[Bug c/106992] New: aarch64: Unexpected maybe-uninitialized warning when compiling fdlibm

[Bug c/106992] New: aarch64: Unexpected maybe-uninitialized warning when compiling fdlibm

[shqking at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106992

            Bug ID: 106992
           Summary: aarch64: Unexpected maybe-uninitialized warning when
                    compiling fdlibm
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: shqking at gmail dot com
  Target Milestone: ---

Unexpected "maybe-uninitialized" warning was emitted when compiling fdlibm
library on Ubuntu 22.04 && AArch64 platform with GCC-12 and GCC-13.

How to reproduce it?

```
wget https://netlib.org/fdlibm/fdlibm.h
wget https://netlib.org/fdlibm/k_rem_pio2.c
gcc -Wall -Wextra -Werror -O2 -std=c11 -fno-strict-aliasing -c k_rem_pio2.c
```

The following warning would be produced.
```
k_rem_pio2.c: In function '__kernel_rem_pio2':
k_rem_pio2.c:188:13: error: this 'for' clause does not guard...
[-Werror=misleading-indentation]
  188 |             for(j=0,fw=0.0;j<=jx;j++) fw += x[j]*f[jx+i-j]; q[i] = fw;
      |             ^~~
k_rem_pio2.c:188:61: note: ...this statement, but the latter is misleadingly
indented as if it were guarded by the 'for'
  188 |             for(j=0,fw=0.0;j<=jx;j++) fw += x[j]*f[jx+i-j]; q[i] = fw;
      |                                                             ^
k_rem_pio2.c:293:24: error: 'fq' may be used uninitialized
[-Werror=maybe-uninitialized]
  293 |                 fw = fq[0]-fw;
      |                      ~~^~~
k_rem_pio2.c:171:27: note: 'fq' declared here
  171 |         double z,fw,f[20],fq[20],q[20];
      |                           ^~
k_rem_pio2.c:293:24: error: 'fq' may be used uninitialized
[-Werror=maybe-uninitialized]
  293 |                 fw = fq[0]-fw;
      |                      ~~^~~
k_rem_pio2.c:171:27: note: 'fq' declared here
  171 |         double z,fw,f[20],fq[20],q[20];
      |                           ^~
cc1: all warnings being treated as errors
```

Note-1: I think it's a false positive because 1) array fq[] is initialized by
the for-stmt at line 276. Elements at index 0 to index "jz" are initialized. 2)
"jz" should be a non-negative value from the comment at line 87.

Note-2: No warning is produced on Ubuntu 22.04 && x86 platform.

Note-3: Known to work: gcc-11. Known to fail: gcc-12.0.1, 12.1.0, 12.2.0,
13.0.0.

Note-4: I filed an example with the C case I mentioned, i.e. k_rem_pio2.c,
after the pre-processing. See https://godbolt.org/z/Ph6r7vWeP. You may try
different GCC versions and x86/aarch64 backends there.

Note-5: It's worth noting that GCC at x86 platform would also produce this
waring message if we remove the option "-fno-strict-aliasing".

[Bug tree-optimization/106992] [12/13 Regression] aarch64: Unexpected maybe-uninitialized warning when compiling fdlibm

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106992

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|aarch64                     |aarch64, x86_64-*-*
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2022-09-21
             Status|UNCONFIRMED                 |NEW
   Target Milestone|---                         |12.3
             Blocks|                            |24639
           Priority|P3                          |P2
            Summary|aarch64: Unexpected         |[12/13 Regression] aarch64:
                   |maybe-uninitialized warning |Unexpected
                   |when compiling fdlibm       |maybe-uninitialized warning
                   |                            |when compiling fdlibm
                 CC|                            |rguenth at gcc dot gnu.org

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed without -fno-strict-aliasing and with -fno-tree-vectorize
-fno-strict-aliasing on x86_64.

It's a bit odd - I traced walk_aliased_vdefs and indeed it didn't come along
a definition of fq.  We diagnose a PRE hoisted fq[0] but even with
-fno-tree-pre -fno-code-hoisting we diagnose that.

-fno-thread-jumps avoids the diagnostic.  The reference we diagnose is
introduced by thread2:

k_rem_pio2.c.195t.thread2:  _508 = fq[0];

This case is

jump thread: (80, 136) incoming edge;  (88, 90) normal (90, 94) nocopy;

  <bb 80> [local count: 437174]:
<L145>:
  if (jz_117 >= 0)
    goto <bb 86>; [89.00%]
  else
    goto <bb 88>; [11.00%]

  <bb 88> [local count: 437174]:
  # fw_363 = PHI <fw_260(87), 0.0(80)>
  if (ih_111 != 0)
    goto <bb 90>; [50.00%]
  else
    goto <bb 89>; [50.00%]

  <bb 90> [local count: 218587]:
  iftmp.2_253 = -fw_363;
  *y_237(D) = iftmp.2_253;
  _85 = fq[0];
  fw_255 = _85 - fw_363;
  if (jz_117 > 0)
    goto <bb 91>; [89.00%]
  else
    goto <bb 94>; [11.00%]

so it might be that we somehow thread a jz < 0 case (which would skip
the init loop) but I still can't see how ...

I've used

-Wall -O2 -std=c11  k_rem_pio2.c -fno-tree-vectorize -fno-strict-aliasing
-fno-code-hoisting -fno-tree-pre

for the above.


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=24639
[Bug 24639] [meta-bug] bug to track all Wuninitialized issues

[Bug tree-optimization/106992] [12/13/14 Regression] aarch64: Unexpected maybe-uninitialized warning when compiling fdlibm

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106992

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|12.3                        |12.4

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 12.3 is being released, retargeting bugs to GCC 12.4.