Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.bugs.dist > #1266071 > unrolled thread

Bug#1118058: scipy: m68k fails: sparsetools/bsr.h unable to find a register to spill in class ‘ADDR_REGS’

Started byDrew Parsons <dparsons@debian.org>
First post2025-10-14 15:10 +0200
Last post2025-11-05 15:10 +0100
Articles 5 — 4 participants

Back to article view | Back to linux.debian.bugs.dist

Contents

  Bug#1118058: scipy: m68k fails: sparsetools/bsr.h unable to find a register to spill in class ‘ADDR_REGS’ Drew Parsons <dparsons@debian.org> - 2025-10-14 15:10 +0200
    Bug#1118058: scpiy m68k build failure Matthias Klose <doko@debian.org> - 2025-10-29 10:40 +0100
    Bug#1118058: scipy: m68k fails: sparsetools/bsr.h unable to find a register to spill in class ‘ADDR_REGS’ Drew Parsons <dparsons@debian.org> - 2025-11-05 14:40 +0100
      Bug#1118058: scipy: m68k fails: sparsetools/bsr.h unable to find a register to spill in class ‘ADDR_REGS’ John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> - 2025-11-05 14:50 +0100
        Bug#1118058: scipy: m68k fails: sparsetools/bsr.h unable to find a register to spill in class ‘ADDR_REGS’ Drew Parsons <dparsons@emerall.com> - 2025-11-05 15:10 +0100

#1266071 — Bug#1118058: scipy: m68k fails: sparsetools/bsr.h unable to find a register to spill in class ‘ADDR_REGS’

FromDrew Parsons <dparsons@debian.org>
Date2025-10-14 15:10 +0200
SubjectBug#1118058: scipy: m68k fails: sparsetools/bsr.h unable to find a register to spill in class ‘ADDR_REGS’
Message-ID<LFMWR-58Up-1@gated-at.bofh.it>
Source: gcc-15
Version: 15.2.0-5
Severity: normal
Control: affects -1 src:scipy

m68k is failing to build scipy.
Both scipy 1.15.3 and 1.16.2 are affected
https://buildd.debian.org/status/fetch.php?pkg=scipy&arch=m68k&ver=1.15.3-1.1&stamp=1757118702&raw=0
https://buildd.debian.org/status/fetch.php?pkg=scipy&arch=m68k&ver=1.16.2-1exp2&stamp=1760305910&raw=0

The error is coming via cython, but cython itself has not changed,
suggesting the problem might be a regression in gcc-15.
And the error log does contain a GCC DUMP

For 1.16.2 the redacted error message is

[96/1448] cc -Iscipy/lib_fortranobject.a.p -Iscipy -I../scipy -I../../../../usr/lib/m68k-linux-gnu/python3-numpy/numpy/_core/include -I../../../../usr/lib/m68k-linux-gnu/python3-numpy/numpy/f2py/src -I/usr/include/python3.13 -I/usr/include/m68k-linux-gnu/python3.13 -fvisibility=hidden -fdiagnostics-color=always -DNDEBUG -Wall -Winvalid-pch -std=c17 -O3 -Wno-unused-but-set-variable -Wno-unused-function -Wno-conversion -Wno-misleading-indentation -g -O2 -Werror=implicit-function-declaration -ffile-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong -Wformat -Werror=format-security -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -DNPY_NO_DEPRECATED_API=NPY_1_9_API_VERSION -MD -MQ scipy/lib_fortranobject.a.p/.._.._.._.._usr_lib_m68k-linux-gnu_python3-numpy_numpy_f2py_src_fortranobject.c.o -MF scipy/lib_fortranobject.a.p/.._.._.._.._usr_lib_m68k-linux-gnu_python3-numpy_numpy_f2py_src_fortranobject.c.o.d -o scipy/lib_fortranobject.a.p/.._.._.._.._usr_lib_m68k-linux-gnu_python3-numpy_numpy_f2py_src_fortranobject.c.o -c ../../../../usr/lib/m68k-linux-gnu/python3-numpy/numpy/f2py/src/fortranobject.c
...


[886/1448] c++ -Iscipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p -Iscipy/sparse/sparsetools -I../scipy/sparse/sparsetools -I../scipy/_build_utils/src -I../../../../usr/lib/m68k-linux-gnu/python3-numpy/numpy/_core/include -I/usr/include/python3.13 -I/usr/include/m68k-linux-gnu/python3.13 -fvisibility=hidden -fvisibility-inlines-hidden -fdiagnostics-color=always -DNDEBUG -Wall -Winvalid-pch -std=c++17 -O3 -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong -Wformat -Werror=format-security -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -DNPY_NO_DEPRECATED_API=NPY_1_9_API_VERSION -MD -MQ scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o -MF scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o.d -o scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o -c ../scipy/sparse/sparsetools/bsr.cxx
FAILED: scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o 
c++ -Iscipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p -Iscipy/sparse/sparsetools -I../scipy/sparse/sparsetools -I../scipy/_build_utils/src -I../../../../usr/lib/m68k-linux-gnu/python3-numpy/numpy/_core/include -I/usr/include/python3.13 -I/usr/include/m68k-linux-gnu/python3.13 -fvisibility=hidden -fvisibility-inlines-hidden -fdiagnostics-color=always -DNDEBUG -Wall -Winvalid-pch -std=c++17 -O3 -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong -Wformat -Werror=format-security -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -DNPY_NO_DEPRECATED_API=NPY_1_9_API_VERSION -MD -MQ scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o -MF scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o.d -o scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o -c ../scipy/sparse/sparsetools/bsr.cxx
In file included from ../scipy/sparse/sparsetools/bsr.cxx:5:
../scipy/sparse/sparsetools/bsr.h: In function ‘void bsr_binop_bsr(I, I, I, I, const I*, const I*, const T*, const I*, const I*, const T*, I*, I*, T2*, const bin_op&) [with I = long long int; T = complex_wrapper<float, npy_cfloat>; T2 = complex_wrapper<float, npy_cfloat>; bin_op = std::multiplies<complex_wrapper<float, npy_cfloat> >]’:
../scipy/sparse/sparsetools/bsr.h:591:1: error: unable to find a register to spill in class ‘ADDR_REGS’
  591 | }
      | ^
../scipy/sparse/sparsetools/bsr.h:591:1: error: this is the insn:
(insn 848 847 3788 54 (set (mem:DI (plus:SI (mult:SI (subreg:SI (reg/v:DI 744 [orig:137 nnz ] [137]) 4)
                    (const_int 8 [0x8]))
                (reg/v/f:SI 392 [ Cj ])) [2 *_176+0 S8 A16])
        (mem:DI (post_inc:SI (reg:SI 6 %d6 [orig:63 ivtmp.100551 ] [63])) [2 MEM[(const long long int *)_814]+0 S8 A16])) "../scipy/sparse/sparsetools/bsr.h":512:25 75 {*m68k.md:1569}
     (expr_list:REG_INC (reg:SI 6 %d6 [orig:63 ivtmp.100551 ] [63])
        (nil)))
../scipy/sparse/sparsetools/bsr.h:591: confused by earlier errors, bailing out
Preprocessed source stored into /tmp/ccduDjZh.out file, please attach this to your bugreport.
=== BEGIN GCC DUMP ===
4017446: // Target: m68k-linux-gnu
4017446: // Configured with: ../src/configure -v --with-pkgversion='Debian 15.2.0-4' --with-bugurl=file:///usr/share/doc/gcc-15/README.Bugs --enable-languages=c,ada,c++,fortran,objc,obj-c++,m2,algol68 --prefix=/usr --with-gcc-major-version-only --program-suffix=-15 --program-prefix=m68k-linux-gnu- --enable-shared --enable-linker-build-id --libexecdir=/usr/libexec --without-included-gettext --enable-threads=posix --libdir=/usr/lib --enable-nls --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new --enable-libstdcxx-backtrace --enable-gnu-unique-object --disable-libssp --disable-libitm --disable-libsanitizer --disable-libquadmath --disable-libquadmath-support --enable-plugin --with-system-zlib --enable-objc-gc=auto --enable-multiarch --disable-werror --disable-werror --disable-multilib --enable-checking=release --build=m68k-linux-gnu --host=m68k-linux-gnu --target=m68k-linux-gnu
4017446: // Thread model: posix
4017446: // Supported LTO compression algorithms: zlib zstd
4017446: // gcc version 15.2.0 (Debian 15.2.0-4) 
4017446: // 
4017446: // In file included from ../scipy/sparse/sparsetools/bsr.cxx:5:
4017446: // ../scipy/sparse/sparsetools/bsr.h: In function ‘void bsr_binop_bsr(I, I, I, I, const I*, const I*, const T*, const I*, const I*, const T*, I*, I*, T2*, const bin_op&) [with I = long long int; T = complex_wrapper<float, npy_cfloat>; T2 = complex_wrapper<float, npy_cfloat>; bin_op = std::multiplies<complex_wrapper<float, npy_cfloat> >]’:
4017446: // ../scipy/sparse/sparsetools/bsr.h:591:1: error: unable to find a register to spill in class ‘ADDR_REGS’
4017446: //   591 | }
4017446: //       | ^
4017446: // ../scipy/sparse/sparsetools/bsr.h:591:1: error: this is the insn:
4017446: // (insn 848 847 3788 54 (set (mem:DI (plus:SI (mult:SI (subreg:SI (reg/v:DI 744 [orig:137 nnz ] [137]) 4)
4017446: //                     (const_int 8 [0x8]))
4017446: //                 (reg/v/f:SI 392 [ Cj ])) [2 *_176+0 S8 A16])
4017446: //         (mem:DI (post_inc:SI (reg:SI 6 %d6 [orig:63 ivtmp.100551 ] [63])) [2 MEM[(const long long int *)_814]+0 S8 A16])) "../scipy/sparse/sparsetools/bsr.h":512:25 75 {*m68k.md:1569}
4017446: //      (expr_list:REG_INC (reg:SI 6 %d6 [orig:63 ivtmp.100551 ] [63])
4017446: //         (nil)))
4017446: // ../scipy/sparse/sparsetools/bsr.h:591: confused by earlier errors, bailing out
4017446: 
4017446: // /usr/libexec/gcc/m68k-linux-gnu/15/cc1plus -quiet -I scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p -I scipy/sparse/sparsetools -I ../scipy/sparse/sparsetools -I ../scipy/_build_utils/src -I ../../../../usr/lib/m68k-linux-gnu/python3-numpy/numpy/_core/include -I /usr/include/python3.13 -I /usr/include/m68k-linux-gnu/python3.13 -imultilib . -imultiarch m68k-linux-gnu -MD scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.d -MF scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o.d -MQ scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/bsr.cxx.o -D_GNU_SOURCE -D NDEBUG -D _LARGEFILE_SOURCE -D _FILE_OFFSET_BITS=64 -D _TIME_BITS=64 -D _FORTIFY_SOURCE=2 -D NPY_NO_DEPRECATED_API=NPY_1_9_API_VERSION ../scipy/sparse/sparsetools/bsr.cxx -quiet -dumpdir scipy/sparse/sparsetools/_sparsetools.cpython-313-m68k-linux-gnu.so.p/ -dumpbase bsr.cxx.cxx -dumpbase-ext .cxx -mcpu=68020 -g -O3 -O2 -Wall -Winvalid-pch -Wformat=1 -Werror=format-security -Wdate-time -std=c++17 -fdiagnostics-color=always -fvisibility=hidden -fvisibility-inlines-hidden -ffile-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong -fPIC -o - -frandom-seed=0 -fdump-noaddr

[toc] | [next] | [standalone]

#1268230 — Bug#1118058: scpiy m68k build failure

FromMatthias Klose <doko@debian.org>
Date2025-10-29 10:40 +0100
SubjectBug#1118058: scpiy m68k build failure
Message-ID<LLaOR-8Rb3-3@gated-at.bofh.it>
In reply to#1266071
Control: tags -1 + moreinfo

please could you attach the complete preprocessed file?  bonus points 
for reducing it to a test case.

[toc] | [prev] | [next] | [standalone]

#1269115

FromDrew Parsons <dparsons@debian.org>
Date2025-11-05 14:40 +0100
Message-ID<LNLTX-aEHq-1@gated-at.bofh.it>
In reply to#1266071
Package: gcc-15
Version: 15.2.0-7
Followup-For: Bug #1118058
X-Debbugs-Cc: debian-68k@lists.debian.org
User: debian-68k@lists.debian.org
Usertags: m68k
Control: affects -1 src:hypre

hypre is also affected by a gcc dump on m68k

https://buildd.debian.org/status/fetch.php?pkg=hypre&arch=m68k&ver=2.33.0-1exp1&stamp=1762349220&raw=0

2399190: // csr_matrix.c: In function ‘hypre_CSRMatrixRead’:
2399190: // csr_matrix.c:520:1: error: insn does not satisfy its constraints:
2399190: //   520 | }
2399190: //       | ^
2399190: // (insn 70 244 71 4 (set (mem:SI (reg:SI 8 %a0) [3 MEM[(HYPRE_Int *)_73 + 4294967292B]+0 S4 A16])
2399190: //         (plus:SI (mem:SI (reg:SI 9 %a1) [3 MEM[(HYPRE_Int *)_6]+0 S4 A16])
2399190: //             (const_int -1 [0xffffffffffffffff]))) "csr_matrix.c":486:19 150 {*addsi3_internal}
2399190: //      (nil))
2399190: // during RTL pass: postreload
2399190: // csr_matrix.c:520:1: internal compiler error: in extract_constrain_insn, at recog.cc:2783


True, in this case it's not reporting a missing register ADDR_REGS.

[toc] | [prev] | [next] | [standalone]

#1269118

FromJohn Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Date2025-11-05 14:50 +0100
Message-ID<LNM3D-aELk-3@gated-at.bofh.it>
In reply to#1269115
On Wed, 2025-11-05 at 14:33 +0100, Drew Parsons wrote:
> Package: gcc-15
> Version: 15.2.0-7
> Followup-For: Bug #1118058
> X-Debbugs-Cc: debian-68k@lists.debian.org
> User: debian-68k@lists.debian.org
> Usertags: m68k
> Control: affects -1 src:hypre
> 
> hypre is also affected by a gcc dump on m68k
> 
> https://buildd.debian.org/status/fetch.php?pkg=hypre&arch=m68k&ver=2.33.0-1exp1&stamp=1762349220&raw=0
> 
> 2399190: // csr_matrix.c: In function ‘hypre_CSRMatrixRead’:
> 2399190: // csr_matrix.c:520:1: error: insn does not satisfy its constraints:
> 2399190: //   520 | }
> 2399190: //       | ^
> 2399190: // (insn 70 244 71 4 (set (mem:SI (reg:SI 8 %a0) [3 MEM[(HYPRE_Int *)_73 + 4294967292B]+0 S4 A16])
> 2399190: //         (plus:SI (mem:SI (reg:SI 9 %a1) [3 MEM[(HYPRE_Int *)_6]+0 S4 A16])
> 2399190: //             (const_int -1 [0xffffffffffffffff]))) "csr_matrix.c":486:19 150 {*addsi3_internal}
> 2399190: //      (nil))
> 2399190: // during RTL pass: postreload
> 2399190: // csr_matrix.c:520:1: internal compiler error: in extract_constrain_insn, at recog.cc:2783
> 
> 
> True, in this case it's not reporting a missing register ADDR_REGS.

It's a different bug, see: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122438

The bug report for the scipy issue is: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122479

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

[toc] | [prev] | [next] | [standalone]

#1269120

FromDrew Parsons <dparsons@emerall.com>
Date2025-11-05 15:10 +0100
Message-ID<LNMdj-aEPd-11@gated-at.bofh.it>
In reply to#1269118
On 2025-11-05 14:47, John Paul Adrian Glaubitz wrote:
> On Wed, 2025-11-05 at 14:33 +0100, Drew Parsons wrote:
>> Package: gcc-15
>> Version: 15.2.0-7
>> Followup-For: Bug #1118058
>> 
>> hypre is also affected by a gcc dump on m68k
>> 
>> True, in this case it's not reporting a missing register ADDR_REGS.
> 
> It's a different bug, see: 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122438
> 
> The bug report for the scipy issue is: 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122479

Thanks for clarifying, Adrian.
Great to see they're both getting fixed.

Drew

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.bugs.dist

csiph-web