[Bug bootstrap/97933] New: [11 Regression] Bootstrap failure on s390x-linux

[Bug bootstrap/97933] New: [11 Regression] Bootstrap failure on s390x-linux

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97933

            Bug ID: 97933
           Summary: [11 Regression] Bootstrap failure on s390x-linux
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
          Assignee: unassigned at gcc dot gnu.org
          Reporter: jakub at gcc dot gnu.org
  Target Milestone: ---

../configure --target s390x-linux-gnu --enable-languages=c,c++
--enable-checking=release --disable-multilib --with-system-zlib
--enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object
--enable-linker-build-id --with-gcc-major-version-only
--with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl
--enable-gnu-indirect-function --with-long-double-128 --with-arch=zEC12
--with-tune=z13 --enable-decimal-float
fails both profiledbootstrap and bootstrap (no LTO in either case), while
building stage2 libgcc.  The ICE is e.g. on decDouble.c:
Program received signal SIGSEGV, Segmentation fault.
0x00000000014b4012 in ipa_sra_function_summaries::duplicate(cgraph_node*,
cgraph_node*, isra_func_summary*, isra_func_summary*) ()
Missing separate debuginfos, use: dnf debuginfo-install gmp-6.2.0-5.fc34.s390x
isl-0.16.1-12.fc33.s390x libmpc-1.2.1-1.fc34.s390x libzstd-1.4.5-6.fc34.s390x
mpfr-4.1.0-2.fc33.s390x zlib-1.2.11-23.fc34.s390x
(gdb) up
#1  0x00000000014bec9c in
function_summary<isra_func_summary*>::symtab_duplication(cgraph_node*,
cgraph_node*, void*) ()
(gdb) 
#2  0x00000000012e7a34 in
symbol_table::call_cgraph_duplication_hooks(cgraph_node*, cgraph_node*) ()
(gdb) down
#1  0x00000000014bec9c in
function_summary<isra_func_summary*>::symtab_duplication(cgraph_node*,
cgraph_node*, void*) ()
(gdb) 
#0  0x00000000014b4012 in ipa_sra_function_summaries::duplicate(cgraph_node*,
cgraph_node*, isra_func_summary*, isra_func_summary*) ()
(gdb) 
Bottom (innermost) frame selected; you cannot go down.
(gdb) bt
#0  0x00000000014b4012 in ipa_sra_function_summaries::duplicate(cgraph_node*,
cgraph_node*, isra_func_summary*, isra_func_summary*) ()
#1  0x00000000014bec9c in
function_summary<isra_func_summary*>::symtab_duplication(cgraph_node*,
cgraph_node*, void*) ()
#2  0x00000000012e7a34 in
symbol_table::call_cgraph_duplication_hooks(cgraph_node*, cgraph_node*) ()
#3  0x00000000012fd466 in cgraph_node::create_virtual_clone(vec<cgraph_edge*,
va_heap, vl_ptr>, vec<ipa_replace_map*, va_gc, vl_embed>*,
ipa_param_adjustments*, char const*, unsigned int) ()
#4  0x00000000014b7b90 in (anonymous
namespace)::pass_ipa_sra::execute(function*) ()
#5  0x0000000001631dac in execute_one_pass(opt_pass*) ()
#6  0x0000000001632dce in execute_ipa_pass_list(opt_pass*) ()
#7  0x00000000012f7140 in symbol_table::compile() [clone .part.0] ()
#8  0x00000000012f9610 in symbol_table::finalize_compilation_unit() ()
#9  0x0000000001706ad6 in compile_file() ()
#10 0x00000000011731bc in toplev::main(int, char**) ()
#11 0x00000000011750d8 in main ()
stage1 cc1 doesn't ICE though, nor can reproduce it in a cross-compiler from
x86_64-linux (not even with valgrind annotations under valgrind), so maybe
stage2 is miscompiled.

[Bug bootstrap/97933] [11 Regression] Bootstrap failure on s390x-linux

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97933

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |vmakarov at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r11-5034-g253c415a1acba50711c82693426391743ac18040

[Bug bootstrap/97933] [11 Regression] Bootstrap failure on s390x-linux

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97933

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
   Target Milestone|---                         |11.0

[Bug bootstrap/97933] [11 Regression] Bootstrap failure on s390x-linux

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97933

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 49610
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49610&action=edit
ipa-sra.ii.xz

unxz ipa-sra.ii.xz
./cc1plus -quiet -fno-PIE -O2 -fno-checking -fno-exceptions -fno-rtti
-fasynchronous-unwind-tables -march=zEC12 -mtune=z13 -mlong-double-128
ipa-sra.ii
with cc1plus with that reload change reverted vs. with that change in results
in:
--- ipa-sra.s1  2020-11-22 10:51:59.367579654 +0100
+++ ipa-sra.s2  2020-11-22 10:52:07.166491603 +0100
@@ -4686,13 +4686,13 @@ _ZN26ipa_sra_function_summaries9duplicat
        .cfi_offset 15, -40
        lay     %r15,-176(%r15)
        .cfi_def_cfa_offset 336
-       stg     %r6,168(%r15)
+       stg     %r6,160(%r15)
        risbgn  %r1,%r1,64-1,128+63,2+1
        risbgn  %r2,%r1,58,58+1-1,64-58-1
        stc     %r2,8(%r6)
        tm      8(%r5),16
        jne     .L1172
-       lg      %r1,168(%r15)
+       lg      %r1,160(%r15)
        lgr     %r13,%r5
        ni      8(%r1),239
        ltg     %r8,0(%r5)
@@ -4728,10 +4728,10 @@ _ZN26ipa_sra_function_summaries9duplicat
        slrk    %r2,%r12,%r1
        jnhe    .L1129
 .L1135:
-       st      %r12,164(%r15)
+       st      %r12,172(%r15)
        ltr     %r12,%r12
        lhi     %r1,1
-       stoce   %r1,164(%r15)
+       stoce   %r1,172(%r15)
        lghi    %r6,0
 .L1130:
        llgfr   %r2,%r6
@@ -4801,7 +4801,7 @@ _ZN26ipa_sra_function_summaries9duplicat
        stg     %r2,8(%r1,%r3)
        brct    %r11,.L1160
 .L1136:
-       l       %r1,164(%r15)
+       l       %r1,172(%r15)
        aghi    %r6,1
        brct    %r1,.L1158
        lmg     %r6,%r15,224(%r15)
@@ -4823,8 +4823,7 @@ _ZN26ipa_sra_function_summaries9duplicat
        lg      %r1,8(%r9,%r8)
        j       .L1139
 .L1158:
-       st      %r1,164(%r15)
-       lg      %r1,168(%r15)
+       lg      %r1,160(%r15)
        lg      %r8,0(%r13)
        lg      %r7,0(%r1)
        j       .L1130
@@ -4855,7 +4854,7 @@ _ZN26ipa_sra_function_summaries9duplicat
        brctg   %r3,.L1131
        j       .L1134
 .L1127:
-       lg      %r11,168(%r15)
+       lg      %r11,160(%r15)
        lghi    %r4,1
        lgr     %r2,%r11
        llgfr   %r3,%r12
difference and sadly the ipa-sra.s2 variant crashes.

[Bug bootstrap/97933] [11 Regression] Bootstrap failure on s390x-linux

[vmakarov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97933

--- Comment #3 from Vladimir Makarov <vmakarov at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #1)
> Started with r11-5034-g253c415a1acba50711c82693426391743ac18040

Sorry for causing this error.  It is clearly my mistake. I've started to work
on this.  The fix will be ready tomorrow.

[Bug bootstrap/97933] [11 Regression] Bootstrap failure on s390x-linux

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97933

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Vladimir Makarov <vmakarov@gcc.gnu.org>:

https://gcc.gnu.org/g:bc8f0f1f88d95a284aa329fbc7e70e0b529eaa2a

commit r11-5320-gbc8f0f1f88d95a284aa329fbc7e70e0b529eaa2a
Author: Vladimir N. Makarov <vmakarov@redhat.com>
Date:   Tue Nov 24 11:25:16 2020 -0500

    [PR97933] LRA: find correctly last empty dest block.

    gcc/

    2020-11-24  Vladimir Makarov  <vmakarov@redhat.com>

            PR bootstrap/97933
            * lra.c (lra_process_new_insns): Stop on the first real insn after
            head of e->dest.

[Bug bootstrap/97933] [11 Regression] Bootstrap failure on s390x-linux

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97933

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Confirmed fixed.