[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[LpSolit at netscape dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

--- Comment #10 from Dmitry G. Dyachenko  ---
r249961 PASS for me

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[LpSolit at netscape dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

--- Comment #10 from Dmitry G. Dyachenko  ---
r249961 PASS for me

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[dimhen at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

--- Comment #10 from Dmitry G. Dyachenko  ---
r249961 PASS for me

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[rsandifo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

rsandifo at gcc dot gnu.org  changed:

   What|Removed |Added

 Status|ASSIGNED|RESOLVED
 Resolution|--- |FIXED

--- Comment #9 from rsandifo at gcc dot gnu.org  
---
Fixed.

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[rsandifo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

--- Comment #8 from rsandifo at gcc dot gnu.org  
---
Author: rsandifo
Date: Tue Jul  4 11:48:44 2017
New Revision: 249961

URL: https://gcc.gnu.org/viewcvs?rev=249961=gcc=rev
Log:
PR 81292: ICE on related strlens after r249880

r249880 installed the result of a strlen in a strinfo if the strinfo
wasn't previously a full string.  But as Jakub says in the PR comments,
we can't just do that in isolation, because there are no vdefs on the
call that would invalidate any related strinfos.

This patch updates the related strinfos if the adjustment is simple and
invalidates them otherwise.  As elsewhere, we treat adjustments of the
form strlen +/- INTEGER_CST as simple but anything else as too complex.

2017-07-04  Richard Sandiford  

gcc/
PR tree-optimization/81292
* tree-ssa-strlen.c (handle_builtin_strlen): When setting
full_string_p, also call adjust_related_strinfos if the adjustment
is simple, otherwise invalidate related strinfos.

gcc/testsuite/
PR tree-optimization/81292
* gcc.dg/pr81292-1.c: New test.
* gcc.dg/pr81292-2.c: Likewise.

Added:
trunk/gcc/testsuite/gcc.dg/pr81292-1.c
trunk/gcc/testsuite/gcc.dg/pr81292-2.c
Modified:
trunk/gcc/ChangeLog
trunk/gcc/testsuite/ChangeLog
trunk/gcc/tree-ssa-strlen.c

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

--- Comment #7 from Jakub Jelinek  ---
Comment on attachment 41671
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41671
Proposed patch

LGTM

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[rsandifo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

--- Comment #6 from rsandifo at gcc dot gnu.org  
---
Created attachment 41671
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41671=edit
Proposed patch

This is the patch I'm testing.

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[rsandifo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

rsandifo at gcc dot gnu.org  changed:

   What|Removed |Added

 Status|NEW |ASSIGNED
   Assignee|unassigned at gcc dot gnu.org  |rsandifo at gcc dot 
gnu.org

--- Comment #5 from rsandifo at gcc dot gnu.org  
---
(In reply to Jakub Jelinek from comment #4)
> The problem is that we hit:
>   if (si != NULL)
> {
>   if (!si->full_string_p && !si->stmt)
> {
>   /* Until now we only had a lower bound on the string length.
>  Install LHS as the actual length.  */
>   si = unshare_strinfo (si);
>   si->nonzero_chars = lhs;
>   si->full_string_p = true;
> }
>   return;
> }
> in handle_builtin_strlen, which changes just that single strinfo record, but
> doesn't do anything for the related ones.  And as strlen doesn't have a vdef
> (it is a pure function), nothing is invalidated, so we end up with a mixture
> of related strinfos where some strinfos are full_string_p and others are
> not,

Bah, sorry, I hadn't considered the lack of a vdef here.

> and e.g. zero_length_string has asserts that this does not happen.
> So, either handle_builtin_strlen needs to adjust also the related strinfos
> if any (note, maybe even if verify_related_strinfos fails we might need to
> do that or invalidate them manually), or we need to invalidate them, or not
> to record this change if we can't adjust or invalidate them all.

Ack.  Will fix.

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

Jakub Jelinek  changed:

   What|Removed |Added

 CC||rsandifo at gcc dot gnu.org

--- Comment #4 from Jakub Jelinek  ---
The problem is that we hit:
  if (si != NULL)
{
  if (!si->full_string_p && !si->stmt)
{
  /* Until now we only had a lower bound on the string length.
 Install LHS as the actual length.  */
  si = unshare_strinfo (si);
  si->nonzero_chars = lhs;
  si->full_string_p = true;
}
  return;
}
in handle_builtin_strlen, which changes just that single strinfo record, but
doesn't do anything for the related ones.  And as strlen doesn't have a vdef
(it is a pure function), nothing is invalidated, so we end up with a mixture of
related strinfos where some strinfos are full_string_p and others are not, and
e.g. zero_length_string has asserts that this does not happen.
So, either handle_builtin_strlen needs to adjust also the related strinfos if
any (note, maybe even if verify_related_strinfos fails we might need to do that
or invalidate them manually), or we need to invalidate them, or not to record
this change if we can't adjust or invalidate them all.

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

Jakub Jelinek  changed:

   What|Removed |Added

 CC||jakub at gcc dot gnu.org

--- Comment #3 from Jakub Jelinek  ---
Better testcase, the above has UB:
void foo (char *x)
{
  x[__builtin_strlen (x)] = ' ';
  x[__builtin_strlen (x)] = ' ';
}

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

Richard Biener  changed:

   What|Removed |Added

   Target Milestone|--- |8.0

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

--- Comment #2 from Marek Polacek  ---
Likely r249880.

[Bug tree-optimization/81292] [8 regression] ICE in zero_length_string, at tree-ssa-strlen.c:822

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81292

Marek Polacek  changed:

   What|Removed |Added

 Status|UNCONFIRMED |NEW
   Last reconfirmed||2017-07-03
 CC||mpolacek at gcc dot gnu.org
 Ever confirmed|0   |1

--- Comment #1 from Marek Polacek  ---
Confirmed.