Re: kern/59165: vio9p(4): expose tag through sysctl or device properties

At Tue, 25 Nov 2025 17:39:19 -0800, "Greg A. Woods" <[email protected]> wrote:
Subject: Re: kern/59165: vio9p(4): expose tag through sysctl or device properties
>
> I see the code is there in sys/dev/pci/vio9p.c, but it doesn't seem to
> work for me:

Ooops!  It does work.  The wrong kernel was booted.  The right one was
installed, but not booted.  I should have seen that in the uname output,
but somehow it didn't click.

> $ uname -a
> NetBSD nbt2.local 10.99.12 NetBSD 10.99.12 (GENERIC) #1: Tue Apr  8 14:29:37 PDT 2025  [email protected]:/Users/woods/build/woods/very.local/trunk-x86_64-amd64-obj/Volumes/work/woods/g-NetBSD-src/sys/arch/amd64/compile/GENERIC amd64

An even newer kernel is now confirmed as booted:

$ uname -a
NetBSD nbt2.local 11.99.4 NetBSD 11.99.4 (GENERIC) #1: Thu Nov 27 10:58:12 PST 2025  [email protected]:/Users/woods/build/woods/very.local/trunk-x86_64-amd64-obj/Volumes/work/woods/g-NetBSD-src/sys/arch/amd64/compile/GENERIC amd64

$ /sbin/sysctl  hw.vio9p
hw.vio9p.vio9p0.tag = share


Sorry for the false report!


In other news though the vio9p device(s) doesn't show up as a disk, and
so its I/O stats don't get reported by systat:

 $ /sbin/sysctl  hw.disknames
hw.disknames = ld0 dk0 dk1 dk2 dk3 dk4 dk5

Is that (partly) because they are character devices?

During a sysinst upgrade with sets on the vio9p device I noted that the
read speed seemed to be around 12MB/s, which couldn't even keep the root
filesystem and disk an more than about 30% busy.

--
					Greg A. Woods <[email protected]>

Kelowna, BC     +1 250 762-7675           RoboHack <[email protected]>
Planix, Inc. <[email protected]>     Avoncote Farms <[email protected]>

-----BEGIN PGP SIGNATURE-----

iF0EABECAB0WIQRuK6dmwVAucmRxuh9mfXG3eL/0fwUCaSi07QAKCRBmfXG3eL/0
f2VuAJ9fHRRH7PdmjbDJ3NcQgSS8sphz1wCgm2bFfA3RHbf+w6QjrKsGQX7TOiE=
=dKsN
-----END PGP SIGNATURE-----