Bug 95852 - Failure to optimize __builtin_mul_overflow pattern
Summary: Failure to optimize __builtin_mul_overflow pattern
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: tree-optimization (show other bugs)
Version: 11.0
: P3 normal
Target Milestone: 11.0
Assignee: Jakub Jelinek
URL:
Keywords: missed-optimization
Depends on:
Blocks: 19987
  Show dependency treegraph
 
Reported: 2020-06-23 21:50 UTC by Gabriel Ravier
Modified: 2024-07-29 17:04 UTC (History)
1 user (show)

See Also:
Host:
Target:
Build:
Known to work:
Known to fail:
Last reconfirmed: 2021-01-07 00:00:00


Attachments
gcc11-pr95852-wip.patch (1.78 KB, patch)
2021-01-07 19:26 UTC, Jakub Jelinek
Details | Diff
gcc11-pr95852.patch (5.02 KB, patch)
2021-01-08 13:45 UTC, Jakub Jelinek
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Gabriel Ravier 2020-06-23 21:50:31 UTC
bool f(unsigned x, unsigned y, unsigned* res)
{
    *res = x * y;
    return x && ((*res / x) != y);
}

This can be optimized to `return __builtin_mul_overflow(x, y, res);`. This transformation is done by LLVM, but not by GCC.
Comment 1 Jakub Jelinek 2021-01-07 19:26:46 UTC
Created attachment 49913 [details]
gcc11-pr95852-wip.patch

Untested WIP patch.  This handles the transformation to mullo but doesn't yet optimize away the x != 0 && (which can be done of course only if the x != 0 guarded basic block contains only specific statements we'd need to pattern recognize.
Furthermore, signed multiplication can be handled similarly, but it will need even more work because of the casts.
Comment 2 Jakub Jelinek 2021-01-08 13:45:21 UTC
Created attachment 49919 [details]
gcc11-pr95852.patch

Full untested patch (so far only for unsigned __builtin_mul_overflow{,_p} pattern matching though).  Signed will be handled incrementally.
Comment 3 GCC Commits 2021-01-11 09:36:37 UTC
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:a2106317cd6673e110b347c70f21e25fbb23379e

commit r11-6579-ga2106317cd6673e110b347c70f21e25fbb23379e
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Jan 11 10:32:07 2021 +0100

    widening_mul: Pattern recognize unsigned multiplication with overflow check [PR95852]
    
    The following patch pattern recognizes some forms of multiplication followed
    by overflow check through division by one of the operands compared to the
    other one, with optional removal of guarding non-zero check for that operand
    if possible.  The patterns are replaced with effectively
    __builtin_mul_overflow or __builtin_mul_overflow_p.  The testcases cover 64
    different forms of that.
    
    2021-01-11  Jakub Jelinek  <jakub@redhat.com>
    
            PR tree-optimization/95852
            * tree-ssa-math-opts.c (maybe_optimize_guarding_check): New function.
            (uaddsub_overflow_check_p): Renamed to ...
            (arith_overflow_check_p): ... this.  Handle also multiplication
            with overflow check.
            (match_uaddsub_overflow): Renamed to ...
            (match_arith_overflow): ... this.  Add cfg_changed argument.  Handle
            also multiplication with overflow check.  Adjust function comment.
            (math_opts_dom_walker::after_dom_children): Adjust callers.  Call
            match_arith_overflow also for MULT_EXPR.
    
            * gcc.target/i386/pr95852-1.c: New test.
            * gcc.target/i386/pr95852-2.c: New test.
Comment 4 GCC Commits 2021-01-11 09:36:42 UTC
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:9febe9e4be7812519258ea3ed4f38bbc1a61624b

commit r11-6580-g9febe9e4be7812519258ea3ed4f38bbc1a61624b
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Jan 11 10:34:07 2021 +0100

    widening_mul: Pattern recognize also signed multiplication with overflow check [PR95852]
    
    On top of the previous widening_mul patch, this one recognizes also
    (non-perfect) signed multiplication with overflow, like:
    int
    f5 (int x, int y, int *res)
    {
      *res = (unsigned) x * y;
      return x && (*res / x) != y;
    }
    The problem with such checks is that they invoke UB if x is -1 and
    y is INT_MIN during the division, but perhaps the code knows that
    those values won't appear.  As that case is UB, we can do for that
    case whatever we want and handling that case as signed overflow
    is the best option.  If x is a constant not equal to -1, then the checks
    are 100% correct though.
    Haven't tried to pattern match bullet-proof checks, because I really don't
    know if users would write it in real-world code like that,
    perhaps
      *res = (unsigned) x * y;
      return x && (x == -1 ? (*res / y) != x : (*res / x) != y);
    ?
    
    https://wiki.sei.cmu.edu/confluence/display/c/INT32-C.+Ensure+that+operations+on+signed+integers+do+not+result+in+overflow
    suggests to use twice as wide multiplication (perhaps we should handle that
    too, for both signed and unsigned), or some very large code
    with 4 different divisions nested in many conditionals, no way one can
    match all the possible variants thereof.
    
    2021-01-11  Jakub Jelinek  <jakub@redhat.com>
    
            PR tree-optimization/95852
            * tree-ssa-math-opts.c (maybe_optimize_guarding_check): Change
            mul_stmts parameter type to vec<gimple *> &.  Before cond_stmt
            allow in the bb any of the stmts in that vector, div_stmt and
            up to 3 cast stmts.
            (arith_cast_equal_p): New function.
            (arith_overflow_check_p): Add cast_stmt argument, handle signed
            multiply overflow checks.
            (match_arith_overflow): Adjust caller.  Handle signed multiply
            overflow checks.
    
            * gcc.target/i386/pr95852-3.c: New test.
            * gcc.target/i386/pr95852-4.c: New test.
Comment 5 Jakub Jelinek 2021-01-11 09:37:17 UTC
Fixed.