[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [PATCH v2 2/8] qemu: Allow coldunplugging of hub device




On 11/7/18 8:17 AM, Han Han wrote:
> 
> 
> On Tue, Nov 6, 2018 at 6:29 AM John Ferlan <jferlan redhat com
> <mailto:jferlan redhat com>> wrote:
> 
> 
> 
>     On 10/12/18 4:50 AM, Han Han wrote:
>     > https://bugzilla.redhat.com/show_bug.cgi?id=1375423
>     >
>     > Signed-off-by: Han Han <hhan redhat com <mailto:hhan redhat com>>
>     > ---
>     >  src/conf/domain_conf.c   | 30 ++++++++++++++++++++++++++++++
>     >  src/conf/domain_conf.h   |  3 +++
>     >  src/libvirt_private.syms |  1 +
>     >  src/qemu/qemu_driver.c   | 10 +++++++++-
>     >  4 files changed, 43 insertions(+), 1 deletion(-)
>     >
> 
>     I assume no concerns over something that could be attached to a port on
>     the hub... I guess for a cold unplug it probably won't matter so much.
> 
> For the usb hub coldunplug, I find a concern. If a usb device attatched
> to a usb-hub, then we coldunplug the hub, how will the hub and usb device
> look like? All disappear OR all remain?

Since you've posted the patches, then I would hope you could answer your
own question regarding what happens.

Still in a way this is different than hot/live unplug. Attempting to
start the guest after you've removed a hub that had defined/listed USB
address/attachments would I assume fail because the hub is gone. But
there is a little self doubt because of the automatic hub add algorithm
pointed out in patch 7.

If it does expectedly fail, then someone would have to cold plug a new
device in using the correct address before booting - which is an
expected operation.

In a way no different that someone changing a guest config to use 100
vCPUs or 100G on a host that couldn't support those amounts. When cold
changing things, as long as the value is valid - we're good. When
starting if we don't have the resources, then we're not so good

John

> 
> As I know, for scsi and scsi controllers, it is all remaining :)
> 
> 
>     Reviewed-by: John Ferlan <jferlan redhat com
>     <mailto:jferlan redhat com>>
> 
>     John
> 
> 
> 
> -- 
> Best regards,
> -----------------------------------
> Han Han
> Quality Engineer
> Redhat.
> 
> Email: hhan redhat com <mailto:hhan redhat com>
> Phone: +861065339333


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]