Re: [PATCH] PR125535 wrong-code for implied-do with allocatable-component, derived type and transformational intrinsic

The master branch has been updated by Jerry DeLisle <[email protected]>:

https://gcc.gnu.org/g:b5a3b874b8df2ea2ca82bd97fcd5906f7ebe82d8

commit r17-1410-gb5a3b874b8df2ea2ca82bd97fcd5906f7ebe82d8
Author: Jerry DeLisle <[email protected]>
Date:   Fri Jun 5 13:20:05 2026 -0700


On 6/7/26 12:11 PM, Jerry D wrote:
> On 6/7/26 6:13 AM, Paul Richard Thomas wrote:
>> Hello Jerry,
>>
>> This patch looks good to me for mainline and gcc-16.
>>
>> However, in common with other test cases with nested array and
>> structure constructors, it leaks memory allocated at line 17 of the
>> testcase. This is something that we will have to work on because it
>> blocks the PDT meta-bug, for example.
>>
>> Please apply it as it is but it should be flagged up as a leaker
>> (another meta-bug?)
>>
> 
> Thanks Paul, I will push this. How many leakers have you identified?
> 
> Jerry
> 
> 
>> Thanks
>>
>> Paul
>>
>> On Sat, 6 Jun 2026 at 02:48, Jerry D <[email protected]> wrote:
>>>
>>> See attached patch.
>>>
>>> This one is fairly simple. This is also the last of the meta-bug PR125515.
>>>
>>> Regression tested on x86_64.
>>>
>>> I will probably simplify the commit message. Thought I should comments before I
>>> do that.
>>>
>>> OK for mainline and later backport to 16.
>>>
>>> Regards,
>>>
>>> Jerry
>>>
>>> ---
>>>
>>>
>>> When a nested implied-do array constructor called a transformational
>>> intrinsic (e.g. RESHAPE) whose result type has allocatable components, the
>>> argument temporaries were freed before the result's allocatable components
>>> were deep-copied, causing a heap-use-after-free and wrong runtime values.
>>>
>>> The root cause: in gfc_conv_procedure_call, the call to
>>> gfc_add_block_to_block (&se->pre, &post) — which frees the allocatable
>>> components of argument temporaries — was issued before the deep-copy loop
>>> for the result's allocatable components.  Transformational library functions
>>> such as RESHAPE do a shallow byte-copy of the source array into the result,
>>> so the result's component pointers alias those of the argument temporaries.
>>> Freeing the temporaries first and then copying yielded use-after-free.
>>>
>>> Fix: move gfc_add_block_to_block (&se->pre, &post) to after the deep-copy
>>> loop, so that the result's allocatable components are copied while the
>>> source storage is still live.
>>>
>>> PR fortran/125535
>>>
>>> Assisted by: Claude Sonnet 4.6
>>>
>>> gcc/fortran/ChangeLog:
>>>
>>>          PR fortran/125535
>>>          * trans-expr.cc (gfc_conv_procedure_call): Move post block append
>>>          to after the deep copy of allocatable components for transformational
>>>          intrinsics, so that argument temporaries are not freed before the
>>>          result components are copied.
>>>
>>> gcc/testsuite/ChangeLog:
>>>
>>>          PR fortran/125535
>>>          * gfortran.dg/implied_do_alloc_comp_1.f90: New test.
>>> ---
>