What were you trying to do that didn't work?

As subject

What is the impact of this issue to you?

under windows server 2022 the default virtio-blk situation is quite
bad, ssd trim is very slow. and defrag a bigger volume like 1TB
disk will always show "memory not enough", even when the volume is
empty.

  I found discussions to change the "discard_granularity" to make trim
happy again. and libvirt support the syntax like below:

   <blockio discard_granularity='2097152'/>

   I also found if I can set "discard_granularity" to zero, then
windows will recognize the device as "traditional hard drive" again
and won't do unnecessary trim to it.

   the sad part is if I setup it under RHEL 9.4 with librirt 10.0 like below:

  <blockio discard_granularity='0'/>

   the line will just disappear when I close "virsh edit"

    so I can only use complex  format with "<qemu:override>" to set"discard_granularity='0'"

    PS: although the parameter seems working fine for windows driver, it seems has no affect for linux. the vm trim capability under rhel 9.4 looks like below no matter the "discard_granularity" is zero or other value.

> lsblk -D
NAME            DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
vda                    0      512B       2G         0

Please provide the package NVR for which the bug is seen:

libvirt-10.0.0-6.7.el9_4.x86_64

How reproducible is this bug?:

100%

Steps to reproduce

1. create a vm with virtio-blk device
2. virsh edit the vm, add lines "<blockio discard_granularity='0'/>" to the device. the parameter is ignored when saved because the value is "0" which looks like the default value.

Expected results

     virsh accept "<blockio discard_granularity='0'/>" as a valid parameter.