On 3/20/22 10:10, Richard Henderson wrote:
Most of the changes here reduce the amount of locking involved
in logging, due to repeated qemu_log calls, each of which takes
and releases the rcu_read_lock.
This makes more use of qemu_log_lock/unlock around code blocks,
which both keeps the output to
Most of the changes here reduce the amount of locking involved
in logging, due to repeated qemu_log calls, each of which takes
and releases the rcu_read_lock.
This makes more use of qemu_log_lock/unlock around code blocks,
which both keeps the output together in the face of threads and
also plays
