On 22.10.21 11:28, Andrew Cooper wrote:
> On 22/10/2021 07:47, Juergen Gross wrote:
>> When booting the xenbus driver will wait for PV devices to have
>> connected to their backends before continuing. The timeout is different
>> between essential and non-essential devices.
>>
>> Non-essential devices are identified by their nodenames directly in the
>> xenbus driver, which requires to update this list in case a new device
>> type being non-essential is added (this was missed for several types
>> in the past).
>>
>> In order to avoid this problem, add a "not_essential" flag to struct
>> xenbus_driver which can be set to "true" by the respective frontend.
>>
>> Set this flag for the frontends currently regarded to be not essential
>> (vkbs and vfb) and use it for testing in the xenbus driver.
>>
>> Signed-off-by: Juergen Gross <jgross@suse.com>
> 
> Wouldn't it be better to annotate essential?  That way, when new misc
> drivers come along, they don't by default block boot.

It isn't as if new drivers would "block boot". Normally the short
timeout for all drivers of 30 seconds is more than enough for all of
them.

I'm a little bit hesitant to have a kind of "white listing" essential
drivers, as there might be different views which drivers should have
that flag. Doing this the other way round is easier: in case of
disagreement such a patch just wouldn't go in, not breaking anything
in that case.

Additionally there might be out-of-tree PV drivers, which could be
hit by not being flagged to be essential. With the not_essential flag
the situation wouldn't change for such a driver.


Juergen