<div dir="ltr"><div><div><div>>What if you had bitcode versions of your C library<br></div>Then the definitions for these functions would be available. Would we still set function attributes to these functions in FunctionAttrs.cpp if their definitions were available?<br><br>>This also seems a bit tailored to malloc/free, and can't work for user-defined allocation functions<br></div>I don't think so. For example, printf would have the flag set, preventing two calls to printfs (without the return value being used) from being interchanged. In the case of user-defined allocation functions, the definitions for those functions are available, and what state they modify is directly visible. I'm not sure I understand how malloc alone would be more powerful. This point however reminds me to add, functions that transitively call functions with HasInaccessibleState must also have the flag set.<br></div></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature"><div dir="ltr"> - Vaivaswatha<br></div></div></div>
<br><div class="gmail_quote">On Fri, Dec 4, 2015 at 10:06 PM, James Molloy <span dir="ltr"><<a href="mailto:james@jamesmolloy.co.uk" target="_blank">james@jamesmolloy.co.uk</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Hi,<div><br></div><div>I'm still a bit dubious about this, I don't think it's bombproof. How does this fit with LTO? What if you had bitcode versions of your C library (not entirely crazy - it might allow a lot of LTO) - you'd collapse those two domains into one in a rather messy way.</div><div><br></div><div>This also seems a bit tailored to malloc/free, and can't work for user-defined allocation functions. Our current attributes mechanism has the ability to infer noalias on such functions, so here you'll be making malloc more powerful than user-defined functions.</div><div><br></div><div>All in all it just smells a bit specialist. I'd welcome it if we could bombproof the semantics and extend the scope somewhat somehow.</div><div><br></div><div>Cheers,</div><div><br></div><div>James</div></div><br><div class="gmail_quote"><div><div class="h5"><div dir="ltr">On Fri, 4 Dec 2015 at 16:31 Vaivaswatha Nagaraj via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>> wrote:<br></div></div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5"><div dir="ltr"><div><div>>there would be two disjoint global states<br></div></div></div><div dir="ltr"><div>In some sense yes, but technically not disjoint. Functions marked with this attribute should still be able to access the globals within the program under compilation, if its not marked with ReadNone.</div></div><div dir="ltr"><div><br><br>>If malloc and free can both use global variables (there is no notion of library in the compiler)<br></div></div><div dir="ltr"><div></div>Inaccessible state here refers to any global that is not visible to the program under compilation. The key idea (behind the new attribute) is to convey that these external functions do things inside that the compiler cannot know about, and hence deal with them conservatively.<br></div><div class="gmail_extra"></div><div class="gmail_extra"><br clear="all"><div><div><div dir="ltr"> - Vaivaswatha<br></div></div></div></div><div class="gmail_extra">
<br><div class="gmail_quote">On Fri, Dec 4, 2015 at 9:51 PM, Mehdi Amini <span dir="ltr"><<a href="mailto:mehdi.amini@apple.com" target="_blank">mehdi.amini@apple.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word"><br><div><span><blockquote type="cite"><div>On Dec 4, 2015, at 8:09 AM, Vaivaswatha Nagaraj <<a href="mailto:vn@compilertree.com" target="_blank">vn@compilertree.com</a>> wrote:</div><br><div><div dir="ltr"><div><div><div>>what is a non-public state that no-one but you can access? (I’d call that private).<br></div>malloc and free could both use global variables that are defined in libc, but are inaccessible to the program under compilation. <br></div></div></div></div></blockquote><div><br></div><div><br></div></span><div>If malloc and free can both use global variables (there is no notion of library in the compiler), then from what I understand you are actually creating another global state: i.e. there would be two disjoint global states: the usual default one and another one that is only accessed by function having this attribute.</div><div><br></div><div>— </div><span><font color="#888888"><div>Mehdi</div></font></span><div><div><div><br></div><div><br></div><br><blockquote type="cite"><div><div dir="ltr"><div><div><br>>if you’re attribute is saying they have some internal state, then malloc() cannot access the state of free() and vice versa.<br></div>Which is why it would be preferable to call it "inaccessible" state rather than "internal".<br><br>>It would prevent to swap two mallocs but not moving freely a malloc with respect to a free.<br></div>No, it would also prevent interchanging the order of malloc and free, since they both maintain states (which can be shared, but not accessible to the program under compilation) and the swapping order could result in a different final state.<br><div><br>>At the cost of some redundancy, I think a new attribute is needed.</div><div>@hal. I'm not sure what this implies. Does the semantics of the attribute in the first mail sound right to you?<br><br>> that'd be redefining the semantics of ReadNone. ReadNone allows elision of a call if its result is unused,<br></div><div>@James. That right. Optimizations should hereafter (if the proposed attribute is accepted) be more careful in interpreting ReadNone. If the call also has HasInaccessibleState, it shouldn't remove the call, even if the call takes no arguments or its return value isn't, because it could be modifying some internal state.<br></div><div><br></div></div><div class="gmail_extra"><br clear="all"><div><div><div dir="ltr"> - Vaivaswatha<br></div></div></div>
<br><div class="gmail_quote">On Fri, Dec 4, 2015 at 9:10 PM, Mehdi Amini <span dir="ltr"><<a href="mailto:mehdi.amini@apple.com" target="_blank">mehdi.amini@apple.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word"><br><div><span><blockquote type="cite"><div>On Dec 3, 2015, at 11:58 PM, Vaivaswatha Nagaraj <<a href="mailto:vn@compilertree.com" target="_blank">vn@compilertree.com</a>> wrote:</div><br><div><div dir="ltr"><div><div>>is this "internal state” supposed to be private to the function?<br></div>It could be private or not. Hence the name "inaccessible", to mean that the program under compilation has no access to the state. So while printf and malloc (for example) could share state in libc, the program under compilation cannot access this state.<br></div></div></div></blockquote><div><br></div></span><div>This is still not clear to me, you’re saying “it could be private or not”: what is a non-public state that no-one but you can access? (I’d call that private).</div><div><br></div><div>Now, from the point of view of the compiler, malloc and free are two separate functions, if you’re attribute is saying they have some internal state, then malloc() cannot access the state of free() and vice versa.</div><span><br><blockquote type="cite"><div><div dir="ltr"><div>>how this flag would prevent the last “optimization” you’re illustrating<br></div>Assuming you are referring to the quoted examples, currently these optimizations are not happening anyway (from what I understand). The issue is that, after malloc/free are tagged with "ReadNone", such transforms may happen. Hence to prevent that, the additional flag denoting that these functions maintain an internal state. <br></div></div></blockquote><div><br></div></span><div>I’m questioning why would this flag solve that, it does not seem to to me. It would prevent to swap two mallocs but not moving freely a malloc with respect to a free.</div><div><br></div><div>— </div><span><font color="#888888"><div>Mehdi</div></font></span><div><div><div><br></div><br><blockquote type="cite"><div><div dir="ltr"><div><br></div></div><div class="gmail_extra"><br clear="all"><div><div><div dir="ltr"> - Vaivaswatha<br></div></div></div>
<br><div class="gmail_quote">On Fri, Dec 4, 2015 at 12:20 PM, Mehdi Amini <span dir="ltr"><<a href="mailto:mehdi.amini@apple.com" target="_blank">mehdi.amini@apple.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word">Hi,<div><br></div><div><div><div><div><blockquote type="cite"><div>On Dec 3, 2015, at 10:33 PM, Vaivaswatha Nagaraj via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>> wrote:</div><br><div><div dir="ltr"><div><div>Hi,<br><br></div>This email is in continuation to the mail thread <a href="http://lists.llvm.org/pipermail/llvm-dev/2015-December/092996.html" target="_blank">http://lists.llvm.org/pipermail/llvm-dev/2015-December/092996.html</a>, to propose a new function attribute that can convey that a function maintains state, but this state is inaccessible to the rest of the program under compilation.<br><br></div>Such a flag could be added to most libc/system calls such as printf/malloc/free. (libc and system calls do access/modify internal variables such as errno).<br><div><div><br></div><div>Example attributes (in addition to what are already set):<br></div><div>malloc/free: HasInaccessibleState, ReadNone<br></div><div>printf: HasInaccessibleState, ArgMemOnly<br></div><div>realloc: HasInaccessibleState, ReadOnly (not sure).<br><br></div><div>The intention behind introducing this attribute is to relax the conditions in GlobalsAA as below:<br></div><div>(this code is in GlobalsAAResult::AnalyzeCallGraph)<br></div><div><pre> if (F->isDeclaration()) {
// Try to get mod/ref behaviour from function attributes.
- if (F->doesNotAccessMemory()) {
+ if (F->doesNotAccessMemory() || F->onlyAccessesArgMemory()) {
// Can't do better than that!
} else if (F->onlyReadsMemory()) {
FunctionEffect |= Ref;
if (!F->isIntrinsic())
// This function might call back into the module and read a global -
// consider every global as possibly being read by this function.
FR.MayReadAnyGlobal = true;
} else {
FunctionEffect |= ModRef;
// Can't say anything useful unless it's an intrinsic - they don't
// read or write global variables of the kind considered here.
KnowNothing = !F->isIntrinsic();
}
continue;
}</pre>This relaxation allows functions that (transitively) call library functions (such as printf/malloc) to still maintain and propagate GlobalsAA info. In general, this adds more precision to the description of these functions. <br><br></div><div>Concerns regarding impact on other optimizations (I'm repeating a few examples that Hal mentioned earlier).<br><br>1. <br>><span>A readnone
function is one whose output is a function only of its inputs, and if
you have this:<br>
><br>
> int *x = malloc(4);<br>
> *x = 2;<br>
> int *y = malloc(4);<br>
> *y = 4;<br>
</span><span>> you certainly don't want EarlyCSE to replace the
second call to malloc with the result of the first (which it will
happily do if you mark malloc as readnone).<br><br></span></div><div><span><span><span>For malloc, even though ReadNone is set now (as proposed above), EarlyCSE should be taught to respect the HasInaccessibleState and not combine functions that maintain internal states.</span> <span>Similarly other optimizations (such as DCE) must be taught to respect the flag.</span><br><br>2.<br></span></span><span><span>>void foo(char * restrict s1, char * restrict s2) {<br>
> printf(s1);<br>
> printf(s2);<br>>}<br>
<br>>If printf is argmemonly, then we could interchange the two printf calls.<br><br></span></span></div><div><span><span>In this example too, printf maintains an internal state, preventing the calls from being internchanged</span>. <span>Also, it is now correct to add ArgMemOnly to printf as it does not access any other program memory.</span> <br><br>3.<br></span>>For malloc this is still a problem, in the following sense, if we have:<br>
><br>> p1 = malloc(really_big);<br>> ...<br>> free(p1);<br>
><br>> p2 = malloc(really_big);<br>> ...<br>> free(p2);<br>
>allowing a transformation into:<br>
> p1 = malloc(really_big);<br>> p2 = malloc(really_big);<br>
> ...<br>
> free(p1); free(p2);<br>
>could be problematic.</div><div><span><br></span></div><div><span><span>Both free and malloc would be marked with having an internal state. This should prevent this kind of an optimization</span>. <span>Note that having the ReadNone attribute should not cause problems anymore.</span><br></span></div></div></div></div></blockquote><div><br></div></div></div><div><div>Something is not clear to me: is this "internal state” supposed to be private to the function?</div><div>How does it deal with malloc/free which can be seen as sharing a state? Especially it is not clear to me how this flag would prevent the last “optimization” you’re illustrating.</div><div><br></div><div>— </div><span><font color="#888888"><div>Mehdi</div><div><br></div><div><br><div><blockquote type="cite"></blockquote></div></div></font></span></div></div><br></div></div></blockquote></div><br></div>
</div></blockquote></div></div></div><br></div></blockquote></div><br></div>
</div></blockquote></div></div></div><br></div></blockquote></div><br></div></div></div><span class="">
_______________________________________________<br>
LLVM Developers mailing list<br>
<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev" rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
</span></blockquote></div>
</blockquote></div><br></div>