[Bug lto/54794] New: [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[Bug lto/54794] New: [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[dimhen at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

             Bug #: 54794
           Summary: [4.8 regression] tree code '�F Hc���' is not supported
                    in LTO streams
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: dimhen@gmail.com


err is similar to PR54632, but i cant trigger it with one file. Only while
build library from two dozens files.

tests from PR54632 c#5, c#11 PASS

1. err triggered with '-O1 -ipa-cp' and absent with -O1 only
2. valgrind ... w/o '--param ggc-min-expand=0' say nothing
lto1: internal compiler error: tree code '�F\x18Hc���' is not supported in LTO
streams

3. valgrind ... with '--param ggc-min-expand=0' say
==25840== Invalid read of size 2
==25840==    at 0x763D00: lto_output_tree(output_block*, tree_node*, bool,
bool) (lto-streamer-out.c:127)
==25840==    by 0x9FB800: streamer_write_tree_body(output_block*, tree_node*,
bool) (tree-streamer-out.c:438)
==25840==    by 0x763FE3: lto_output_tree(output_block*, tree_node*, bool,
bool) (lto-streamer-out.c:340)
==25840==    by 0x9FC1EF: streamer_write_tree_body(output_block*, tree_node*,
bool) (tree-streamer-out.c:677)
==25840==    by 0x763FE3: lto_output_tree(output_block*, tree_node*, bool,
bool) (lto-streamer-out.c:340)
==25840==    by 0x9FC153: streamer_write_tree_body(output_block*, tree_node*,
bool) (tree-streamer-out.c:675)
==25840==    by 0x763FE3: lto_output_tree(output_block*, tree_node*, bool,
bool) (lto-streamer-out.c:340)
==25840==    by 0x75EBD5: output_symtab() (lto-cgraph.c:1416)
==25840==    by 0x76545C: lto_output() (lto-streamer-out.c:1018)
==25840==    by 0x79C320: ipa_write_optimization_summaries_1(opt_pass*,
lto_out_decl_state*) (passes.c:2388)
==25840==    by 0x79D18E:
ipa_write_optimization_summaries(lto_symtab_encoder_d*) (passes.c:2435)
==25840==    by 0x4D29E2: lto_main() (lto.c:2637)
==25840==  Address 0x1700000000 is not stack'd, malloc'd or (recently) free'd

$ gdb /usr/local/gcc_current/libexec/gcc/x86_64-unknown-linux-gnu/4.8.0/lto1    
GNU gdb (GDB) Fedora (7.4.50.20120120-50.fc17)
Copyright (C) 2012 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from
/usr/local/gcc_current/libexec/gcc/x86_64-unknown-linux-gnu/4.8.0/lto1...done.
(gdb) r -dumpdir ./ -dumpbase libp.so.wpa -m64 -mtune=generic -march=x86-64
-mtune=generic -march=x86-64 -auxbase RI_EncodeAndSign -fPIC
-fltrans-output-list=libp.so.ltrans.out -fwpa -fresolution=-lm.res
@/tmp/cceC92I4 -O1 -fipa-cp --param ggc-min-expand=0
Starting program:
/usr/local/gcc_current/libexec/gcc/x86_64-unknown-linux-gnu/4.8.0/lto1 -dumpdir
./ -dumpbase libp.so.wpa -m64 -mtune=generic -march=x86-64 -mtune=generic
-march=x86-64 -auxbase RI_EncodeAndSign -fPIC
-fltrans-output-list=libp.so.ltrans.out -fwpa -fresolution=-lm.res
@/tmp/cceC92I4 -O1 -fipa-cp --param ggc-min-expand=0
Reading object files: .libs/RI_EncodeAndSign.o {GC 4346k -> 4139k}
.libs/RI_GenerateKeyPair.o {GC 5388k -> 5388k} .libs/RI_DecodeAndVerify.o {GC
8781k -> 8781k} .libs/CA_CMP_Util.o {GC 11906k -> 11906k}
.libs/CA_CMP_ProcessReq.o {GC 13719k -> 13674k} .libs/CA_CMP_Crl.o {GC 14874k
-> 14697k} .libs/CA_CMP_Root.o {GC 16479k -> 15714k}
.libs/CA_CMP_RevAnnContent.o {GC 17342k -> 16108k} .libs/CA_CMP_Chain.o {GC
17600k -> 17277k} .libs/CA_CMP_Store.o {GC 18251k -> 18089k} .libs/1235.o {GC
18815k -> 18686k} .libs/CA_CMP_CrlGen.o {GC 19664k -> 19185k} .libs/ASN1Blob.o
{GC 19768k -> 19075k} .libs/ASN1Util.o {GC 20769k -> 19831k} .libs/StoreUtil.o
{GC 23895k -> 23014k} .libs/ChainUtil.o {GC 25928k -> 23906k} .libs/EncodeAnd.o
{GC 24933k -> 23866k} .libs/Chain.o {GC 27841k -> 25715k} .libs/ChainBase.o {GC
27401k -> 26761k} .libs/ChainCache.o {GC 27811k -> 27551k}
.libs/ChainCandidate.o {GC 28564k -> 28296k} .libs/UrlRetrieve.o {GC 29382k ->
29020k} .libs/ASN1Types.o {GC 32695k -> 31972k} .libs/ASN1Traits.o {GC 34239k
-> 33693k}
Reading the callgraph
Merging declarations
 {GC 38144k -> 34488k}Reading summaries
 {GC 38236k -> 35369k}Materializing decls:
 <whole-program> {GC 35446k -> 35379k} <profile_estimate> <cp> {GC 35493k ->
35487k} <cdtor> {GC 35505k -> 35502k} {GC 35502k -> 35502k} {GC 35502k ->
35502k} {GC 35502k -> 35502k} {GC 35502k -> 35502k} {GC 35502k -> 35502k} {GC
35502k -> 35502k} <inline> {GC 35502k -> 35502k} {GC 35502k -> 35502k} {GC
35916k -> 35710k} {GC 35710k -> 35710k} {GC 35710k -> 35710k} {GC 35710k ->
35710k} {GC 35710k -> 35710k} <pure-const> {GC 35710k -> 35710k} {GC 35710k ->
35710k} {GC 35710k -> 35710k} {GC 35710k -> 35710k} {GC 35710k -> 35710k} {GC
35710k -> 35710k} <static-var> {GC 35710k -> 35710k} {GC 35710k -> 35710k} {GC
35710k -> 35710k} {GC 35710k -> 35710k} {GC 35710k -> 35710k} {GC 35710k ->
35710k}
Streaming out libp.so.ltrans0.o ( 8537 insns) libp.so.ltrans1.o ( 7341 insns)
libp.so.ltrans2.o ( 7290 insns) libp.so.ltrans3.o ( 6141 insns)
libp.so.ltrans4.o ( 6125 insns) libp.so.ltrans5.o ( 5859 insns)
libp.so.ltrans6.o ( 5835 insns) libp.so.ltrans7.o ( 5778 insns)
libp.so.ltrans8.o ( 5741 insns)
Program received signal SIGSEGV, Segmentation fault.
lto_output_tree (ob=0x1c71880, expr=0x1700000000, ref_p=true, this_ref_p=true)
at /home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:389
389      if (this_ref_p && tree_is_indexable (expr))
(gdb) bt
#0  lto_output_tree (ob=0x1c71880, expr=0x1700000000, ref_p=true,
this_ref_p=true) at /home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:389
#1  0x00000000009fb801 in write_ts_common_tree_pointers (ref_p=true,
expr=0x7ffff420f6e0, ob=0x1c71880) at
/home/dimhen/src/gcc-current/gcc/tree-streamer-out.c:438
#2  streamer_write_tree_body (ob=0x1c71880, expr=0x7ffff420f6e0,
ref_p=<optimized out>) at
/home/dimhen/src/gcc-current/gcc/tree-streamer-out.c:821
#3  0x0000000000763fe4 in lto_write_tree (ref_p=<optimized out>,
expr=0x7ffff420f6e0, ob=0x1c71880) at
/home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:340
#4  lto_output_tree (ob=ob@entry=0x1c71880, expr=0x7ffff420f6e0,
ref_p=ref_p@entry=true, this_ref_p=this_ref_p@entry=true) at
/home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:427
#5  0x00000000009fc1f0 in write_ts_exp_tree_pointers (ref_p=true,
expr=0x7ffff6d4b2d0, ob=0x1c71880) at
/home/dimhen/src/gcc-current/gcc/tree-streamer-out.c:677
#6  streamer_write_tree_body (ob=0x1c71880, expr=0x7ffff6d4b2d0,
ref_p=<optimized out>) at
/home/dimhen/src/gcc-current/gcc/tree-streamer-out.c:860
#7  0x0000000000763fe4 in lto_write_tree (ref_p=<optimized out>,
expr=0x7ffff6d4b2d0, ob=0x1c71880) at
/home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:340
#8  lto_output_tree (ob=0x1c71880, expr=0x7ffff6d4b2d0, ref_p=<optimized out>,
this_ref_p=<optimized out>) at
/home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:427
#9  0x00000000009fc154 in write_ts_exp_tree_pointers (ref_p=true,
expr=0x7ffff4205f80, ob=0x1c71880) at
/home/dimhen/src/gcc-current/gcc/tree-streamer-out.c:675
#10 streamer_write_tree_body (ob=0x1c71880, expr=0x7ffff4205f80,
ref_p=<optimized out>) at
/home/dimhen/src/gcc-current/gcc/tree-streamer-out.c:860
#11 0x0000000000763fe4 in lto_write_tree (ref_p=<optimized out>,
expr=0x7ffff4205f80, ob=0x1c71880) at
/home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:340
#12 lto_output_tree (ob=0x1c71880, expr=0x7ffff4205f80, ref_p=<optimized out>,
this_ref_p=<optimized out>) at
/home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:427
#13 0x000000000075ebd6 in output_node_opt_summary (encoder=0x198d650,
node=0x7ffff62b24a0, ob=0x1c71880) at
/home/dimhen/src/gcc-current/gcc/lto-cgraph.c:1416
#14 output_cgraph_opt_summary () at
/home/dimhen/src/gcc-current/gcc/lto-cgraph.c:1459
#15 output_symtab () at /home/dimhen/src/gcc-current/gcc/lto-cgraph.c:767
#16 0x000000000076545d in lto_output () at
/home/dimhen/src/gcc-current/gcc/lto-streamer-out.c:1018
#17 0x000000000079c321 in ipa_write_optimization_summaries_1 (pass=0x1473c40,
state=state@entry=0x1bf65d0) at /home/dimhen/src/gcc-current/gcc/passes.c:2388
#18 0x000000000079d18f in ipa_write_optimization_summaries (encoder=0x198d650)
at /home/dimhen/src/gcc-current/gcc/passes.c:2435
#19 0x00000000004d29e3 in lto_wpa_write_files () at
/home/dimhen/src/gcc-current/gcc/lto/lto.c:2637
#20 do_whole_program_analysis () at
/home/dimhen/src/gcc-current/gcc/lto/lto.c:3180
#21 lto_main () at /home/dimhen/src/gcc-current/gcc/lto/lto.c:3302
#22 0x000000000088b48c in compile_file () at
/home/dimhen/src/gcc-current/gcc/toplev.c:546
#23 0x000000000088d0aa in do_compile () at
/home/dimhen/src/gcc-current/gcc/toplev.c:1863
#24 toplev_main (argc=44, argv=0x1525c60) at
/home/dimhen/src/gcc-current/gcc/toplev.c:1939
#25 0x0000003536621735 in __libc_start_main (main=0x10638a0 <main(int,
char**)>, argc=21, ubp_av=0x7fffffffddb8, init=<optimized out>, fini=<optimized
out>, rtld_fini=<optimized out>, 
    stack_end=0x7fffffffdda8) at libc-start.c:226
#26 0x00000000004b10e9 in _start ()


$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/local/gcc_current/libexec/gcc/x86_64-unknown-linux-gnu/4.8.0/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: /home/dimhen/src/gcc-current/configure
--prefix=/usr/local/gcc_current --with-multilib-list=m64 --enable-__cxa_atexit
--enable-shared --enable-checking=yes,df,fold,rtl,tree
--enable-gnu-unique-object --enable-linker-build-id
--enable-languages=c,c++,lto --enable-plugin
--enable-version-specific-runtime-libs
Thread model: posix
gcc version 4.8.0 20121003 (experimental) [trunk revision 192029] (GCC)

[Bug lto/54794] [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

Markus Trippelsdorf <markus at trippelsdorf dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |markus at trippelsdorf dot
                   |                            |de

--- Comment #1 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-10-03 13:47:36 UTC ---
Please see:
http://gcc.gnu.org/wiki/A_guide_to_testcase_reduction#Reducing_LTO_bugs

[Bug lto/54794] [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[dimhen at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

--- Comment #2 from Dmitry G. Dyachenko <dimhen at gmail dot com> 2012-10-03 14:04:06 UTC ---
i have 23 src-files hence 23 ii-files with about 50Mb code
if i remove any file from build then err not triggered
(at beginning there were 29 files)

i'll run creduce tonight. Oh, cycle throw 23 files :)

[Bug lto/54794] [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |x86_64-*-*
                 CC|                            |dehao at gcc dot gnu.org,
                   |                            |rguenth at gcc dot gnu.org
   Target Milestone|---                         |4.8.0

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-10-04 12:15:50 UTC ---
Thanks for any help in reducing this (very likely some random BLOCK again) ;)
If you don't end up with a significantly smaller testcase please also attach
all the original 23 ii-files ...

[Bug lto/54794] [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

--- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> 2012-10-04 15:34:04 UTC ---
This may be related to PR 54795.

[Bug lto/54794] [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[dimhen at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

--- Comment #5 from Dmitry G. Dyachenko <dimhen at gmail dot com> 2012-10-07 10:26:14 UTC ---
(In reply to comment #3)
> Thanks for any help in reducing this (very likely some random BLOCK again) ;)
> If you don't end up with a significantly smaller testcase please also attach
> all the original 23 ii-files ...

creduce works fine, but slowly :(
i'll need 2--4 days more to creduce it

may be smb is interesting : r192178 FAIL

[Bug lto/54794] [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[dimhen at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

--- Comment #6 from Dmitry G. Dyachenko <dimhen at gmail dot com> 2012-10-10 15:12:19 UTC ---
creduc'ing it was very slow.
r192289 FAIL
r192295 PASS

probably r192293 fix it

[Bug lto/54794] [4.8 regression] tree code '�F Hc���' is not supported in LTO streams

[dimhen at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54794

Dmitry G. Dyachenko <dimhen at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |FIXED

--- Comment #7 from Dmitry G. Dyachenko <dimhen at gmail dot com> 2012-10-10 15:14:21 UTC ---
r192295 PASS