Am 25.03.2020 um 18:23 hat Kevin Wolf geschrieben: > The recent fix didn't actually fix the whole problem. Operations can't > only wait for themselves, but we can also end up with circular > dependencies like two operations waiting for each other to complete. > > This reverts the first fix and implements another approach.
Hm, somehow this seems to break iotests 151. I don't actually understand the backtrace, because that's during job initialisation, so my changes shouldn't have had any effect yet: (gdb) bt #0 0x00007fba6d85057f in raise () at /lib64/libc.so.6 #1 0x00007fba6d83a895 in abort () at /lib64/libc.so.6 #2 0x00005624d94d109a in bitmap_new (nbits=<optimized out>) at /home/kwolf/source/qemu/include/qemu/bitmap.h:103 #3 0x00005624d94d109a in mirror_run (job=0x5624dc8d5560, errp=<optimized out>) at block/mirror.c:922 #4 0x00005624d988053f in job_co_entry (opaque=0x5624dc8d5560) at job.c:878 #5 0x00005624d998d3bb in coroutine_trampoline (i0=<optimized out>, i1=<optimized out>) at util/coroutine-ucontext.c:115 #6 0x00007fba6d866250 in __start_context () at /lib64/libc.so.6 #7 0x00007fffa2d48130 in () #8 0x0000000000000000 in () Something to check tomorrow. Kevin