[Bug tree-optimization/85168] [7 Regression] ICE in tree-ssa-coalesce.c: SSA corruption: Unable to coalesce ssa_names which are marked as MUST COALESCE when -O2 is used
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85168 --- Comment #7 from Ivo Raisr --- Fix backported to 7.3.0 correctly compiles the original (huge) source module. Thank you, Richard.
[Bug tree-optimization/85168] ICE in tree-ssa-coalesce.c: SSA corruption: Unable to coalesce ssa_names which are marked as MUST COALESCE when -O2 is used
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85168 Ivo Raisr changed: What|Removed |Added CC||ivo.raisr at oracle dot com --- Comment #1 from Ivo Raisr --- Debug output for rtl-expand shows: ... Partition map Partition 0 (h_5(ab) - 5 ) Partition 1 (h_6(ab) - 6 ) Partition 2 (h_7 - 7 ) Partition 3 (h_12(ab) - 12 ) Partition 4 (h_14(ab) - 14 ) Partition 5 (h_20 - 20 ) Partition 6 (pretmp_21 - 21 ) Partition 7 (pretmp_22 - 22 ) Partition 8 (pretmp_23 - 23 ) Partition 9 (prephitmp_24 - 24 ) Conflict graph: 3: 4 4: 3 ... Abnormal coalesce: (6)h_6(ab) & (12)h_12(ab) [map: 1, 3] : Success -> 1 Abnormal coalesce: (6)h_6(ab) & (14)h_14(ab) [map: 1, 4] : Fail due to conflict I am not sure if the conflict is because h_12 and h_14 are still live at the same time or that h_7 and h_20 are not marked abnormal.
[Bug tree-optimization/85168] New: ICE in tree-ssa-coalesce.c: SSA corruption: Unable to coalesce ssa_names which are marked as MUST COALESCE when -O2 is used
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85168
Bug ID: 85168
Summary: ICE in tree-ssa-coalesce.c: SSA corruption: Unable to
coalesce ssa_names which are marked as MUST COALESCE
when -O2 is used
Product: gcc
Version: 8.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: ivo.raisr at oracle dot com
Target Milestone: ---
Created attachment 43815
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=43815=edit
reproducer.c
With gcc 8.0.1 built recently from SVN, I receive internal compiler error: SSA
corruption when -O2 is used:
void g(void) {
^
0xdcec2e fail_abnormal_edge_coalesce
../../src/gcc/tree-ssa-coalesce.c:1012
0xdcec2e coalesce_partitions
../../src/gcc/tree-ssa-coalesce.c:1401
0xdcec2e coalesce_ssa_name()
../../src/gcc/tree-ssa-coalesce.c:1888
0xd6450e remove_ssa_form
../../src/gcc/tree-outof-ssa.c:948
0xd6450e rewrite_out_of_ssa(ssaexpand*)
../../src/gcc/tree-outof-ssa.c:1174
0x8b2820 execute
../../src/gcc/cfgexpand.c:6227
The full source module is not publicly available. I have creduce'd it to the
following minimal form:
typedef struct {
struct {
char a;
} b;
} c;
int d, f;
c *e;
extern void i(void);
void g(void) {
c *h = e;
if (f) {
i();
h--;
if (d)
if (h->b.a)
i();
}
if (h->b.a)
__builtin_setjmp(0);
e = h;
}
The problem is reproducible with:
$ gcc -fdump-rtl-expand-details -O2 -c
but not with:
$ gcc -fdump-rtl-expand-details -O1 -c
[Bug sanitizer/82545] [7 Regression] -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545 --- Comment #11 from Ivo Raisr --- (In reply to Martin Liška from comment #9) > Fixed on trunk, queued for backports. I wonder if you had a chance for backporting this into gcc-7 branch? Let me know if I can be of assistance here.
[Bug sanitizer/82545] [7 Regression] -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545 --- Comment #10 from Ivo Raisr --- Thank you for the fix. Works on the full source file as well.
[Bug sanitizer/82545] [7/8 Regression] -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545 --- Comment #7 from Ivo Raisr --- If no viable solution to splitting an abnormal edge is found, I wonder if we could actually prevent getting into this situation?
[Bug tree-optimization/82545] -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545 --- Comment #4 from Ivo Raisr --- The reproducer was C-reduce'd from a large source file (30 MB) which is not publicly available.
[Bug tree-optimization/82545] -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545
--- Comment #3 from Ivo Raisr ---
$ cat reduced.c
extern void c(int);
extern void d(void);
void a(void) {
{
int b;
__builtin_setjmp(0);
c(b);
}
d();
}
[Bug tree-optimization/82545] -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545 --- Comment #2 from Ivo Raisr --- Created attachment 42352 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42352=edit C-reduced reproducer, not reformatted
[Bug tree-optimization/82545] -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545 Ivo Raisr changed: What|Removed |Added CC||ivo.raisr at oracle dot com --- Comment #1 from Ivo Raisr --- Created attachment 42351 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42351=edit reproducer, reformatted for better readability
[Bug tree-optimization/82545] New: -O2: ICE in gimple_split_edge, at tree-cfg.c:2838, during GIMPLE pass sanopt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82545
Bug ID: 82545
Summary: -O2: ICE in gimple_split_edge, at tree-cfg.c:2838,
during GIMPLE pass sanopt
Product: gcc
Version: 8.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: ivo.raisr at oracle dot com
Target Milestone: ---
gcc 8 and 7 produce an internal compiler error on the following program (also)
attached.
$ gcc --version
gcc (GCC) 8.0.0 20171012 (experimental)
$ uname -a
Linux 4.10.0-37-generic #41-Ubuntu SMP x86_64 x86_64 x86_64 GNU/Linux
$ gcc -O2 -fsanitize=address -c reduced.c
during GIMPLE pass: sanopt
reduced.c: In function ‘a’:
reduced.c:4:6: internal compiler error: in gimple_split_edge, at
tree-cfg.c:2838
void a(void) {
^
0x6b9315 gimple_split_edge
../../src/gcc/tree-cfg.c:2838
0x99a5cd split_edge(edge_def*)
../../src/gcc/cfghooks.c:638
0xafc5d7 gimple_find_edge_insert_loc
../../src/gcc/gimple-iterator.c:793
0xafd36d gsi_commit_one_edge_insert(edge_def*, basic_block_def**)
../../src/gcc/gimple-iterator.c:884
0xafd43e gsi_commit_edge_inserts()
../../src/gcc/gimple-iterator.c:863
0x6b0d11 execute
../../src/gcc/sanopt.c:1359
gcc 8: FAIL, ICE produced. SVN version 253695.
gcc 7: FAIL, ICE produced. SVN version 253695, gcc-7-branch.
gcc 6: PASS, ICE not produced. SVN version 253695, gcc-6-branch.
[Bug tree-optimization/82337] [5/6/7/8 Regression] ICE: SSA corruption at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82337 --- Comment #10 from Ivo Raisr --- (In reply to Bill Schmidt from comment #9) I confirm this fixes the problem also in the original full-blown source.
[Bug tree-optimization/82337] -O2: ICE: SSA corruption at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82337 --- Comment #2 from Ivo Raisr --- Created attachment 42241 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42241=edit reduced-3.c (slightly reformatted source to make it more readable)
[Bug tree-optimization/82337] -O2: ICE: SSA corruption at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82337 Ivo Raisr changed: What|Removed |Added CC||ivo.raisr at oracle dot com --- Comment #1 from Ivo Raisr --- Created attachment 42240 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42240=edit reduced-3.bare.c (output produced by C-reduce)
[Bug tree-optimization/82337] New: -O2: ICE: SSA corruption at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82337
Bug ID: 82337
Summary: -O2: ICE: SSA corruption at tree-ssa-coalesce.c:1010
Product: gcc
Version: 8.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: ivo.raisr at oracle dot com
Target Milestone: ---
With gcc 8.0, 7.2 and 6.3.1 we hit internal compiler error:
Unable to coalesce ssa_names 7 and 15 which are marked as MUST COALESCE.
i_7(ab) and i_15(ab)
during RTL pass: expand
reduced-3.c:12:6: internal compiler error: SSA corruption
void g(void) {
^
0xe24c8c fail_abnormal_edge_coalesce
../../gcc-trunk/gcc/tree-ssa-coalesce.c:1010
0xe24c8c coalesce_partitions
../../gcc-trunk/gcc/tree-ssa-coalesce.c:1399
0xe24c8c coalesce_ssa_name()
../../gcc-trunk/gcc/tree-ssa-coalesce.c:1886
0xdca93b remove_ssa_form
../../gcc-trunk/gcc/tree-outof-ssa.c:948
0xdca93b rewrite_out_of_ssa(ssaexpand*)
../../gcc-trunk/gcc/tree-outof-ssa.c:1172
0x992d70 execute
../../gcc-trunk/gcc/cfgexpand.c:6164
Full pre-processed sources are not publicly available.
C-reduce was able to produce "reduced-3.bare.c" which I slightly reformatted to
make it more readable: "reduced-3.c". Both files attached.
Reproduce as:
$ gcc -c reduced-3.c -O2
[Bug tree-optimization/82276] -O2: ICE: SSA corruption during RTL pass: expand; at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82276 --- Comment #3 from Ivo Raisr --- Reproducible with gcc 8.0 and 7.2. I was not able to reproduce it with 6.3.0.
[Bug tree-optimization/82276] -O2: ICE: SSA corruption during RTL pass: expand; at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82276 --- Comment #2 from Ivo Raisr --- Created attachment 42216 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42216=edit slightly reformatted output from C-reduce to make it more readable
[Bug tree-optimization/82276] -O2: ICE: SSA corruption during RTL pass: expand; at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82276 Ivo Raisr changed: What|Removed |Added CC||ivo.raisr at oracle dot com --- Comment #1 from Ivo Raisr --- Created attachment 42215 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42215=edit reduced-2.bare.c (output produced by C-reduce)
[Bug tree-optimization/82276] New: -O2: ICE: SSA corruption during RTL pass: expand; at tree-ssa-coalesce.c:1010
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82276
Bug ID: 82276
Summary: -O2: ICE: SSA corruption during RTL pass: expand; at
tree-ssa-coalesce.c:1010
Product: gcc
Version: c++-concepts
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: ivo.raisr at oracle dot com
Target Milestone: ---
With gcc 8.0 (SVN revision 252973 from 2017-09-19) we hit internal compiler
error:
Unable to coalesce ssa_names 8 and 28 which are marked as MUST COALESCE.
f_8(ab) and f_28(ab)
during RTL pass: expand
reduced-2.c:12:6: internal compiler error: SSA corruption
void c(void) {
^
0xe23c9c fail_abnormal_edge_coalesce
../../gcc-trunk/gcc/tree-ssa-coalesce.c:1010
0xe23c9c coalesce_partitions
../../gcc-trunk/gcc/tree-ssa-coalesce.c:1399
0xe23c9c coalesce_ssa_name()
../../gcc-trunk/gcc/tree-ssa-coalesce.c:1886
0xdc997b remove_ssa_form
../../gcc-trunk/gcc/tree-outof-ssa.c:948
0xdc997b rewrite_out_of_ssa(ssaexpand*)
../../gcc-trunk/gcc/tree-outof-ssa.c:1172
0x992a90 execute
../../gcc-trunk/gcc/cfgexpand.c:6164
Full pre-processed source is not publicly available but it is the same one as
for bug #82244.
C-reduce was able to produce "reduced-2.bare.c" which I slightly reformatted to
make it more readable: "reduced-2.c". Both files attached.
Reproduce as:
$ gcc -c reduced-2.c -O2
[Bug tree-optimization/82244] -O2: ICE: tree check: expected ssa_name, have integer_cst in replace_uses_by, at tree-cfg.c:1904
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82244 --- Comment #2 from Ivo Raisr --- Reproducible with gcc 8.0 and 7.2. I was not able to reproduce it with 6.3.0.
[Bug tree-optimization/82244] -O2: ICE: tree check: expected ssa_name, have integer_cst in replace_uses_by, at tree-cfg.c:1904
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82244 Ivo Raisr changed: What|Removed |Added CC||ivo.raisr at oracle dot com --- Comment #1 from Ivo Raisr --- Created attachment 42203 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42203=edit reduced.bare.c (output produced by C-reduce)
[Bug tree-optimization/82244] New: -O2: ICE: tree check: expected ssa_name, have integer_cst in replace_uses_by, at tree-cfg.c:1904
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82244
Bug ID: 82244
Summary: -O2: ICE: tree check: expected ssa_name, have
integer_cst in replace_uses_by, at tree-cfg.c:1904
Product: gcc
Version: 8.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: ivo.raisr at oracle dot com
Target Milestone: ---
Created attachment 42202
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42202=edit
reduced.c (reformatted file to make it more readable)
With gcc 8.0 (source snapshot gcc-8-20170827) we hit internal compiler error:
during GIMPLE pass: vrp
reduced.c:10:5: internal compiler error: tree check: expected ssa_name, have
integer_cst in replace_uses_by, at tree-cfg.c:1904
int c(void) {
^
0x78cc94 tree_check_failed(tree_node const*, char const*, int, char const*,
...)
../../gcc-8-20170827/gcc/tree.c:9076
0x6b6046 tree_check(tree_node*, char const*, int, char const*, tree_code)
../../gcc-8-20170827/gcc/tree.h:3098
0x6b6046 replace_uses_by(tree_node*, tree_node*)
../../gcc-8-20170827/gcc/tree-cfg.c:1904
0x7864c2 remove_range_assertions
../../gcc-8-20170827/gcc/tree-vrp.c:7044
0xf27c33 execute_vrp
../../gcc-8-20170827/gcc/tree-vrp.c:11463
0xf27c33 execute
../../gcc-8-20170827/gcc/tree-vrp.c:11547
Full pre-processed sources are not publicly available.
C-reduce was able to produce "reduced.bare.c" which I slightly reformatted to
make it more readable: "reduced.c". Both files attached.
Reproduce as:
$ gcc -c reduced.c -O2
