[cfe-dev] PR16833 - Analyzer hangs trying to analyze large case ranges

[Jordan Rose]

Hm, okay. Here's some basic structure for what happens now:

1. SimpleConstraintManager::canReasonAbout decides whether or not a symbolic expression is even worth processing.
2. SimpleConstraintManager::assumeAux (the one that takes a NonLoc) does some preprocessing and canonicalizing so that it can identify the parts of a constraint.
3. SimpleConstraintManager::assumeSymRel makes sure all the numbers involved are compatible, and then dispatches to the various RangeConstraintManager methods
4. RangeConstraintManager::assumeSym* handle the six possible relational operators and the minute tricky differences between them, applying the new constraints as intersections with the existing range sets.

If you really want to put this into the constraint manager, I think I would still do this the "sneaky way", by canonicalizing your "LOW < $sym < HIGH" into "$sym - LOW < HIGH - low" in SimpleConstraintManager::assumeAux. That achieves maximal code reuse. If you weren't to do it this way, I think you'd have to introduce a new assumeInRange method that could be overridden; this would probably be easier to get right but I'm not sure it would actually add value.

As a side note, I would consider writing your disjunction with & rather than &&. Because of &&'s short-circuiting, you'd never actually see that expression show up in assume() otherwise, and handling & would make it possible to handle similar cases in user code in the future.

Please continue to e-mail with any questions you may have!
Jordan


On Aug 18, 2014, at 7:41 , Zach Davis <zdavkeos at gmail.com> wrote:

> Let me see if I understand what you are saying:
> 
> We are able to symbolically evaluate (correctly) the "x >= lower && x
> <= upper" expression with SVal's. However, we can't 'remember' this
> result because we don't construct a Constraint on the expression, and
> even if we did the existing RangeConstraint manager doesn't support
> it.
> 
> I feel like the RangeConstraint manager is really the best way to fix
> this, even if it's not easy.  It doesn't seem to be a huge issue
> anyway, if someone _really_ wants to analyse a file with such an
> expression, they can use the patch i supplied in the mean time.
> 
> 
> Unfortunately, looking through the Constraint manager(s) code, I have
> no idea where to start.
> 
> Zach
> 
> On Tue, Aug 12, 2014 at 9:38 PM, Jordan Rose <jordan_rose at apple.com> wrote:
>> That FIXME, and indeed the code to support case ranges, actually predates
>> RangeConstraintManager, which implements most of what the FIXME is referring
>> to. (That is, the technical/historical reason we couldn't do it symbolically
>> is "it wasn't implemented yet".) So your change is in fact on the right
>> track.
>> 
>> However, what you've done is construct the expression "$x >= LOWER && $x <=
>> UPPER". This is correct, but it's not a form that RangeConstraintManager
>> will actually handle right now:
>> 
>> void clang_analyzer_eval(int);
>> void test(int i) {
>> switch (i) {
>> case 1 ... 5:
>> return;
>> }
>> clang_analyzer_eval(x == 2); // should print FALSE, but will print UNKNOWN
>> with your patch.
>> }
>> 
>> (clang_analyzer_eval is a special function used to probe the analyzer state,
>> handled by the debug.ExprInspection checker.)
>> 
>> The easiest way to fix this would be to use assumeDual on each expression
>> individually. Then you end up with three cases: $x < LOWER, $x is in range,
>> and $x > UPPER. Unfortunately, you'd then have to restructure the whole
>> function, which currently assumes that you end up with exactly one state
>> that can end up in the "default" branch. That wouldn't be too hard.
>> 
>> The harder way to solve this is to actually teach [Range]ConstraintManager
>> to recognize constraints like this, and do the right thing. The
>> implementation is already set up to handle disjoint sets of ranges, so you
>> could actually do this without changing everything. It's probably more messy
>> than you'd like, though.
>> 
>> The sneaky way to do this is to realize that "$x >= LOWER && $x <= UPPER" is
>> actually equivalent to "$x - (LOWER - MIN) <= UPPER - (LOWER - MIN)", where
>> MIN is the minimum value for the integer type, for integers with wraparound
>> semantics. (RangeConstraintManager actually takes advantage of this in
>> reverse to translate "$x - OFFSET <= VALUE" back into a pair of range
>> disjunctions.) The only trouble with doing something like this is that
>> wraparound is Hard™ and you/we will want to be very careful that this
>> actually handles everything correctly, for all the integer types we care
>> about. A definite benefit of this is that we're back to only having one
>> state, which prevents potential exponential explosion as analysis continues.
>> 
>> What do you think we should do here?
>> Jordan
>> 
>> 
>> On Aug 12, 2014, at 12:39 , Zach Davis <zdavkeos at gmail.com> wrote:
>> 
>> All-
>> 
>> I was looking over 16833 and had a few questions.
>> 
>> The issue is the analyzer tries to concretize all the values in a case
>> statement with a range value:
>> 
>>   switch (a) {
>>       case 1 ... 0xfffff: // hang's trying each value in the range
>>          return 1;
>>       default:
>>          return 0;
>>   }
>> 
>> Is there any reason we can't symbolically evaluate 1 <= a <= 0xfffff,
>> rather than trying each 'a == v for v in 1..0xfffff'? I've attached a
>> patch that attempts to do this and it seems to work (it no longer
>> hangs anyway).  Is there a technical/historical reason for doing it
>> the current way?
>> 
>> The FIXME above the block mentions using '"ranges" for NonLVals'. I'm
>> not sure what this means, or if any work has been done toward this.
>> If not, would there be any value in pursing a patch like this in the
>> meantime to resolve the performance issue?
>> 
>> Thanks,
>> Zach
>> <16833_hang_on_large_case_range.patch>_______________________________________________
>> cfe-dev mailing list
>> cfe-dev at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
>> 
>> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20140821/78f3313d/attachment.html>