<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 12/16/11 4:45 PM, Chris Lattner wrote:
<blockquote
cite="mid:ECA41BF4-D1D5-446A-BD6D-87ED2CEC5D73@apple.com"
type="cite">
<meta http-equiv="Content-Type" content="text/html;
charset=ISO-8859-1">
<br>
<div>
<div>On Dec 16, 2011, at 2:41 PM, John Criswell wrote:</div>
<br class="Apple-interchange-newline">
<blockquote type="cite">
<div bgcolor="#FFFFFF" text="#000000"> On 12/16/11 4:14 PM,
Chris Lattner wrote:
<blockquote
cite="mid:19D8F13C-C5A8-44D3-861A-2CFF72BC2D94@apple.com"
type="cite">
<div>
<div>On Dec 16, 2011, at 12:39 PM, Kostya Serebryany
wrote:</div>
<blockquote type="cite">
<div class="gmail_quote">
<blockquote class="gmail_quote" style="margin-top:
0px; margin-right: 0px; margin-bottom: 0px;
margin-left: 0.8ex; border-left-width: 1px;
border-left-color: rgb(204, 204, 204);
border-left-style: solid; padding-left: 1ex;
position: static; z-index: auto; ">
<div class="HOEnZb">
<div class="h5">> Do we consider the above
transformation legal?<br>
</div>
</div>
</blockquote>
</div>
</blockquote>
<div><br>
</div>
<div>Yes, the transformation is perfectly legal for the
normal compiler.</div>
</div>
</blockquote>
<br>
So how do you guarantee that the behavior is predictable
regardless of hardware platform if you don't define what the
behavior should be?<br>
</div>
</blockquote>
<div><br>
</div>
<div>I'm not sure what you mean. What isn't defined?</div>
</div>
</blockquote>
<br>
The alloca in question allocates 22 bytes. The 64-bit load in
Kostya's original email is accessing two additional bytes past the
end of the alloca (i.e., it is accessing array "elements" a[22] and
a[23]). Accessing that memory with a read or write is undefined
behavior. The program could fault, read zeros, read arbitrary bit
patterns, etc.<br>
<br>
In other words, the compiler is transforming this:<br>
<br>
return a[16] + a[21];<br>
<br>
into something like this:<br>
<br>
unsigned long * p = &(a[16]);<br>
unsigned long v = *p; // This accesses memory locations a[22] and
a[23]; doing so is undefined behavior<br>
(do some bit fiddling to extra a[16] and a[17] from v)<br>
<br>
The original code is memory safe and exhibits defined behavior. You
can do whatever crazy, semantics-preserving optimization you want,
run it on any crazy architecture you want, and it'll always exhibit
the same behavior.<br>
<br>
The optimized code exhibits undefined behavior. On most systems, it
just reads garbage data that is ignored by the compiler, but that's
really just a side-effect of how most OS's and architectures do
things. If you do some crazy transforms or run on some obscure
architecture, the optimized code may break.<br>
<br>
<blockquote
cite="mid:ECA41BF4-D1D5-446A-BD6D-87ED2CEC5D73@apple.com"
type="cite">
<div><br>
[snip]<br>
<blockquote type="cite">
<div bgcolor="#FFFFFF" text="#000000">What if you have a funky
architecture that someone is porting LLVM to, or someone is
using x86-32 segments in an interesting way?<br>
</div>
</blockquote>
<div><br>
</div>
<div>We'll burn that bridge when we get to it ;-)</div>
</div>
</blockquote>
<br>
ASAN got burnt; SAFECode probably got burnt, too. If we work around
it, some poor researcher or developer may get burnt by it, too, and
spend some time figuring out why his correct-looking program is not
acting properly. In other words, you're burning someone else's
bridge.<br>
<br>
Granted, perhaps the benefits of an incorrect optimization may
outweigh the loss of using LLVM on novel systems, but are you sure
that making the optimization work properly is going to be so
detrimental?<br>
<br>
<blockquote
cite="mid:ECA41BF4-D1D5-446A-BD6D-87ED2CEC5D73@apple.com"
type="cite">
<div><br>
<blockquote type="cite">
<div bgcolor="#FFFFFF" text="#000000">Moreover, I don't really
understand the rationale for allowing a transform to
introduce undefined behavior into programs that exhibit no
undefined behavior.</div>
</blockquote>
<div><br>
</div>
<div>There is no undefined behavior here. This is exactly
analogous to the code you get for bitfield accesses. If you
have an uninitialized struct and start storing into its fields
(to initialize it) you get a series of "load + mask + or +
store" operations. These are loading and touching "undefined"
bits in a completely defined way.</div>
</div>
</blockquote>
<br>
I'll agree that they're both undefined behavior, but I don't think
they fall within the same category. The bit-mask initializing issue
is a compromise you made because there either isn't an alternative
way to do it that has defined behavior, or such an alternative is
too expensive and difficult to implement and/or use.<br>
<br>
This appears to be a different case. Fixing the optimization looks
simple enough to me (am I missing something?), and I'm not convinced
that fixing it would hurt performance (although since I haven't run
an experiment, that is conjecture).<br>
<br>
So, perhaps I should ask this: if someone took the time to fix the
transform so that it checks both the alignment *and* the allocation
size and measures the resulting performance change, how much would
performance need to suffer before the cure was deemed worse than the
disease?<br>
<br>
-- John T.<br>
<br>
<br>
<br>
<br>
<blockquote
cite="mid:ECA41BF4-D1D5-446A-BD6D-87ED2CEC5D73@apple.com"
type="cite">
<div>
<div><br>
</div>
<div>-Chris</div>
</div>
<br>
</blockquote>
<br>
</body>
</html>