqy66sPdzMvNXe7Q@mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div><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 text="#000000" bgcolor="#FFFFFF"><span class=""><br>
<blockquote type="cite">
<div dir="ltr">
<div>
<div><br>
</div>
After a bit of thought, it is correct in the
general case, but definitively something
stricter is needed here. Looking at <tt><span>inaccessiblememonly</span></tt>
I'm not sure this is what is needed. What if the
memory allocator is defined is the current
module ?<br>
</div>
</div>
</blockquote>
</span> At the moment, inaccessiblememonly would require
separate compilation of the allocation function. <br>
<span class="">
<blockquote type="cite">
<div dir="ltr">
<div><br>
</div>
This leads me to conclude this is way more linked
to the memory allocation pass than I expected it
to be in the first place. Can I ask what you plan
to use <tt><span>inaccessiblememonly</span></tt>
for ? Should the semantic be refined to fit the
bill better ?<br>
</div>
</blockquote>
</span> Well, I didn't introduce the attribute, so I
can't speak for the original intent. For me, I plan on
applying it to some of our out of line allocation
functions and other helper routines which modify runtime
state, but not java visible state. <br>
<br>
If you have specific suggestions for how to refine the
semantics, please make them. Getting the details right
is always the hard part. :)<br>
<br>
You might also consider using a variant of your
allocation function which takes a pointer to the global
state it needs to modify. Doing this would allow you to
use argmemonly to restrict the aliasing while still
allowing whole program optimization. I haven't tried
this in practice, but it seems like it would probably
work...<span class=""><br>
</span></div>
</blockquote>
<div><br>
</div>
<div>I do not wish to make suggestion before I understand
where this is coming from. So far, from what I've
collected, use cases are:<br>
</div>
<div> - Memory allocation<br>
</div>
<div> - Runtime isolation for managed languages.<br>
</div>
<div><br>
</div>
<div>I have some more though to put into this, but to boot,
would that be possible to only use this attribute on
method that are declared, but not defined and remove it
when merging modules ? </div>
</div>
</div>
</div>
</blockquote>
This seems semi reasonable. I haven't thought through the
implications, but it might be worth considering. <br>
<br>
<blockquote
cite="mid:CANGV3T2CLd_cocdkhw5DgUPpO1qXuVc1teAqy66sPdzMvNXe7Q@mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div>It doesn't look like it is necessary to have it when
the function may be exposed depending on the way the
software is built.<br>
</div>
</div>
</div>
</div>
</blockquote>
Er, not sure what you meant here. <br>
<br>
Philip<br>
<br>
</body>
</html>