PNaCl: Extend ExpandMulWithOverflow pass to handle uadd.with.overflow too

PNaCl: Extend ExpandMulWithOverflow pass to handle uadd.with.overflow too

It turned out that umul.with.overflow wasn't the only *.with.overflow
intrinsic usage introduced by Clang.

I knew that Clang's CGExprCXX.cpp generates umul.with.overflow for an
overflow check for C++'s "new Foo[]".  The same code for handling "new
Foo[]" also generates uadd.with.overflow in some cases.  This happens
if class Foo has a destructor or a delete[] operator that takes a size
argument.  In those cases, the C++ ABI adds a "cookie" to the
allocation which contains the array's size.

Rename the pass to "ExpandArithWithOverflow" and rename files
accordingly.

Also enable the pass.

BUG=https://code.google.com/p/nativeclient/issues/detail?id=3434
TEST=*.ll tests + trybots + GCC torture tests

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-llvm.git;a=commit;h=9a6f5fa

[eliben, 2013-05-28 17:59:54]

https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
File lib/Transforms/NaCl/ExpandArithWithOverflow.cpp (right):

https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
lib/Transforms/NaCl/ExpandArithWithOverflow.cpp:56: "Expand out some uses of
*.with.overflow intrinsics",
What does "some" mean here? You either expand or report fatal error, no? So this
can be seen as "expand out uses", without "some".

https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
lib/Transforms/NaCl/ExpandArithWithOverflow.cpp:140: bool Modified;
= false

[Mark Seaborn, 2013-05-28 18:41:22]

https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
File lib/Transforms/NaCl/ExpandArithWithOverflow.cpp (right):

https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
lib/Transforms/NaCl/ExpandArithWithOverflow.cpp:56: "Expand out some uses of
*.with.overflow intrinsics",
On 2013/05/28 17:59:55, eliben wrote:
> What does "some" mean here? You either expand or report fatal error, no? So
this
> can be seen as "expand out uses", without "some".

Well, it doesn't touch usub.with.overflow or s{add,sub,mul}.with.overflow (the
signed variants).

I think saying "Expand out uses of *.with.overflow intrinsics" would be a little
misleading since it implies "all".

"Some" is at least accurate.  This is a short description after all, and if
someone is interested in the details they can look at the pass itself.

https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
lib/Transforms/NaCl/ExpandArithWithOverflow.cpp:140: bool Modified;
On 2013/05/28 17:59:55, eliben wrote:
> = false

Thanks, fixed.

[eliben, 2013-05-28 18:45:13]

On 2013/05/28 18:41:22, Mark Seaborn wrote:
>
https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
> File lib/Transforms/NaCl/ExpandArithWithOverflow.cpp (right):
> 
>
https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
> lib/Transforms/NaCl/ExpandArithWithOverflow.cpp:56: "Expand out some uses of
> *.with.overflow intrinsics",
> On 2013/05/28 17:59:55, eliben wrote:
> > What does "some" mean here? You either expand or report fatal error, no? So
> this
> > can be seen as "expand out uses", without "some".
> 
> Well, it doesn't touch usub.with.overflow or s{add,sub,mul}.with.overflow (the
> signed variants).
> 
> I think saying "Expand out uses of *.with.overflow intrinsics" would be a
little
> misleading since it implies "all".
> 
> "Some" is at least accurate.  This is a short description after all, and if
> someone is interested in the details they can look at the pass itself.
> 

Yes, this makes sense.

>
https://codereview.chromium.org/15688011/diff/1/lib/Transforms/NaCl/ExpandAri...
> lib/Transforms/NaCl/ExpandArithWithOverflow.cpp:140: bool Modified;
> On 2013/05/28 17:59:55, eliben wrote:
> > = false
> 
> Thanks, fixed.

[eliben, 2013-05-28 18:45:20]

lgtm

[Mark Seaborn, 2013-05-28 18:54:09]

Committed patchset #3 manually as r9a6f5fa (presubmit successful).