Andreas Schwab <sch...@suse.de> writes:
> On Jul 03 2023, Andreas Schwab wrote: > >> This is a regression in 8.0. It causes the testsuite of mold to fail: >> >> https://build.opensuse.org/package/live_build_log/openSUSE:Factory:RISCV/mold/standard/riscv64 >> >> + out/test/elf/riscv64/section-start/exe2 >> ** >> ERROR:../accel/tcg/cpu-exec.c:1028:cpu_exec_setjmp: assertion failed: (cpu >> == current_cpu) >> ** >> ERROR:../accel/tcg/cpu-exec.c:1028:cpu_exec_setjmp: assertion failed: (cpu >> == current_cpu) > > This bisects down to commit 86f04735ac ("linux-user: Fix brk() to > release pages"). See the attached test case. > > $ ./qemu-riscv64 ../exe1 > ** > ERROR:../qemu/accel/tcg/cpu-exec.c:1027:cpu_exec_setjmp: assertion failed: > (cpu == current_cpu) > Bail out! ERROR:../qemu/accel/tcg/cpu-exec.c:1027:cpu_exec_setjmp: assertion > failed: (cpu == current_cpu) > ** > ERROR:../qemu/accel/tcg/cpu-exec.c:1027:cpu_exec_setjmp: assertion failed: > (cpu == current_cpu) > Bail out! ERROR:../qemu/accel/tcg/cpu-exec.c:1027:cpu_exec_setjmp: assertion > failed: (cpu == current_cpu) I wonder if applying c5ffd16ba4c8fd fixes this? I tried to double check with the test case but I can't run it as it's a dynamic binary so needs an ld.so. -- Alex Bennée Virtualisation Tech Lead @ Linaro
