llvm.noalias patches
Hal Finkel via llvm-commits
llvm-commits at lists.llvm.org
Sun Jul 10 08:54:53 PDT 2016
Hi again,
Thanks to David Majnemer for reviewing many of the patches. David made an important observation: Instead of making the "looking through the intrinsic" logic about the llvm.noalias intrinsic specifically, we could add the 'returned' attribute to the relevant parameter and make the patches about looking through calls with a 'returned' parameter. I like this approach much better, and I'll rework the patches to do this. Also, we don't seem to infer the returned attribute; I'll post a patch for that too.
Another change I need to make is that I need to add a second metadata parameter to the intrinsic, analogous to the alias.scope metadata we have currently. While the semantics as they stand are fine, they're not completely sufficient for representing local restrict-qualified pointers in C where the pointer variable itself is updated. For example, if we have:
void foo(int *a) {
int *restrict r = a;
*r++ = 1;
*r++ = 2;
--r;
*r = 3;
}
With the Clang CodeGen patch, the resulting IR looks like this (after optimization):
define void @foo(i32* nocapture %a) local_unnamed_addr #0 {
entry:
%0 = tail call i32* @llvm.noalias.p0i32(i32* %a, metadata !1)
%incdec.ptr = getelementptr inbounds i32, i32* %0, i64 1
%1 = tail call i32* @llvm.noalias.p0i32(i32* %incdec.ptr, metadata !1)
store i32 1, i32* %0, align 4, !tbaa !4, !noalias !1
%incdec.ptr1 = getelementptr inbounds i32, i32* %1, i64 1
%2 = tail call i32* @llvm.noalias.p0i32(i32* %incdec.ptr1, metadata !1)
store i32 2, i32* %1, align 4, !tbaa !4, !noalias !1
%incdec.ptr2 = getelementptr inbounds i32, i32* %2, i64 -1
%3 = tail call i32* @llvm.noalias.p0i32(i32* %incdec.ptr2, metadata !1)
store i32 3, i32* %3, align 4, !tbaa !4, !noalias !1
ret void
}
which yields this:
NoAlias: store i32 2, i32* %1, align 4, !tbaa !4, !noalias !1 <-> store i32 1, i32* %0, align 4, !tbaa !4, !noalias !1
NoAlias: store i32 3, i32* %3, align 4, !tbaa !4, !noalias !1 <-> store i32 1, i32* %0, align 4, !tbaa !4, !noalias !1
NoAlias: store i32 3, i32* %3, align 4, !tbaa !4, !noalias !1 <-> store i32 2, i32* %1, align 4, !tbaa !4, !noalias !1
but that's not correct. The common case is that a restrict-qualified pointer is initialized once and then used as a base pointer for further computations. However, nothing prevents updating the pointer variable to point to an object it pointed to previously. The rule, as it stands, is that an access with a noalias scope using a pointer based on the intrinsic's return value (with a compatible scope parameter) does not alias with an access using a pointer with a compatible noalias scope not based on the intrinsics's return value. With a second metadata parameter, this will grow to be: ... not based on the return value from any intrinsic with the same second metadata parameter (the alias.scope parameter). MetadataAsValue objects have use lists, so these will be easy to find (the implementation already takes advantage of this fact for the existing metadata parameter).
-Hal
----- Original Message -----
> From: "Hal Finkel via llvm-commits" <llvm-commits at lists.llvm.org>
> To: "Daniel Berlin" <dberlin at dberlin.org>, "chandlerc" <chandlerc at gmail.com>, "Philip Reames"
> <listmail at philipreames.com>, "John McCall" <rjmccall at apple.com>
> Cc: "llvm-commits" <llvm-commits at lists.llvm.org>
> Sent: Saturday, July 9, 2016 4:32:34 PM
> Subject: llvm.noalias patches
>
> Hi everyone,
>
> After a year's hiatus, I've resumed working on the llvm.noalias
> patches. In case you don't recall the motivation, and would like to,
> please see:
> http://lists.llvm.org/pipermail/llvm-dev/2014-November/078755.html
> -- I've now rebased and updated the patches in response to the last
> round of review comments (embarrassingly, from last July). There are
> currently 19 patches in the set, mostly small. These patches should
> apply bottom-up (starting from the bottom of the list). I've put a
> (*) next to the non-trivial ones.
>
> Clang:
>
> llvm.noalias - Clang CodeGen - check restrict var map only for
> restrict-qualified lvalues
> http://reviews.llvm.org/D22189
>
> llvm.noalias - Clang CodeGen for local restrict-qualified
> pointers (*)
> http://reviews.llvm.org/D9403
>
> LLVM:
>
> llvm.noalias - don't look through it to simplify GEPs
> http://reviews.llvm.org/D22184
>
> llvm.noalias - isPointerDereferenceable should look through them
> http://reviews.llvm.org/D9384
>
> llvm.noalias - don't prevent loop vectorization
> http://reviews.llvm.org/D9382
>
> llvm.noalias - SCEV can look through it
> http://reviews.llvm.org/D9381
>
> llvm.noalias - Use noalias intrinsics when inlining (and update
> them when cloning metadata) (*)
> http://reviews.llvm.org/D9400
>
> llvm.noalias - Add IRBuilder support
> http://reviews.llvm.org/D9378
>
> llvm.noalias - Pointer comparison folding should look through
> them
> http://reviews.llvm.org/D9387
>
> llvm.noalias - computeKnownBits should look through them
> http://reviews.llvm.org/D9397
>
> llvm.noalias - CodeGen support
> http://reviews.llvm.org/D9380
>
> llvm.noalias - CaptureTracking needs to look through them
> http://reviews.llvm.org/D9386
>
> llvm.noalias - handling of dead intrinsics
> http://reviews.llvm.org/D9376
>
> llvm.noalias - don't interfere with llvm.assume
> http://reviews.llvm.org/D9379
>
> llvm.noalias - don't block EarlyCSE
> http://reviews.llvm.org/D9377
>
> llvm.noalias - AA impl (*)
> http://reviews.llvm.org/D9401
>
> llvm.noalias - look through with basicaa
> http://reviews.llvm.org/D9383
>
> llvm.noalias - GetUnderlyingObject intrinsic collection (*)
> http://reviews.llvm.org/D9398
>
> llvm.noalias - add the intrinsic (*)
> http://reviews.llvm.org/D9375
>
> Thanks again,
> Hal
>
> --
> Hal Finkel
> Assistant Computational Scientist
> Leadership Computing Facility
> Argonne National Laboratory
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
--
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory
More information about the llvm-commits
mailing list