device: 0.991 risc-v: 0.983 graphic: 0.982 vnc: 0.981 performance: 0.978 debug: 0.977 semantic: 0.975 virtual: 0.974 ppc: 0.973 network: 0.972 permissions: 0.971 hypervisor: 0.969 socket: 0.969 user-level: 0.968 kernel: 0.968 register: 0.968 assembly: 0.968 PID: 0.966 peripherals: 0.965 architecture: 0.965 x86: 0.963 VMM: 0.962 arm: 0.958 mistranslation: 0.958 KVM: 0.954 boot: 0.947 TCG: 0.944 i386: 0.941 files: 0.917 Backup with large RBD disk is slow since QEMU 6.2.0 (since commit 0347a8fd) Description of problem: Since commit 0347a8fd4c ("block/rbd: implement bdrv_co_block_status"), there is a big slowdown for large RBD images for backup. Steps to reproduce: I used the following script ``` root@pve701 ~ # cat rbdbackup.sh #!/bin/bash rbd create emptytestA -p rbdkvm --size $2 rbd create emptytestB -p rbdkvm --size $2 $1 \ -qmp stdio \ -drive file=rbd:rbdkvm/emptytestA:conf=/etc/pve/ceph.conf:id=admin:keyring=/etc/pve/priv/ceph/rbdkvm.keyring,if=none,id=driveA,format=raw \ -drive file=rbd:rbdkvm/emptytestB:conf=/etc/pve/ceph.conf:id=admin:keyring=/etc/pve/priv/ceph/rbdkvm.keyring,if=none,id=driveB,format=raw \ <