# Re: \[PATCH 0/5\] xen: cleanup detection of non-essential pv devices

- From: Juergen Gross \<jgross@suse.com\>
- Date: 2021-10-22 07:34:10 
- List: linux-input
- Page: https://ratatoskr.run/linux-input/2021/10/1550490
- Thread: https://ratatoskr.run/alsa-devel/2021/10/1550456/t.md

```text
On 22.10.21 09:24, Jan Beulich wrote:
> On 22.10.2021 08:47, Juergen Gross wrote:
>> Today the non-essential pv devices are hard coded in the xenbus driver
>> and this list is lacking multiple entries.
>>
>> This series reworks the detection logic of non-essential devices by
>> adding a flag for that purpose to struct xenbus_driver.
>
> I'm wondering whether it wouldn't better be the other way around: The
> (hopefully few) essential ones get flagged, thus also making it more
> prominent during patch review that a flag gets added (and justification
> provided), instead of having to spot the lack of a flag getting set.

Not flagging a non-essential one is less problematic than not flagging
an essential driver IMO.

For some drivers I'm on the edge, BTW. The pv 9pfs driver ought to be
non-essential in most cases, but there might be use cases where it is
needed, so I didn't set its non_essential flag.

Same applies to pv-usb and maybe pv-scsi, while pv-tpm probably really
is essential.

With the current series I'm ending up with 6 non-essential drivers and
6 essential ones, so either way needs the same number of drivers
modified.


Juergen
```

## Attachments

- [OpenPGP\_signature](https://ratatoskr.run/linux-input/2021/10/1550490/attachment/0) (application/pgp-signature, 495 bytes)
