[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 Jakub Jelinek changed: What|Removed |Added Status|REOPENED|RESOLVED CC||jakub at gcc dot gnu.org Resolution|--- |DUPLICATE --- Comment #12 from Jakub Jelinek --- Dup. *** This bug has been marked as a duplicate of bug 89084 ***
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 --- Comment #11 from Jerry DeLisle --- The missing local variable exists in the fortran dump and it shows as the first item in the namespace passed to gfc_process_block_locals. However, it has no backend decl. I do not understand enough to proceed
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009
--- Comment #10 from Jerry DeLisle ---
In the tree dump, the declaration for idxs is missing for the first subroutine.
I have modified to the following example so one can see it is not symbol name
conflicts.
MODULE sparse_matrix_csx_benchmark_utils
IMPLICIT NONE
CONTAINS
SUBROUTINE sparse_matrix_csr_benchmark ( )
WRITE(*,*) 'At*x: t1'
block
integer, dimension(1), parameter :: idxs1=[1]
integer :: i, idx
do i = 1, size(idxs1)
idx = idxs1(i)
enddo
end block
END SUBROUTINE sparse_matrix_csr_benchmark
SUBROUTINE sparse_matrix_csc_benchmark ( )
WRITE(*,*) 'An*x: t2'
block
integer, dimension(1), parameter :: idxs2=[1]
integer :: i, idx
do i = 1, size(idxs2)
idx = idxs2(i)
enddo
end block
END SUBROUTINE sparse_matrix_csc_benchmark
END MODULE sparse_matrix_csx_benchmark_utils
program main
use sparse_matrix_csx_benchmark_utils
call sparse_matrix_csr_benchmark ()
call sparse_matrix_csc_benchmark ()
end program
and the tree dump:
sparse_matrix_csc_benchmark ()
{
{
integer(kind=4) i;
integer(kind=4) idx;
{
struct __st_parameter_dt dt_parm.0;
dt_parm.0.common.filename = &"pr82009.f03"[1]{lb: 1 sz: 1};
dt_parm.0.common.line = 15;
dt_parm.0.common.flags = 128;
dt_parm.0.common.unit = 6;
_gfortran_st_write (_parm.0);
_gfortran_transfer_character_write (_parm.0, &"An*x: t2"[1]{lb: 1 sz:
1}, 8);
_gfortran_st_write_done (_parm.0);
}
i = 1;
while (1)
{
{
logical(kind=4) D.3777;
D.3777 = i > 1;
if (D.3777) goto L.3;
idx = idxs2[(integer(kind=8)) i + -1];
L.2:;
i = i + 1;
}
}
L.3:;
L.1:;
}
}
sparse_matrix_csr_benchmark ()
{
{
static integer(kind=4) idxs2[1] = {1};
integer(kind=4) i;
integer(kind=4) idx;
{
struct __st_parameter_dt dt_parm.1;
dt_parm.1.common.filename = &"pr82009.f03"[1]{lb: 1 sz: 1};
dt_parm.1.common.line = 5;
dt_parm.1.common.flags = 128;
dt_parm.1.common.unit = 6;
_gfortran_st_write (_parm.1);
_gfortran_transfer_character_write (_parm.1, &"At*x: t1"[1]{lb: 1 sz:
1}, 8);
_gfortran_st_write_done (_parm.1);
}
i = 1;
while (1)
{
{
logical(kind=4) D.3786;
D.3786 = i > 1;
if (D.3786) goto L.6;
idx = idxs1[(integer(kind=8)) i + -1];
L.5:;
i = i + 1;
}
}
L.6:;
L.4:;
}
}
MAIN__ ()
{
sparse_matrix_csr_benchmark ();
sparse_matrix_csc_benchmark ();
}
__attribute__((externally_visible))
main (integer(kind=4) argc, character(kind=1) * * argv)
{
static integer(kind=4) options.2[7] = {2116, 4095, 0, 1, 1, 0, 31};
_gfortran_set_args (argc, argv);
_gfortran_set_options (7, [0]);
MAIN__ ();
return 0;
}
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 Jerry DeLisle changed: What|Removed |Added Status|RESOLVED|REOPENED Resolution|FIXED |--- --- Comment #9 from Jerry DeLisle --- (In reply to Dominique d'Humieres from comment #8) > > Fixed on trunk. Closing > > When compiling the original test, the ICE is gone but I get: > > Undefined symbols for architecture x86_64: > "_idxs.3783", referenced from: > ___sparse_matrix_csx_benchmark_utils_MOD_sparse_matrix_csc_benchmark > in ccP98FL2.o > "_idxs.3792", referenced from: > ___sparse_matrix_csx_benchmark_utils_MOD_sparse_matrix_csr_benchmark > in ccP98FL2.o > > This looks weird. Agree, I added a main program to that test case and now I see similar, so reopening.
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 --- Comment #8 from Dominique d'Humieres --- > Fixed on trunk. Closing When compiling the original test, the ICE is gone but I get: Undefined symbols for architecture x86_64: "_idxs.3783", referenced from: ___sparse_matrix_csx_benchmark_utils_MOD_sparse_matrix_csc_benchmark in ccP98FL2.o "_idxs.3792", referenced from: ___sparse_matrix_csx_benchmark_utils_MOD_sparse_matrix_csr_benchmark in ccP98FL2.o This looks weird.
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 Jerry DeLisle changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #7 from Jerry DeLisle --- Fixed on trunk. Closing
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 --- Comment #6 from Jerry DeLisle --- Author: jvdelisle Date: Wed Jul 4 18:08:16 2018 New Revision: 262416 URL: https://gcc.gnu.org/viewcvs?rev=262416=gcc=rev Log: 2018-07-04 Jerry DeLisle PR fortran/82009 * trans-decl.c (gfc_process_block_locals): Delete assert and set saved_local_decls = NULL_TREE. * gfortran.dg/block_16.f08. New test. Added: trunk/gcc/testsuite/gfortran.dg/block_16.f08 Modified: trunk/gcc/fortran/ChangeLog trunk/gcc/fortran/trans-decl.c trunk/gcc/testsuite/ChangeLog
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009
Jerry DeLisle changed:
What|Removed |Added
CC||jvdelisle at gcc dot gnu.org
--- Comment #5 from Jerry DeLisle ---
This seems on the verge of silly. The ICE is at an assert:
assert (saved_local_decls == NULL_TREE);
When the function gfc_process_block_locals concludes it does
saved_local_decls = NULL_TREE;
So as I looked at this I thought it is expected to be already set at NULL_TREE
and it leaves it that way. Why not just set it instead of doing the assert.
So I did this:
diff --git a/gcc/fortran/trans-decl.c b/gcc/fortran/trans-decl.c
index 254768c5828..08c1ebd2d4b 100644
--- a/gcc/fortran/trans-decl.c
+++ b/gcc/fortran/trans-decl.c
@@ -6751,7 +6751,7 @@ gfc_process_block_locals (gfc_namespace* ns)
{
tree decl;
- gcc_assert (saved_local_decls == NULL_TREE);
+ saved_local_decls = NULL_TREE;
has_coarray_vars = false;
generate_local_vars (ns);
The problem goes away and it regression tests fine.
saved_local_decls is used only in one other place where it is set in
add_decl_as_local(). It is not reset anywhere else. So I think the pacth above
is likely correct.
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 kargl at gcc dot gnu.org changed: What|Removed |Added Priority|P3 |P4 --- Comment #4 from kargl at gcc dot gnu.org --- (In reply to Valery Weber from comment #0) > hi all > > the following code is ICEing with gcc 7.2.0 > > thanks > > v > > cat sparse_matrix_csx_benchmark_utils.F90 > MODULE sparse_matrix_csx_benchmark_utils > IMPLICIT NONE > CONTAINS > SUBROUTINE sparse_matrix_csr_benchmark ( ) > WRITE(*,*) 'At*x: t' > block > integer, dimension(1), parameter :: idxs=[1] > integer :: i, idx > do i = 1, size(idxs) > idx = idxs(i) > enddo > end block > END SUBROUTINE sparse_matrix_csr_benchmark > SUBROUTINE sparse_matrix_csc_benchmark ( ) > WRITE(*,*) 'An*x: t' > block > integer, dimension(1), parameter :: idxs=[1] > integer :: i, idx > do i = 1, size(idxs) > idx = idxs(i) > enddo > end block > END SUBROUTINE sparse_matrix_csc_benchmark > END MODULE sparse_matrix_csx_benchmark_utils This one is weird. If I move the 2nd WRITE in sparse_matrix_csr_benchmark to after the BLOCK constructor, the code compiles. If I leave the 2nd WRITE in its current location, and comment out the 1st WRITE statement in sparse_matrix_csr_benchmark, I still get the ICE. I'm perplexed!
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 --- Comment #3 from kargl at gcc dot gnu.org --- (In reply to kargl from comment #2) > (In reply to janus from comment #1) > > Confirmed. ICEs with every gfortran version I tried, from 4.7 up to trunk. > > > > Slight reduction: > > > > MODULE csx > > IMPLICIT NONE > > CONTAINS > > SUBROUTINE csr > > block > > end block > > END > > SUBROUTINE csc > > WRITE(*,*) 'An*x: t' > > block > > integer, dimension(1), parameter :: idxs=[1] > > print *, idxs(1) > > end block > > END > > END > > The above code compiles for me on trunk (gcc version 9.0.0 20180609) Original testcase still causes an ICE.
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 kargl at gcc dot gnu.org changed: What|Removed |Added CC||kargl at gcc dot gnu.org --- Comment #2 from kargl at gcc dot gnu.org --- (In reply to janus from comment #1) > Confirmed. ICEs with every gfortran version I tried, from 4.7 up to trunk. > > Slight reduction: > > MODULE csx > IMPLICIT NONE > CONTAINS > SUBROUTINE csr > block > end block > END > SUBROUTINE csc > WRITE(*,*) 'An*x: t' > block > integer, dimension(1), parameter :: idxs=[1] > print *, idxs(1) > end block > END > END The above code compiles for me on trunk (gcc version 9.0.0 20180609)
[Bug fortran/82009] [F08] ICE with block construct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82009 janus at gcc dot gnu.org changed: What|Removed |Added Keywords||ice-on-valid-code Status|UNCONFIRMED |NEW Last reconfirmed||2017-08-28 CC||janus at gcc dot gnu.org Summary|ICE with block construct|[F08] ICE with block ||construct Ever confirmed|0 |1 --- Comment #1 from janus at gcc dot gnu.org --- Confirmed. ICEs with every gfortran version I tried, from 4.7 up to trunk. Slight reduction: MODULE csx IMPLICIT NONE CONTAINS SUBROUTINE csr block end block END SUBROUTINE csc WRITE(*,*) 'An*x: t' block integer, dimension(1), parameter :: idxs=[1] print *, idxs(1) end block END END
