Pavel Dovgalyuk <dovga...@ispras.ru> writes:
>> From: Alex Bennée [mailto:alex.ben...@linaro.org] >> Pavel Dovgalyuk <dovga...@ispras.ru> writes: >> >> >> From: Alex Bennée [mailto:alex.ben...@linaro.org] >> >> Pavel Dovgalyuk <pavel.dovga...@ispras.ru> writes: >> >> >> >> > Ping? >> >> >> >> I started having a look but I ran into this straight away. First I >> >> recorded a boot of the kernel: >> >> >> >> ./aarch64-softmmu/qemu-system-aarch64 -machine virt,graphics=on,gic- >> >> version=3,virtualization=on -cpu cortex-a53 --serial mon:stdio -display >> >> none -kernel >> >> ../images/aarch64-current-linux-initrd-guest.img -icount >> shift=7,rr=record,rrfile=replay.bin >> >> >> >> Then played back: >> >> >> >> ./aarch64-softmmu/qemu-system-aarch64 -machine virt,graphics=on,gic- >> >> version=3,virtualization=on -cpu cortex-a53 --serial mon:stdio -display >> >> none -kernel >> >> ../images/aarch64-current-linux-initrd-guest.img -icount >> shift=7,rr=replay,rrfile=replay.bin - >> >> s -S >> > >> > This looks ok, but... >> > >> >> And did the following on gdb: >> >> >> >> (gdb) i >> >> 0x0000000040000004 in ?? () >> >> => 0x40000004: mov x1, xzr >> >> 0x40000008: mov x2, xzr >> >> 0x4000000c: mov x3, xzr >> >> (gdb) >> >> 0x0000000040000008 in ?? () >> >> => 0x40000008: mov x2, xzr >> >> 0x4000000c: mov x3, xzr >> >> 0x40000010: ldr x4, 0x40000020 >> >> (gdb) >> >> 0x000000004000000c in ?? () >> >> => 0x4000000c: mov x3, xzr >> >> 0x40000010: ldr x4, 0x40000020 >> >> 0x40000014: br x4 >> >> (gdb) >> >> 0x0000000040000010 in ?? () >> >> => 0x40000010: ldr x4, 0x40000020 >> >> 0x40000014: br x4 >> >> 0x40000018: .inst 0x44000000 ; undefined >> >> (gdb) >> >> 0x0000000040000014 in ?? () >> >> => 0x40000014: br x4 >> >> 0x40000018: .inst 0x44000000 ; undefined >> >> 0x4000001c: .inst 0x00000000 ; undefined >> >> (gdb) p/x $x4 >> >> $1 = 0x40080000 >> >> (gdb) reverse-stepi >> >> warning: Remote failure reply: E14 >> >> >> >> Surely this is the simple case and doesn't require any snapshots for >> >> block devices as there are none. Am I missing something? >> > >> > Reverse debugging requires the snapshotting. QEMU can't revert the VM >> > state without the >> snapshots. >> > You can try adding an empty qcow2 image to allow snapshotting there. >> >> I suspect a recent patch has broken locking again: >> >> Starting program: >> /home/alex/lsrc/qemu/qemu.git/aarch64-softmmu/qemu-system-aarch64 -machine >> virt,graphics=on,gic-version=3,virtualization=on -cpu cortex-a53 --serial >> mon:stdio -display >> none -kernel ../images/aarch64-current-linux-initrd-guest.img -icount >> shift=7,rr=replay,rrfile=replay.bin,rrsnapshot=debug -drive >> file=rr.qcow2,if=none,snapshot,id=rr -s -S >> [Thread debugging using libthread_db enabled] >> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". >> [New Thread 0x7fffd8536700 (LWP 32452)] >> [New Thread 0x7fffd5bb3700 (LWP 32453)] >> [New Thread 0x7fffd4eab700 (LWP 32454)] >> ** >> ERROR:replay/replay-time.c:49:replay_read_clock: assertion failed: >> (replay_file && >> replay_mutex_locked()) > > Have you recorded it with the attached disk before replaying? Yes. I assume the -drive doesn't actually have to be visible to the guest, it's just the mechanism rr needs for saving snapshots? > Are you using the latest version? > If the both answers are 'yes', then can you share the kernel? My i386 > runs work normally. I'll have a go with x86 first as aarch64 hasn't been proven yet. > >> Once I have the linux-user TCG tests merged I'm planning on focusing on >> the system emulation tests and we should be able to add some >> record/replay tests to defend the behaviour. > > That will be great. > There are some Ciro's attempts on that: > https://github.com/cirosantilli/qemu-test/blob/master/arm/rr > > Pavel Dovgalyuk -- Alex Bennée
