Bug #1988710 “qemu FTBFS with GCC 12” : Bugs : qemu package : Ubuntu

Paride Legovini (paride) on 2022-09-05

Changed in qemu (Ubuntu):
importance:	Undecided → High

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-05:

#1

Hmm,
0x48002002 < 0xFFFFFFFF

Is that a false positive or am I looking at the wrong target for this warning.
There are various reports about this and so far most I found where different kinds of false positives.
- https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101977
- https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104657
- https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99578

If you consider this a false positive at the end of your check consider for now doing a wrapper with just this warning disabled on those two assignments:

#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Warray-bounds"
code
#pragma GCC diagnostic pop

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-05:

#2

I agree it looks like a false positive, I'll try that pragma and comment back.

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-05 (last edit on 2022-09-05):

#3

Download full text (3.2 KiB)

I added those pragmas via a quilt patch, but I'm unsure if and how the patch should be submitted upstream. I was thinking about opening an openbios bug (as the patched code come from https://github.com/openbios/, brought into the qemu source tree as a git submodule), but apparently Debian can compile qemu 7.0+dfsg-7 with GCC 12, see:

https://buildd.debian.org/status/fetch.php?pkg=qemu&arch=amd64&ver=1%3A7.0%2Bdfsg-7%2Bb1&stamp=1660766914&raw=0

so I'm not sure we have a solid case for an upstream bug.

In any case I tried recompiling with those pragmas but qemu still ftbfs, with a different error this time. Again it only failed on amd64.

PPA: https://launchpad.net/~paride/+archive/ubuntu/qemu-lp1921664/+packages

Full amd64 build log: https://launchpadlibrarian.net/621847743/buildlog_ubuntu-kinetic-amd64.qemu_1%3A7.0+dfsg-7ubuntu2~ppa2_BUILDING.txt.gz

Log excerpt:

console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
130 | vga[0] = 'H' + attr;
| ~~~~~~~^~~~~~~~~~~~
0x137917b internal_error(char const*, ...)
???:0
0x5a9326 fancy_abort(char const*, int, char const*)
???:0
0xe0a692 alpha_split_const_mov(machine_mode, rtx_def**)
???:0
0xe0a801 alpha_expand_mov(machine_mode, rtx_def**)
???:0
0x112e2dd gen_movv4hi(rtx_def*, rtx_def*)
???:0
0x7eedeb emit_move_insn_1(rtx_def*, rtx_def*)
???:0
0x7ef1d7 emit_move_insn(rtx_def*, rtx_def*)
???:0
0xe0d016 alpha_expand_movmisalign(machine_mode, rtx_def**)
???:0
0x112e5aa gen_movmisalignv4hi(rtx_def*, rtx_def*)
???:0
0xa278fc expand_insn(insn_code, unsigned int, expand_operand*)
???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"' -c -o console-low.o console-low.S
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o ps2port.o ps2port.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o pci.o pci.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o vgaio.o vgaio.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o vgatables.o vgatables.c
Preprocessed source stored into /tmp/cchX2bbT.out file, please attach this to your bugreport.
=== BEGIN GCC DUMP ===

[long dump]

=== END GCC DUMP ===
make[3]: *** [<builtin>: console.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[3]: Leaving directory '/<<PKGBUILDDIR>>/b/qemu-palcode'
make[2]: *** [debian/rules:560: b/qemu-palcode/palcode-clipper] Error 2
make[2]: Leaving directory '/<<PKGBUILDDIR>>'
make[1]: *** [debian/rules:657: i...

I added those pragmas via a quilt patch, but I'm unsure if and how the patch should be submitted upstream. I was thinking about opening an openbios bug (as the patched code come from https://github.com/openbios/, brought into the qemu source tree as a git submodule), but apparently Debian can compile qemu 7.0+dfsg-7 with GCC 12, see:

https://buildd.debian.org/status/fetch.php?pkg=qemu&arch=amd64&ver=1%3A7.0%2Bdfsg-7%2Bb1&stamp=1660766914&raw=0

so I'm not sure we have a solid case for an upstream bug.

In any case I tried recompiling with those pragmas but qemu still ftbfs, with a different error this time. Again it only failed on amd64.

PPA: https://launchpad.net/~paride/+archive/ubuntu/qemu-lp1921664/+packages

Full amd64 build log: https://launchpadlibrarian.net/621847743/buildlog_ubuntu-kinetic-amd64.qemu_1%3A7.0+dfsg-7ubuntu2~ppa2_BUILDING.txt.gz

Log excerpt:

console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
  130 |     vga[0] = 'H' + attr;
      |     ~~~~~~~^~~~~~~~~~~~
0x137917b internal_error(char const*, ...)
	???:0
0x5a9326 fancy_abort(char const*, int, char const*)
	???:0
0xe0a692 alpha_split_const_mov(machine_mode, rtx_def**)
	???:0
0xe0a801 alpha_expand_mov(machine_mode, rtx_def**)
	???:0
0x112e2dd gen_movv4hi(rtx_def*, rtx_def*)
	???:0
0x7eedeb emit_move_insn_1(rtx_def*, rtx_def*)
	???:0
0x7ef1d7 emit_move_insn(rtx_def*, rtx_def*)
	???:0
0xe0d016 alpha_expand_movmisalign(machine_mode, rtx_def**)
	???:0
0x112e5aa gen_movmisalignv4hi(rtx_def*, rtx_def*)
	???:0
0xa278fc expand_insn(insn_code, unsigned int, expand_operand*)
	???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"'  -c -o console-low.o console-low.S
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o ps2port.o ps2port.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o pci.o pci.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o vgaio.o vgaio.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o vgatables.o vgatables.c
Preprocessed source stored into /tmp/cchX2bbT.out file, please attach this to your bugreport.
=== BEGIN GCC DUMP ===

[long dump]

=== END GCC DUMP ===
make[3]: *** [<builtin>: console.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[3]: Leaving directory '/<<PKGBUILDDIR>>/b/qemu-palcode'
make[2]: *** [debian/rules:560: b/qemu-palcode/palcode-clipper] Error 2
make[2]: Leaving directory '/<<PKGBUILDDIR>>'
make[1]: *** [debian/rules:657: install] Error 2
make[1]: Leaving directory '/<<PKGBUILDDIR>>'
make: *** [debian/rules:657: binary] Error 2

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-07:

#4

Since in the latter FTBFS the compiler crashes you might want to get Doko involved?

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-07:

#5

+1, I see you already subscribed him, thanks. I just retriggered the PPA build, let's see if crashes consistently:

https://launchpad.net/~paride/+archive/ubuntu/qemu-lp1921664/+build/24337146

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-07:

#6

FTBFS, same compiler crash:

https://launchpadlibrarian.net/622189783/buildlog_ubuntu-kinetic-amd64.qemu_1%3A7.0+dfsg-7ubuntu2~ppa2_BUILDING.txt.gz

Paride Legovini (paride) on 2022-09-07

tags:

added: rls-kk-incoming

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-07:

#7

Added tags and tasks to get a triage round from Foundations.

summary:

- FTBFS with GCC 12
+ qemu FTBFS with GCC 12

Matthieu Clemenceau (mclemenceau) on 2022-09-08

tags:

added: fr-2662

Revision history for this message

Brian Murray (brian-murray) wrote on 2022-09-08:

#8

During the Foundations team meeting schopin indicated he had fixed a similar issue in lxc so you might want to look at the changes in lxc version 1:5.0.0~git2209-g5a7b9ce67-0ubuntu3 and ubuntu2.

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-09:

#9

By looking at that diff [1] looks like it's meant to fix an array-bounds warning, which is the issue described in this bug's description. However we I managed to silence that warning in qemu by using the pragmas Christian suggested, but then the package FTBFS with a compiler crash which looks unrelated (see comment 3 of this bug).

[1] http://launchpadlibrarian.net/622217429/lxc_1%3A5.0.0~git2209-g5a7b9ce67-0ubuntu2_1%3A5.0.0~git2209-g5a7b9ce67-0ubuntu3.diff.gz

Brian Murray (brian-murray) on 2022-09-15

tags:

added: foundations-todo
removed: rls-kk-incoming

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-15:

#10

The latest FTBFS is with gcc-12_12.2.0-1ubuntu1, I see that now we have 12.2.0-2ubuntu1 in the archive, and it has a bunch of fixes. I'm retrying a build with it.

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-15:

#11

Download full text (20.9 KiB)

Same compile crash, see:

https://launchpadlibrarian.net/623533933/buildlog_ubuntu-kinetic-amd64.qemu_1%3A7.0+dfsg-7ubuntu2~ppa2_BUILDING.txt.gz

Build environment shows: gcc-12_12.2.0-2ubuntu1

Excerpt:

[17/17] Generating bios.bin with a custom command
cp -al roms/qemu-palcode b/
/usr/bin/make -C /<<PKGBUILDDIR>>/b/qemu-palcode CROSS=alpha-linux-gnu-
make[3]: Entering directory '/<<PKGBUILDDIR>>/b/qemu-palcode'
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"' -c -o pal.o pal.S
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"' -c -o sys-clipper.o sys-clipper.S
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o init.o init.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o crb.o crb.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o uart.o uart.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
130 | vga[0] = 'H' + attr;
| ~~~~~~~^~~~~~~~~~~~
0x137917b internal_error(char const*, ...)
???:0
0x5a9326 fancy_abort(char const*, int, char const*)
???:0
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"' -c -o console-low.o console-low.S
0xe0a692 alpha_split_const_mov(machine_mode, rtx_def**)
???:0
0xe0a801 alpha_expand_mov(machine_mode, rtx_def**)
???:0
0x112e2dd gen_movv4hi(rtx_def*, rtx_def*)
???:0
0x7eedeb emit_move_insn_1(rtx_def*, rtx_def*)
???:0
0x7ef1d7 emit_move_insn(rtx_def*, rtx_def*)
???:0
0xe0d016 alpha_expand_movmisalign(machine_mode, rtx_def**)
???:0
0x112e5aa gen_movmisalignv4hi(rtx_def*, rtx_def*)
???:0
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o ps2port.o ps2port.c
0xa278fc expand_insn(insn_code, unsigned int, expand_operand*)
???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o pci.o pci.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o vgaio.o vgaio.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-...

Same compile crash, see:

https://launchpadlibrarian.net/623533933/buildlog_ubuntu-kinetic-amd64.qemu_1%3A7.0+dfsg-7ubuntu2~ppa2_BUILDING.txt.gz

Build environment shows: gcc-12_12.2.0-2ubuntu1

Excerpt:

[17/17] Generating bios.bin with a custom command
cp -al roms/qemu-palcode b/
/usr/bin/make -C /<<PKGBUILDDIR>>/b/qemu-palcode CROSS=alpha-linux-gnu-
make[3]: Entering directory '/<<PKGBUILDDIR>>/b/qemu-palcode'
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"'  -c -o pal.o pal.S
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"'  -c -o sys-clipper.o sys-clipper.S
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o init.o init.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o crb.o crb.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o uart.o uart.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
  130 |     vga[0] = 'H' + attr;
      |     ~~~~~~~^~~~~~~~~~~~
0x137917b internal_error(char const*, ...)
	???:0
0x5a9326 fancy_abort(char const*, int, char const*)
	???:0
alpha-linux-gnu-gcc -Wa,-m21264 -Wa,--noexecstack -DSYSTEM_H='"sys-clipper.h"'  -c -o console-low.o console-low.S
0xe0a692 alpha_split_const_mov(machine_mode, rtx_def**)
	???:0
0xe0a801 alpha_expand_mov(machine_mode, rtx_def**)
	???:0
0x112e2dd gen_movv4hi(rtx_def*, rtx_def*)
	???:0
0x7eedeb emit_move_insn_1(rtx_def*, rtx_def*)
	???:0
0x7ef1d7 emit_move_insn(rtx_def*, rtx_def*)
	???:0
0xe0d016 alpha_expand_movmisalign(machine_mode, rtx_def**)
	???:0
0x112e5aa gen_movmisalignv4hi(rtx_def*, rtx_def*)
	???:0
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o ps2port.o ps2port.c
0xa278fc expand_insn(insn_code, unsigned int, expand_operand*)
	???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o pci.o pci.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o vgaio.o vgaio.c
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o vgatables.o vgatables.c
Preprocessed source stored into /tmp/ccLfxzvp.out file, please attach this to your bugreport.
=== BEGIN GCC DUMP ===
146766: // Target: alpha-linux-gnu
146766: // Configured with: ../src/configure -v --with-pkgversion='Ubuntu 12.2.0-2ubuntu1' --with-bugurl=file:///usr/share/doc/gcc-12/README.Bugs --enable-languages=c,ada,c++,go,fortran,objc,obj-c++,m2 --prefix=/usr --with-gcc-major-version-only --program-suffix=-12 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new --enable-gnu-unique-object --disable-libssp --disable-libsanitizer --disable-libquadmath --disable-libquadmath-support --enable-plugin --with-system-zlib --enable-multiarch --disable-werror --with-long-double-128 --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=alpha-linux-gnu --program-prefix=alpha-linux-gnu- --includedir=/usr/alpha-linux-gnu/include
146766: // Thread model: posix
146766: // Supported LTO compression algorithms: zlib zstd
146766: // gcc version 12.2.0 (Ubuntu 12.2.0-2ubuntu1) 
146766: // 
146766: // during RTL pass: expand
146766: // console.c: In function ‘do_console’:
146766: // console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
146766: //   130 |     vga[0] = 'H' + attr;
146766: //       |     ~~~~~~~^~~~~~~~~~~~
146766: // 0x137917b internal_error(char const*, ...)
146766: // 	???:0
146766: // 0x5a9326 fancy_abort(char const*, int, char const*)
146766: // 	???:0
146766: // 0xe0a692 alpha_split_const_mov(machine_mode, rtx_def**)
146766: // 	???:0
146766: // 0xe0a801 alpha_expand_mov(machine_mode, rtx_def**)
146766: // 	???:0
146766: // 0x112e2dd gen_movv4hi(rtx_def*, rtx_def*)
146766: // 	???:0
146766: // 0x7eedeb emit_move_insn_1(rtx_def*, rtx_def*)
146766: // 	???:0
146766: // 0x7ef1d7 emit_move_insn(rtx_def*, rtx_def*)
146766: // 	???:0
146766: // 0xe0d016 alpha_expand_movmisalign(machine_mode, rtx_def**)
146766: // 	???:0
146766: // 0x112e5aa gen_movmisalignv4hi(rtx_def*, rtx_def*)
146766: // 	???:0
146766: // 0xa278fc expand_insn(insn_code, unsigned int, expand_operand*)
146766: // 	???:0
146766: // Please submit a full bug report, with preprocessed source (by using -freport-bug).
146766: // Please include the complete backtrace with any bug report.
146766: // See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
146766: 
146766: // /usr/lib/gcc-cross/alpha-linux-gnu/12/cc1 -quiet -imultilib . -imultiarch alpha-linux-gnu -D SYSTEM_H="sys-clipper.h" console.c -quiet -dumpbase console.c -dumpbase-ext .c -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -g1 -O2 -Wall -fvisibility=hidden -fno-strict-aliasing -o - -frandom-seed=0 -fdump-noaddr
146766: 
146766: # 0 "console.c"
146766: # 1 "/<<PKGBUILDDIR>>/b/qemu-palcode//"
146766: # 0 "<built-in>"
146766: # 0 "<command-line>"
146766: # 1 "/usr/include/stdc-predef.h" 1 3 4
146766: # 0 "<command-line>" 2
146766: # 1 "console.c"
146766: # 21 "console.c"
146766: # 1 "protos.h" 1
146766: # 27 "protos.h"
146766: typedef unsigned char uint8_t;
146766: typedef unsigned short uint16_t;
146766: typedef unsigned int uint32_t;
146766: typedef unsigned long uint64_t;
146766: typedef unsigned long size_t;
146766: 
146766: 
146766: 
146766: 
146766: 
146766: 
146766: 
146766: typedef __builtin_va_list va_list;
146766: 
146766: 
146766: 
146766: 
146766: 
146766: 
146766: extern void *memset(void *, int, size_t);
146766: extern void *memcpy(void *, const void *, size_t);
146766: extern size_t strlen(const char *);
146766: 
146766: 
146766: 
146766: 
146766: 
146766: static inline void wrent(void *cb, unsigned long which)
146766: {
146766:   register void *a0 __asm__("$16") = cb;
146766:   register unsigned long a1 __asm__("$17") = which;
146766: 
146766:   asm volatile ("call_pal 0x34"
146766:   : "+r"(a0), "+r"(a1)
146766:   : : "$1", "$22", "$23", "$24", "$25");
146766: }
146766: 
146766: static inline unsigned long swpipl(unsigned long newipl)
146766: {
146766:   register unsigned long v0 __asm__("$0");
146766:   register unsigned long a0 __asm__("$16") = newipl;
146766: 
146766:   asm volatile ("call_pal 0x35"
146766:   : "=r"(v0), "+r"(a0)
146766:   : : "$1", "$22", "$23", "$24", "$25");
146766: 
146766:   return v0;
146766: }
146766: 
146766: static inline unsigned long rdps(void)
146766: {
146766:   register unsigned long v0 __asm__("$0");
146766: 
146766:   asm volatile ("call_pal 0x36"
146766:   : "=r"(v0) : : "$1", "$22", "$23", "$24", "$25");
146766: 
146766:   return v0;
146766: }
146766: 
146766: static inline void wrkgp(void)
146766: {
146766:   asm volatile ("mov $29, $16\n\tcall_pal 0x37"
146766:   : : : "$16", "$1", "$22", "$23", "$24", "$25");
146766: }
146766: 
146766: static inline unsigned long wtint(unsigned long skip)
146766: {
146766:   register unsigned long v0 __asm__("$0");
146766:   register unsigned long a0 __asm__("$16") = skip;
146766: 
146766:   asm volatile ("call_pal 0x3e"
146766:   : "=r"(v0), "+r"(a0)
146766:   : : "$1", "$22", "$23", "$24", "$25");
146766: 
146766:   return v0;
146766: }
146766: 
146766: 
146766: 
146766: 
146766: 
146766: static inline unsigned long ldq_p(unsigned long addr)
146766: {
146766:   register unsigned long v0 __asm__("$0");
146766:   register unsigned long a0 __asm__("$16") = 1;
146766:   register unsigned long a1 __asm__("$17") = addr;
146766: 
146766:   asm volatile ("call_pal 9"
146766:   : "=r"(v0), "+r"(a0), "+r"(a1) :
146766:   : "$18", "$19", "$20", "$21");
146766: 
146766:   return v0;
146766: }
146766: 
146766: static inline unsigned long stq_p(unsigned long port, unsigned long val)
146766: {
146766:   register unsigned long v0 __asm__("$0");
146766:   register unsigned long a0 __asm__("$16") = 2;
146766:   register unsigned long a1 __asm__("$17") = port;
146766:   register unsigned long a2 __asm__("$18") = val;
146766: 
146766:   asm volatile ("call_pal 9"
146766:   : "=r"(v0), "+r"(a0), "+r"(a1), "+r"(a2) :
146766:   : "$19", "$20", "$21");
146766: 
146766:   return v0;
146766: }
146766: 
146766: static inline unsigned long get_wall_time(void)
146766: {
146766:   register unsigned long v0 __asm__("$0");
146766:   register unsigned long a0 __asm__("$16") = 3;
146766: 
146766:   asm("call_pal 9" : "=r"(v0), "+r"(a0) : : "$17", "$18", "$19", "$20", "$21");
146766: 
146766:   return v0;
146766: }
146766: 
146766: static inline unsigned long get_alarm(void)
146766: {
146766:   register unsigned long v0 __asm__("$0");
146766:   register unsigned long a0 __asm__("$16") = 4;
146766: 
146766:   asm("call_pal 9" : "=r"(v0), "+r"(a0) : : "$17", "$18", "$19", "$20", "$21");
146766: 
146766:   return v0;
146766: }
146766: 
146766: static inline void set_alarm_rel(unsigned long nsec)
146766: {
146766:   register unsigned long a0 __asm__("$16") = 5;
146766:   register unsigned long a1 __asm__("$17") = nsec;
146766: 
146766:   asm volatile ("call_pal 9"
146766:   : "+r"(a0), "+r"(a1)
146766:   : : "$0", "$18", "$19", "$20", "$21");
146766: }
146766: 
146766: static inline void set_alarm_abs(unsigned long nsec)
146766: {
146766:   register unsigned long a0 __asm__("$16") = 6;
146766:   register unsigned long a1 __asm__("$17") = nsec;
146766: 
146766:   asm volatile ("call_pal 9"
146766:   : "+r"(a0), "+r"(a1)
146766:   : : "$0", "$18", "$19", "$20", "$21");
146766: }
146766: 
146766: 
146766: 
146766: 
146766: 
146766: extern void *pci_io_base;
146766: extern void *pci_mem_base;
146766: 
146766: static inline uint8_t inb(unsigned long port)
146766: {
146766:   return *(volatile uint8_t *)(pci_io_base + port);
146766: }
146766: 
146766: static inline uint16_t inw(unsigned long port)
146766: {
146766:   return *(volatile uint16_t *)(pci_io_base + port);
146766: }
146766: 
146766: static inline uint32_t inl(unsigned long port)
146766: {
146766:   return *(volatile uint32_t *)(pci_io_base + port);
146766: }
146766: 
146766: static inline void outb(uint8_t val, unsigned long port)
146766: {
146766:   *(volatile uint8_t *)(pci_io_base + port) = val;
146766: }
146766: 
146766: static inline void outw(uint16_t val, unsigned long port)
146766: {
146766:   *(volatialpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o vgafonts.o vgafonts.c
le uint16_t *)(pci_io_base + port) = val;
146766: }
146766: 
146766: static inline void outl(uint32_t val, unsigned long port)
146766: {
146766:   *(volatile uint32_t *)(pci_io_base + port) = val;
146766: }
146766: 
146766: 
146766: 
146766: 
146766: 
146766: extern unsigned long crb_dispatch(long select, long a1, long a2,
146766:                                   long a3, long a4);
146766: extern unsigned long crb_fixup(unsigned long vptptr, unsigned long hwrpb);
146766: 
146766: 
146766: 
146766: 
146766: 
146766: extern _Bool have_vga;
146766: extern unsigned int pci_vga_bus;
146766: extern unsigned int pci_vga_dev;
146766: 
146766: extern void do_console(void);
146766: extern void entInt(void);
146766: 
146766: 
146766: 
146766: 
146766: 
146766: extern int printf(const char *, ...);
146766: extern void ndelay(unsigned long nsec);
146766: 
146766: static inline void udelay(unsigned long msec)
146766: {
146766:   ndelay(msec * 1000);
146766: }
146766: 
146766: 
146766: 
146766: 
146766: extern void ps2port_setup(void);
146766: extern void pci_setup(void);
146766: extern void vgahw_init(void);
146766: # 22 "console.c" 2
146766: # 1 "console.h" 1
146766: # 44 "console.h"
146766: extern unsigned long crb_getc(long unit);
146766: extern unsigned long crb_process_keycode(long unit, long keycode, long again);
146766: extern unsigned long crb_puts(long unit, const char *buf, unsigned long length);
146766: extern unsigned long crb_reset_term(long unit);
146766: 
146766: extern unsigned long crb_open(const char *devstr, unsigned long length);
146766: extern unsigned long crb_close(long channel);
146766: extern unsigned long crb_read(long channel, unsigned long length,
146766:                               char *buf, unsigned long block);
146766: extern unsigned long crb_write(long channel, unsigned long length,
146766:                                const char *buf, unsigned long block);
146766: 
146766: extern unsigned long crb_get_env(unsigned long id, char *buf,
146766:                                  unsigned long length);
146766: extern unsigned long crb_set_env(unsigned long id, const char *buf,
146766:                                  unsigned long length);
146766: # 23 "console.c" 2
146766: # 1 "vgatables.h" 1
146766: # 13 "vgatables.h"
146766: typedef uint8_t u8;
146766: typedef uint16_t u16;
146766: # 84 "vgatables.h"
146766: struct __attribute__((packed)) VideoParam_s {
146766:     u8 twidth;
146766:     u8 theightm1;
146766:     u8 cheight;
146766:     u16 slength;
146766:     u8 sequ_regs[4];
146766:     u8 miscreg;
146766:     u8 crtc_regs[25];
146766:     u8 actl_regs[20];
146766:     u8 grdc_regs[9];
146766: };
146766: 
146766: struct vgamode_s {
146766:     u8 svgamode;
146766:     struct VideoParam_s *vparam;
146766:     u8 memmodel;
146766:     u8 pixbits;
146766:     u16 sstart;
146766:     u8 pelmask;
146766:     u8 *dac;
146766:     u16 dacsize;
146766: };
146766: 
146766: struct saveVideoHardware {
146766:     u8 sequ_index;
146766:     u8 crtc_index;
146766:     u8 grdc_index;
146766:     u8 actl_index;
146766:     u8 feature;
146766:     u8 sequ_regs[4];
146766:     u8 sequ0;
146766:     u8 crtc_regs[25];
146766:     u8 actl_regs[20];
146766:     u8 grdc_regs[9];
146766:     u16 crtc_addr;
146766:     u8 plane_latch[4];
146766: };
146766: 
146766: struct saveBDAstate {
146766:     u8 video_mode;
146766:     u16 video_cols;
146766:     u16 video_pagesize;
146766:     u16 crtc_address;
146766:     u8 video_rows;
146766:     u16 char_height;
146766:     u8 video_ctl;
146766:     u8 video_switches;
146766:     u8 modeset_ctl;
146766:     u16 cursor_type;
146766:     u16 cursor_pos[8];
146766:     u16 video_pagestart;
146766:     u8 video_page;
146766: 
146766: 
146766: 
146766: 
146766: 
146766: };
146766: 
146766: struct saveDACcolors {
146766:     u8 rwmode;
146766:     u8 peladdr;
146766:     u8 pelmask;
146766:     u8 dac[768];
146766:     u8 color_select;
146766: };
146766: 
146766: 
146766: struct vgamode_s *find_vga_entry(u8 mode);
146766: extern u16 video_save_pointer_table[];
146766: extern struct VideoParam_s video_param_table[];
146766: extern u8 static_functionality[];
146766: 
146766: 
146766: extern const u8 vgafont8[];
146766: extern const u8 vgafont14[];
146766: extern const u8 vgafont16[];
146766: extern const u8 vgafont14alt[];
146766: extern const u8 vgafont16alt[];
146766: 
146766: 
146766: struct carattr {
146766:     u8 car, attr, use_attr;
146766: };
146766: struct cursorpos {
146766:     u8 x, y, page;
146766: };
146766: 
146766: 
146766: void clear_screen(struct vgamode_s *vmode_g);
146766: void vgafb_scroll(int nblines, int attr
146766:                   , struct cursorpos ul, struct cursorpos lr);
146766: void vgafb_write_char(struct cursorpos cp, struct carattr ca);
146766: struct carattr vgafb_read_char(struct cursorpos cp);
146766: void vgafb_write_pixel(u8 color, u16 x, u16 y);
146766: u8 vgafb_read_pixel(u16 x, u16 y);
146766: void vgafb_load_font(u16 seg, void *src_far, u16 count
146766:                      , u16 start, u8 destflags, u8 fontsize);
146766: 
146766: 
146766: void vgahw_screen_disable(void);
146766: void vgahw_screen_enable(void);
146766: void vgahw_set_border_color(u8 color);
146766: void vgahw_set_overscan_border_color(u8 color);
146766: u8 vgahw_get_overscan_border_color(void);
146766: void vgahw_set_palette(u8 palid);
146766: void vgahw_set_single_palette_reg(u8 reg, u8 val);
146766: u8 vgahw_get_single_palette_reg(u8 reg);
146766: void vgahw_set_all_palette_reg(u8 *data);
146766: void vgahw_get_all_palette_reg(u8 *data);
146766: void vgahw_toggle_intensity(u8 flag);
146766: void vgahw_select_video_dac_color_page(u8 flag, u8 data);
146766: void vgahw_read_video_dac_state(u8 *pmode, u8 *curpage);
146766: void vgahw_set_dac_regs(u8 *data, u8 start, int count);
146766: void vgahw_get_dac_regs(u8 *data, u8 start, int count);
146766: void vgahw_set_pel_mask(u8 val);
146766: u8 vgahw_get_pel_mask(void);
146766: void vgahw_save_dac_state(struct saveDACcolors *info);
146766: void vgahw_restore_dac_state(struct saveDACcolors *info);
146766: void vgahw_sequ_write(u8 index, u8 value);
146766: void vgahw_grdc_write(u8 index, u8 value);
146766: void vgahw_set_text_block_specifier(u8 spec);
146766: void get_font_access(void);
146766: void release_font_access(void);
146766: void vgahw_set_cursor_shape(u8 start, u8 end);
146766: void vgahw_set_active_page(u16 address);
146766: void vgahw_set_cursor_pos(u16 address);
146766: void vgahw_set_scan_lines(u8 lines);
146766: u16 vgahw_get_vde(void);
146766: void vgahw_save_state(struct saveVideoHardware *info);
146766: void vgahw_restore_state(struct saveVideoHardware *info);
146766: void vgahw_set_mode(struct VideoParam_s *vparam_g);
146766: void vgahw_enable_video_addressing(u8 disable);
146766: void vgahw_init(void);
146766: 
146766: 
146766: void cirrus_set_video_mode(u8 mode);
146766: void cirrus_init(void);
146766: # 24 "console.c" 2
146766: 
146766: 
146766: static void
146766: output_crnl(void)
146766: {
146766:   crb_puts(0, "\r\n", 2);
146766: }
146766: 
146766: static void
146766: output_bell(void)
146766: {
146766:   crb_puts(0, "\a", 1);
146766: }
146766: 
146766: static void
146766: backspace_and_erase(void)
146766: {
146766:   crb_puts(0, "\b \b", 3);
146766: }
146766: 
146766: static unsigned long
146766: getline(char *buf, unsigned long bufsize)
146766: {
146766:   unsigned long len = 0;
146766:   long c;
146766: 
146766:   while (1)
146766:     {
146766:       c = crb_getc(0);
146766:       if (c < 0)
146766:  continue;
146766:       switch ((int)c)
146766:  {
146766:  case '\r':
146766:  case '\n':
146766:    output_crnl();
146766:    buf[len] = 0;
146766:    return len;
146766: 
146766:         case '\b':
146766:  case 0x7f:
146766:           if (len > 0)
146766:      {
146766:        backspace_and_erase();
146766:               len--;
146766:             }
146766:    else
146766:      output_bell();
146766:           break;
146766: 
146766:         default:
146766:    if (len + 1 < bufsize)
146766:      {
146766:        buf[len] = c;
146766:               crb_puts(0, buf+len, 1);
146766:        len++;
146766:      }
146766:    else
146766:      output_bell();
146766:    break;
146766:         }
146766:     }
146766: }
146766: 
146766: static inline void set_console_alarm(void)
146766: {
146766: 
146766:   set_alarm_rel(10 * 1000 * 1000);
146766: }
146766: 
146766: void
146766: do_entInt(unsigned long type, unsigned long vector)
146766: {
146766:   switch (type)
146766:     {
146766:     case 0:
146766: 
146766: 
146766:       break;
146766:     case 1:
146766: 
146766:       set_console_alarm();
146766:       break;
146766:     case 2:
146766: 
146766: 
146766:       break;
146766:     }
146766: }
146766: 
146766: void
146766: do_console(void)
146766: {
146766:   char line[256];
146766:   unsigned long len;
146766: 
146766:   wrkgp();
146766:   wrent(entInt, 0);
146766:   set_console_alarm();
146766:   swpipl(0);
146766: 
146766:   if (have_vga)
146766:   {
146766:     unsigned short *vga, attr;
146766:     vga = pci_mem_base + 0xB800 *16;
146766:     attr = 0x2000;
146766:     vga[0] = 'H' + attr;
146766:     vga[1] = 'e' + attr;
146766:     vga[2] = 'l' + attr;
146766:     vga[3] = 'l' + attr;
146766:     vga[4] = 'o' + attr;
146766:   }
146766: 
146766:   while (1)
146766:     {
146766:       crb_puts(0, ">>> ", 4);
146766:       len = getline(line, sizeof(line));
146766:       crb_puts(0, "got: ", 5);
146766:       crb_puts(0, line, len);
146766:       output_crnl();
146766:     }
146766: }
=== END GCC DUMP ===

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#12

This is reproducible in sbuild and occurs in the cross build for hppa firmware.

Failing command:
$ alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o console.o console.c

But to reproduce this way easier you might run just:
$ git clone https://gitlab.com/qemu-project/qemu-palcode.git
$ apt install gcc-alpha-linux-gnu
$ make CROSS=alpha-linux-gnu-

Was there an upstream gcc bug filed that we could attach this to?

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#13

Mitigation

I tried putting some less load on the register transfer and found that using "-O1" (down from 2) already would work around the issue. This does in no means make the issue go away, but it helps to service qemu until resolved.

Internally this is set statically and can be overwritten from d/rules to use -O1 instead.

--- a/debian/rules
+++ b/debian/rules
@@ -557,7 +557,7 @@ sysdata-components += qboot
build-palcode-clipper: b/qemu-palcode/palcode-clipper
b/qemu-palcode/palcode-clipper: | b
        cp -al roms/qemu-palcode b/
- ${MAKE} -C ${CURDIR}/b/qemu-palcode CROSS=${ALPHAEV67_CROSSPFX}
+ ${MAKE} -C ${CURDIR}/b/qemu-palcode CROSS=${ALPHAEV67_CROSSPFX} OPT=-O1
        ${ALPHAEV67_CROSSPFX}strip b/qemu-palcode/palcode-clipper
install-palcode-clipper: b/qemu-palcode/palcode-clipper
        install -m 0644 $< ${sysdataidir}/palcode-clipper

This is already emulation of alpha code, the performance impact is tolerable to make it overall build again.

This worked locally already, currently building a test in PPA:
https://launchpad.net/~paelzer/+archive/ubuntu/lp-1988710-gcc-crash-qemu/+packages

Revision history for this message

In GCC Bugzilla #106966, Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#17

Created attachment 53591
Full interim state generated with -freport-bug

Hi,
this was found as part of Ubuntu's build of qemu for the upcoming 22.10 release.
That already uses gcc 12.2. But gladly it can be reproduced much easier as it breaks in cross-building one of the firmware blobs.

Repro steps in Ubuntu 22.10:
$ git clone https://gitlab.com/qemu-project/qemu-palcode.git
$ cd qemu-palcode/
$ apt install gcc-alpha-linux-gnu
$ make CROSS=alpha-linux-gnu-
...

alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
130 | vga[0] = 'H' + attr;
| ~~~~~~~^~~~~~~~~~~~
0x137917b internal_error(char const*, ...)
???:0
0x5a9326 fancy_abort(char const*, int, char const*)
???:0
0xe0a692 alpha_split_const_mov(machine_mode, rtx_def**)
???:0
0xe0a801 alpha_expand_mov(machine_mode, rtx_def**)
???:0
0x112e2dd gen_movv4hi(rtx_def*, rtx_def*)
???:0
0x7eedeb emit_move_insn_1(rtx_def*, rtx_def*)
???:0
0x7ef1d7 emit_move_insn(rtx_def*, rtx_def*)
???:0
0xe0d016 alpha_expand_movmisalign(machine_mode, rtx_def**)
???:0
0x112e5aa gen_movmisalignv4hi(rtx_def*, rtx_def*)
???:0
0xa278fc expand_insn(insn_code, unsigned int, expand_operand*)
???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
make: *** [<builtin>: console.o] Error 1

Note: Sadly I have no more recent e.g. gcc-snapshot of the cross toolchain to test.

The issue occurred on my laptop on the builders of Ubuntu, so I assume it is pretty generic.

Details about how GCC was built can be fetched from the Ubuntu build logs at https://launchpad.net/ubuntu/+source/gcc-12/12.2.0-2ubuntu1

I have tried to compare older builds using gcc on Jammy. The version 11.3.0-1ubuntu1~22.04 there works fine. So it seems to be a regression between that and 12.2.0-2ubuntu1.

Note: I have found that setting -O1 (instead of the -O2 default) will mitigate the issue.

Created attachment 53591
Full interim state generated with -freport-bug

Hi,
this was found as part of Ubuntu's build of qemu for the upcoming 22.10 release.
That already uses gcc 12.2. But gladly it can be reproduced much easier as it breaks in cross-building one of the firmware blobs.

Repro steps in Ubuntu 22.10:
$ git clone https://gitlab.com/qemu-project/qemu-palcode.git
$ cd qemu-palcode/
$ apt install gcc-alpha-linux-gnu
$ make CROSS=alpha-linux-gnu-
...

alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
  130 |     vga[0] = 'H' + attr;
      |     ~~~~~~~^~~~~~~~~~~~
0x137917b internal_error(char const*, ...)
	???:0
0x5a9326 fancy_abort(char const*, int, char const*)
	???:0
0xe0a692 alpha_split_const_mov(machine_mode, rtx_def**)
	???:0
0xe0a801 alpha_expand_mov(machine_mode, rtx_def**)
	???:0
0x112e2dd gen_movv4hi(rtx_def*, rtx_def*)
	???:0
0x7eedeb emit_move_insn_1(rtx_def*, rtx_def*)
	???:0
0x7ef1d7 emit_move_insn(rtx_def*, rtx_def*)
	???:0
0xe0d016 alpha_expand_movmisalign(machine_mode, rtx_def**)
	???:0
0x112e5aa gen_movmisalignv4hi(rtx_def*, rtx_def*)
	???:0
0xa278fc expand_insn(insn_code, unsigned int, expand_operand*)
	???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
make: *** [<builtin>: console.o] Error 1

Note: Sadly I have no more recent e.g. gcc-snapshot of the cross toolchain to test.

The issue occurred on my laptop on the builders of Ubuntu, so I assume it is pretty generic.

Details about how GCC was built can be fetched from the Ubuntu build logs at https://launchpad.net/ubuntu/+source/gcc-12/12.2.0-2ubuntu1

I have tried to compare older builds using gcc on Jammy. The version 11.3.0-1ubuntu1~22.04 there works fine. So it seems to be a regression between that and 12.2.0-2ubuntu1.

Note: I have found that setting -O1 (instead of the -O2 default) will mitigate the issue.

Revision history for this message

In GCC Bugzilla #106966, Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#18

I compared a few more cross-gcc's I could get hold of.
Thereby I can state this was already broken with 12.1.0 on Ubuntu 22.04 and Fedora 36.

Note: I'm only listing where the instructions for these differ

1. Fedora 36
gcc version 12.1.1 20220507 (Red Hat Cross 12.1.1-1) (GCC)

$ dnf install gcc-alpha-linux-gnu make git
...
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
130 | vga[0] = 'H' + attr;
| ~~~~~~~^~~~~~~~~~~~
Please submit a full bug report, with preprocessed source (by using -freport-bug).
See <http://bugzilla.redhat.com/bugzilla/> for instructions.
Preprocessed source stored into /tmp/cchVNtxj.out file, please attach this to your bugreport.
make: *** [<builtin>: console.o] Error 1

2. Ubuntu 22.04 gcc-12
gcc version 12.1.0 (Ubuntu 12.1.0-2ubuntu1~22.04)

$ apt install gcc-12-alpha-linux-gnu
...
alpha-linux-gnu-gcc-12 -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"' -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
130 | vga[0] = 'H' + attr;
| ~~~~~~~^~~~~~~~~~~~
0x13767bb internal_error(char const*, ...)
???:0
0x5a92b0 fancy_abort(char const*, int, char const*)
???:0
0xe09a32 alpha_split_const_mov(machine_mode, rtx_def**)
???:0
0xe09ba1 alpha_expand_mov(machine_mode, rtx_def**)
???:0
0x112c7fd gen_movv4hi(rtx_def*, rtx_def*)
???:0
0x7ee88b emit_move_insn_1(rtx_def*, rtx_def*)
???:0
0x7eec77 emit_move_insn(rtx_def*, rtx_def*)
???:0
0xe0c3b6 alpha_expand_movmisalign(machine_mode, rtx_def**)
???:0
0x112caca gen_movmisalignv4hi(rtx_def*, rtx_def*)
???:0
0xa272fc expand_insn(insn_code, unsigned int, expand_operand*)
???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
make: *** [<builtin>: console.o] Error 1

I compared a few more cross-gcc's I could get hold of.
Thereby I can state this was already broken with 12.1.0 on Ubuntu 22.04 and Fedora 36.

Note: I'm only listing where the instructions for these differ

1. Fedora 36
gcc version 12.1.1 20220507 (Red Hat Cross 12.1.1-1) (GCC)

$ dnf install gcc-alpha-linux-gnu make git
...
alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
  130 |     vga[0] = 'H' + attr;
      |     ~~~~~~~^~~~~~~~~~~~
Please submit a full bug report, with preprocessed source (by using -freport-bug).
See <http://bugzilla.redhat.com/bugzilla/> for instructions.
Preprocessed source stored into /tmp/cchVNtxj.out file, please attach this to your bugreport.
make: *** [<builtin>: console.o] Error 1

2. Ubuntu 22.04 gcc-12
gcc version 12.1.0 (Ubuntu 12.1.0-2ubuntu1~22.04)

$ apt install gcc-12-alpha-linux-gnu
...
alpha-linux-gnu-gcc-12 -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67 -DSYSTEM_H='"sys-clipper.h"'  -c -o console.o console.c
during RTL pass: expand
console.c: In function ‘do_console’:
console.c:130:12: internal compiler error: in emit_move_insn, at expr.cc:4010
  130 |     vga[0] = 'H' + attr;
      |     ~~~~~~~^~~~~~~~~~~~
0x13767bb internal_error(char const*, ...)
	???:0
0x5a92b0 fancy_abort(char const*, int, char const*)
	???:0
0xe09a32 alpha_split_const_mov(machine_mode, rtx_def**)
	???:0
0xe09ba1 alpha_expand_mov(machine_mode, rtx_def**)
	???:0
0x112c7fd gen_movv4hi(rtx_def*, rtx_def*)
	???:0
0x7ee88b emit_move_insn_1(rtx_def*, rtx_def*)
	???:0
0x7eec77 emit_move_insn(rtx_def*, rtx_def*)
	???:0
0xe0c3b6 alpha_expand_movmisalign(machine_mode, rtx_def**)
	???:0
0x112caca gen_movmisalignv4hi(rtx_def*, rtx_def*)
	???:0
0xa272fc expand_insn(insn_code, unsigned int, expand_operand*)
	???:0
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
make: *** [<builtin>: console.o] Error 1

Revision history for this message

Paride Legovini (paride) wrote on 2022-09-19:

#14

Thanks Christian. I couldn't find a relevant upstream GCC bug for this issue.

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#15

Since I have not found this issue in qemu, gcc nor palcode I have filed it for some more awareness than just this LP report.

- https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106966
- https://github.com/qemu/qemu-palcode/issues/1

FYI I have in the meantime also recreated this in more release/versions:
- Bad: Kinetic - default cross gcc - gcc version 12.2.0 (Ubuntu 12.2.0-2ubuntu1)
- Bad: Jammy - cross gcc-12 - gcc version 11.3.0 (Ubuntu 11.3.0-1ubuntu1~22.04)
- Bad: Fedora 36 - cross gcc - gcc version 12.1.1 20220507 (Red Hat Cross 12.1.1-1) (GCC)
- Good: Jammy - default cross gcc - gcc version 12.2.0 (Ubuntu 12.2.0-2ubuntu1)

Changed in qemu (Ubuntu):
status:	New → In Progress

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#16

FYI Qemu mitigation (also to unblock other fixes to land)

PPA: https://launchpad.net/~paelzer/+archive/ubuntu/lp-1988710-gcc-crash-qemu/+packages
MP: https://code.launchpad.net/~paelzer/ubuntu/+source/qemu/+git/qemu/+merge/430100

Bug Watch Updater (bug-watch-updater) on 2022-09-19

Changed in gcc:
importance:	Unknown → Medium
status:	Unknown → New

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#19

After this is good, next is:

/<<PKGBUILDDIR>>/roms/opensbi/lib/sbi/sbi_tlb.c: Assembler messages:
/<<PKGBUILDDIR>>/roms/opensbi/lib/sbi/sbi_tlb.c:190: Error: unrecognized opcode `fence.i', extension `zifencei' required
make[3]: *** [Makefile:431: /<<PKGBUILDDIR>>/b/opensbi/lib/sbi/sbi_tlb.o] Error 1

Revision history for this message

Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#20

That is reproducible with the snapshot in qemu, but ok with latest branch from https://gitlab.com/qemu-project/opensbi.git

Fix:
https://gitlab.com/qemu-project/opensbi/-/commit/5d53b55aa77ffeefd4012445dfa6ad3535e1ff2c

MP and PPA updated for that.

Revision history for this message

In GCC Bugzilla #106966, Ubizjak (ubizjak) wrote on 2022-09-19:

#22

(In reply to Christian Ehrhardt from comment #0)

> alpha-linux-gnu-gcc -O2 -g1 -Wall -fvisibility=hidden -fno-strict-aliasing
> -msmall-text -msmall-data -mno-fp-regs -mbuild-constants -mcpu=ev67

Just drop -mbuild-constants.

(There is a problem in alpha_emit_set_long_const which is not prepared to handle V4HImode target).

Revision history for this message

In GCC Bugzilla #106966, Christian Ehrhardt  (paelzer) wrote on 2022-09-19:

#23

> Just drop -mbuild-constants.

Thanks for the hint Uroš, but I'm not sure if one can do that, this option is from [1]. I do not have the background on this, but it reads as there was a reason "Use -mbuild-constants to prevent the compiler using static data" to set this which seems more breaking than my current workaround (reduce -O2 to -O1).

[1]: https://github.com/qemu/qemu-palcode/commit/0830e72f0bce29bdf1de0d67ad503a9a8b99c968

Christian Ehrhardt  (paelzer) on 2022-09-20

tags:	added: server-todo
Changed in qemu (Ubuntu):
assignee:	nobody → Christian Ehrhardt  (paelzer)

Revision history for this message

In GCC Bugzilla #106966, Ubizjak (ubizjak) wrote on 2022-09-20:

#24

(In reply to Christian Ehrhardt from comment #3)
> > Just drop -mbuild-constants.
>
> Thanks for the hint Uroš, but I'm not sure if one can do that, this option
> is from [1]. I do not have the background on this, but it reads as there was
> a reason "Use -mbuild-constants to prevent the compiler using static data"
> to set this which seems more breaking than my current workaround (reduce -O2
> to -O1).
>
> [1]:
> https://github.com/qemu/qemu-palcode/commit/
> 0830e72f0bce29bdf1de0d67ad503a9a8b99c968

-mbuild-constants is used here together with -msmall-data (which limits relocations to the small data area to 16 bits). -mbuild-constants option tries harder to construct 64-bit constants using arithmetic with immediates instead of loads from the small data area. So, if the linker does not error out with relocation overflow, it can be safely omitted.

The fix to the compiler is relatively straightforward, but I have no access to alpha system any more to properly test the fix (and I have no patience to run the testsuite on the simulator). The -mbuild-constant option is rarely used, and the proposed workaround should be safe.

Revision history for this message

Launchpad Janitor (janitor) wrote on 2022-09-21:

#21

This bug was fixed in the package qemu - 1:7.0+dfsg-7ubuntu2

---------------
qemu (1:7.0+dfsg-7ubuntu2) kinetic; urgency=medium

  [ Paride Legovini ]
  * d/rules: disable LTO on non-amd64 builds (LP: #1921664)
  * GCC-12 FTBFS (LP: #1988710)
    - d/p/u/lp1988710-silence-openbios-array-bounds-false-positive.patch.
      Silence -Warray-bounds false positive (treated as error)

  [ Christian Ehrhardt ]
  * More on GCC-12 FTBFS (LP 1988710)
    - d/rules: set -O1 for alpha firmware build
    - d/p/u/lp1988710-opensbi-Makefile-fix-build-with-binutils-2.38.patch:
      further FTBFS fixup

-- Christian Ehrhardt <email address hidden> Mon, 19 Sep 2022 08:07:24 +0200

Changed in qemu (Ubuntu):
status:	In Progress → Fix Released

Revision history for this message

In GCC Bugzilla #106966, Christian Ehrhardt  (paelzer) wrote on 2022-09-22:

#25

Thanks Uroš for the explanation.

I've tried and it does not bail out with a relocation overflow when dropping -mbuild-constants.

I've suggested to upstream to consider that change.
=> https://gitlab.com/qemu-project/qemu-palcode/-/merge_requests/1

Revision history for this message

In GCC Bugzilla #106966, Ubizjak (ubizjak) wrote on 2022-09-23:

#26

Testcase:

void
do_console (unsigned short *vga)
{
  vga[0] = 'H';
  vga[1] = 'e';
  vga[2] = 'l';
  vga[3] = 'l';
  vga[4] = 'o';
}

cc1 -O2 -mbuild-constants:

during RTL pass: expand
pr106966.c: In function ‘do_console’:
pr106966.c:4:10: internal compiler error: in emit_move_insn, at expr.cc:4052
    4 | vga[0] = 'H';
      | ~~~~~~~^~~~~
0x5ed664 emit_move_insn(rtx_def*, rtx_def*)
        ../../git/gcc/gcc/expr.cc:4052
0x113a708 alpha_emit_set_long_const
        ../../git/gcc/gcc/config/alpha/alpha.cc:2090
0x113abfc alpha_split_const_mov(machine_mode, rtx_def**)
        ../../git/gcc/gcc/config/alpha/alpha.cc:2204
0x113ade7 alpha_expand_mov(machine_mode, rtx_def**)
        ../../git/gcc/gcc/config/alpha/alpha.cc:2250
0x15510a8 gen_movv4hi(rtx_def*, rtx_def*)
        ../../git/gcc/gcc/config/alpha/alpha.md:4463
...

(gdb) bt
#0 internal_error (gmsgid=gmsgid@entry=0x1a9cf85 "in %s, at %s:%d") at ../../git/gcc/gcc/diagnostic.cc:2107
#1 0x00000000007918d6 in fancy_abort (file=file@entry=0x1910cf9 "../../git/gcc/gcc/expr.cc", line=line@entry=4052,
function=function@entry=0x1a23a78 "emit_move_insn") at ../../git/gcc/gcc/diagnostic.cc:2211
#2 0x00000000005ed665 in emit_move_insn (x=<optimized out>, y=<optimized out>) at ../../git/gcc/gcc/expr.cc:4052
#3 0x000000000113a709 in alpha_emit_set_long_const (target=0x7fffea2ef9f0, c1=7077888) at ../../git/gcc/gcc/config/alpha/alpha.cc:2090
#4 0x000000000113abfd in alpha_split_const_mov (mode=E_V4HImode, operands=0x7fffffffd5a0) at ../../git/gcc/gcc/config/alpha/alpha.cc:2204
#5 0x000000000113ade8 in alpha_expand_mov (mode=E_V4HImode, operands=0x7fffffffd5a0) at ../../git/gcc/gcc/config/alpha/alpha.cc:2250
#6 0x00000000015510a9 in gen_movv4hi (operand0=0x7fffea2ef9f0, operand1=0x7fffea29ebc0) at ../../git/gcc/gcc/config/alpha/alpha.md:4463

(gdb) f 3
#3 0x000000000113a709 in alpha_emit_set_long_const (target=0x7fffea2ef9f0, c1=7077888) at ../../git/gcc/gcc/config/alpha/alpha.cc:2090
2090 emit_move_insn (target, gen_rtx_PLUS (DImode, target, GEN_INT (d3)));
(gdb) p debug_rtx (target)
(reg:V4HI 71)
$1 = void

Testcase:

void
do_console (unsigned short *vga)
{
  vga[0] = 'H';
  vga[1] = 'e';
  vga[2] = 'l';
  vga[3] = 'l';
  vga[4] = 'o';
}

cc1 -O2 -mbuild-constants:

during RTL pass: expand
pr106966.c: In function ‘do_console’:
pr106966.c:4:10: internal compiler error: in emit_move_insn, at expr.cc:4052
    4 |   vga[0] = 'H';
      |   ~~~~~~~^~~~~
0x5ed664 emit_move_insn(rtx_def*, rtx_def*)
        ../../git/gcc/gcc/expr.cc:4052
0x113a708 alpha_emit_set_long_const
        ../../git/gcc/gcc/config/alpha/alpha.cc:2090
0x113abfc alpha_split_const_mov(machine_mode, rtx_def**)
        ../../git/gcc/gcc/config/alpha/alpha.cc:2204
0x113ade7 alpha_expand_mov(machine_mode, rtx_def**)
        ../../git/gcc/gcc/config/alpha/alpha.cc:2250
0x15510a8 gen_movv4hi(rtx_def*, rtx_def*)
        ../../git/gcc/gcc/config/alpha/alpha.md:4463
...

(gdb) bt
#0  internal_error (gmsgid=gmsgid@entry=0x1a9cf85 "in %s, at %s:%d") at ../../git/gcc/gcc/diagnostic.cc:2107
#1  0x00000000007918d6 in fancy_abort (file=file@entry=0x1910cf9 "../../git/gcc/gcc/expr.cc", line=line@entry=4052, 
    function=function@entry=0x1a23a78 "emit_move_insn") at ../../git/gcc/gcc/diagnostic.cc:2211
#2  0x00000000005ed665 in emit_move_insn (x=<optimized out>, y=<optimized out>) at ../../git/gcc/gcc/expr.cc:4052
#3  0x000000000113a709 in alpha_emit_set_long_const (target=0x7fffea2ef9f0, c1=7077888) at ../../git/gcc/gcc/config/alpha/alpha.cc:2090
#4  0x000000000113abfd in alpha_split_const_mov (mode=E_V4HImode, operands=0x7fffffffd5a0) at ../../git/gcc/gcc/config/alpha/alpha.cc:2204
#5  0x000000000113ade8 in alpha_expand_mov (mode=E_V4HImode, operands=0x7fffffffd5a0) at ../../git/gcc/gcc/config/alpha/alpha.cc:2250
#6  0x00000000015510a9 in gen_movv4hi (operand0=0x7fffea2ef9f0, operand1=0x7fffea29ebc0) at ../../git/gcc/gcc/config/alpha/alpha.md:4463

(gdb) f 3
#3  0x000000000113a709 in alpha_emit_set_long_const (target=0x7fffea2ef9f0, c1=7077888) at ../../git/gcc/gcc/config/alpha/alpha.cc:2090
2090            emit_move_insn (target, gen_rtx_PLUS (DImode, target, GEN_INT (d3)));
(gdb) p debug_rtx (target)
(reg:V4HI 71)
$1 = void

Bug Watch Updater (bug-watch-updater) on 2022-09-25

Changed in gcc:
status:	New → Confirmed

Julian Andres Klode (juliank) on 2022-09-29

tags:	removed: foundations-todo
Changed in gcc-12 (Ubuntu):
importance:	Undecided → Medium

Revision history for this message

Dominique Pellé (dominique-pelle) wrote on 2023-02-18:

#27

Can gcc-12 be updated in Ubuntu-22.04?

Currently as of Feb 18 2023, gcc-12 (Ubuntu 12.1.0-2ubuntu1~22.04) 12.1.0 is packaged in Ubuntu 22.04.

Upstream, gcc-12.2 was released on Aug 2022, so many months ago according to https://gcc.gnu.org/gcc-12/

I'm asking because I'm also observing crashes when building C++ projects with gcc-12.1 from Ubuntu-22.04 and gcc-12.2 fixes several of such bugs according to the list of bug fixes in gcc-12.2 at https://gcc.gnu.org/bugzilla/buglist.cgi?bug_status=RESOLVED&resolution=FIXED&target_milestone=12.2

Christian Ehrhardt  (paelzer) on 2023-03-08

tags:

removed: server-todo

Revision history for this message

Launchpad Janitor (janitor) wrote on 2023-04-30:

#28

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in gcc-12 (Ubuntu):
status:	New → Confirmed

Revision history for this message

In GCC Bugzilla #106966, Ubizjak (ubizjak) wrote on 2023-07-08:

#29

*** Bug 110597 has been marked as a duplicate of this bug. ***

Revision history for this message

In GCC Bugzilla #106966, Ubizjak (ubizjak) wrote on 2023-07-08:

#30

Created attachment 55504
Proposed patch.

Can someone please bootstrap and test the attached patch?

Revision history for this message

In GCC Bugzilla #106966, Matoro-gcc-bugzilla (matoro-gcc-bugzilla) wrote on 2023-07-10:

#31

(In reply to Uroš Bizjak from comment #8)
> Created attachment 55504 [details]
> Proposed patch.
>
> Can someone please bootstrap and test the attached patch?

I can queue this up to test on real hardware. By bootstrap, do you mean with --enable-bootstrap, and by test do you mean a full testsuite run or just checking that it doesn't ICE on the reproducer here?

Revision history for this message

In GCC Bugzilla #106966, Ubizjak (ubizjak) wrote on 2023-07-10:

#32

(In reply to matoro from comment #9)
> (In reply to Uroš Bizjak from comment #8)
> > Created attachment 55504 [details]
> > Proposed patch.
> >
> > Can someone please bootstrap and test the attached patch?
>
> I can queue this up to test on real hardware. By bootstrap, do you mean
> with --enable-bootstrap, and by test do you mean a full testsuite run or
> just checking that it doesn't ICE on the reproducer here?

I have already checked that it doesn't ICE with a crosscompiler. Regarding the bootstrap, please note that --enable-bootstrap is the default nowadays, so the option it is not needed to make a full build ("bootstrap"). Also, "make -j N -k check-gcc" should be enough to test the compiler, since other parts of the compiler never exercise -mbuild-constants (N in the command should be substituted with a number of processors to parallelize the testsuite run).

Revision history for this message

In GCC Bugzilla #106966, Matoro-gcc-bugzilla (matoro-gcc-bugzilla) wrote on 2023-07-10:

#33

(In reply to Uroš Bizjak from comment #10)
> (In reply to matoro from comment #9)
> > (In reply to Uroš Bizjak from comment #8)
> > > Created attachment 55504 [details]
> > > Proposed patch.
> > >
> > > Can someone please bootstrap and test the attached patch?
> >
> > I can queue this up to test on real hardware. By bootstrap, do you mean
> > with --enable-bootstrap, and by test do you mean a full testsuite run or
> > just checking that it doesn't ICE on the reproducer here?
>
> I have already checked that it doesn't ICE with a crosscompiler. Regarding
> the bootstrap, please note that --enable-bootstrap is the default nowadays,
> so the option it is not needed to make a full build ("bootstrap"). Also,
> "make -j N -k check-gcc" should be enough to test the compiler, since other
> parts of the compiler never exercise -mbuild-constants (N in the command
> should be substituted with a number of processors to parallelize the
> testsuite run).

I was just checking because alpha hardware is quite antique, it's single-core so a full compile with bootstrap takes around ~72 hours. I don't know how long the test suite takes because I've never run it on this machine, but I would expect it to add a couple additional days to that. If that's fine I'll kick it off, just want to be clear on what exactly you'd like to see.

Revision history for this message

In GCC Bugzilla #106966, Doko-v (doko-v) wrote on 2023-07-10:

#34

I'll apply the proposed patch for the next gcc-13 Debian upload, then reporting back test results.

Revision history for this message

In GCC Bugzilla #106966, Doko-v (doko-v) wrote on 2023-07-13:

#35

search for "test-summary"
https://buildd.debian.org/status/logs.php?pkg=gcc-12&arch=alpha

12.3.0-5 is gcc-12 branch 20230630
12.3.0-6 is gcc-12 branch 20230707 with the proposed patch applied

comparing gcc summary (-5 to -6):

=== gcc Summary ===

# of expected passes 136199
# of unexpected failures 657
# of unexpected successes 16
# of expected failures 1112
# of unresolved testcases 8
# of unsupported tests 2539
/<<PKGBUILDDIR>>/build/gcc/xgcc version 12.3.0 (Debian 12.3.0-5)

=== gcc Summary ===

# of expected passes 136261
# of unexpected failures 652
# of unexpected successes 16
# of expected failures 1112
# of unresolved testcases 8
# of unsupported tests 2539
/<<PKGBUILDDIR>>/build/gcc/xgcc version 12.3.0 (Debian 12.3.0-6)

Revision history for this message

In GCC Bugzilla #106966, Cvs-commit (cvs-commit) wrote on 2023-07-13:

#36

The master branch has been updated by Uros Bizjak <email address hidden>:

https://gcc.gnu.org/g:337649c1660211db733c1ba34ae260b8c66a3578

commit r14-2503-g337649c1660211db733c1ba34ae260b8c66a3578
Author: Uros Bizjak <email address hidden>
Date: Thu Jul 13 18:32:15 2023 +0200

alpha: Fix computation mode in alpha_emit_set_long_cost [PR106966]

PR target/106966

gcc/ChangeLog:

* config/alpha/alpha.cc (alpha_emit_set_long_const):
Always use DImode when constructing long const.

gcc/testsuite/ChangeLog:

* gcc.target/alpha/pr106966.c: New test.

Revision history for this message

In GCC Bugzilla #106966, Cvs-commit (cvs-commit) wrote on 2023-07-13:

#37

The releases/gcc-13 branch has been updated by Uros Bizjak <email address hidden>:

https://gcc.gnu.org/g:27e421319efcf47280339fbc17c263f36c92eee6

commit r13-7561-g27e421319efcf47280339fbc17c263f36c92eee6
Author: Uros Bizjak <email address hidden>
Date: Thu Jul 13 18:32:15 2023 +0200

alpha: Fix computation mode in alpha_emit_set_long_cost [PR106966]

PR target/106966

gcc/ChangeLog:

* config/alpha/alpha.cc (alpha_emit_set_long_const):
Always use DImode when constructing long const.

gcc/testsuite/ChangeLog:

* gcc.target/alpha/pr106966.c: New test.

(cherry picked from commit 337649c1660211db733c1ba34ae260b8c66a3578)

Revision history for this message

In GCC Bugzilla #106966, Cvs-commit (cvs-commit) wrote on 2023-07-13:

#38

The releases/gcc-12 branch has been updated by Uros Bizjak <email address hidden>:

https://gcc.gnu.org/g:4520e2dbc73262028ad556f732871565101ef615

commit r12-9770-g4520e2dbc73262028ad556f732871565101ef615
Author: Uros Bizjak <email address hidden>
Date: Thu Jul 13 18:32:15 2023 +0200

alpha: Fix computation mode in alpha_emit_set_long_cost [PR106966]

PR target/106966

gcc/ChangeLog:

* config/alpha/alpha.cc (alpha_emit_set_long_const):
Always use DImode when constructing long const.

gcc/testsuite/ChangeLog:

* gcc.target/alpha/pr106966.c: New test.

(cherry picked from commit 337649c1660211db733c1ba34ae260b8c66a3578)

Revision history for this message

In GCC Bugzilla #106966, Ubizjak (ubizjak) wrote on 2023-07-13:

#39

Thanks for helping with tests!

Fixed for gcc-12.4+

Bug Watch Updater (bug-watch-updater) on 2023-07-19

Changed in gcc:
status:	Confirmed → Fix Released

Ubuntu
qemu package

qemu FTBFS with GCC 12

Bug Description

Related branches

Other bug subscribers

Remote bug watches

	Status	Importance	Assigned to
gcc	Fix Released	Medium	gcc-bugzilla #106966
gcc-12 (Ubuntu)	Confirmed	Medium	Unassigned
qemu (Ubuntu)	Fix Released	High	Christian Ehrhardt 

Ubuntuqemu package

qemu FTBFS with GCC 12

Bug Description

Related branches

Other bug subscribers

Remote bug watches

Ubuntu
qemu package