please explain ...

Tomohiro Kusumi kusumi.tomohiro at gmail.com
Thu Jul 19 07:09:09 PDT 2018


Aaron LI <aly at aaronly.me>:
>
> On Thu, 19 Jul 2018 15:16:46 +0200
> Holm Tiffe <holm at freibergnet.de> wrote:
>
> > Tomohiro Kusumi wrote:
> >
> > >  Aaron LI <aly at aaronly.me>:
> > > >
> > > > On Thu, 19 Jul 2018 11:39:06 +0200
> > > > Holm Tiffe <holm at freibergnet.de> wrote:
> > > >
> > > > Hi Holm,
> > > >
> > > > Thanks for trying DFly :)
> > > >
> > > > > Jul 19 11:12:34 fly kernel: [drm] fb depth is 24
> > > > > Jul 19 11:12:34 fly kernel: [drm]    pitch is 5888
> > > > > Jul 19 11:12:34 fly kernel: drm0: debug: Error during ACPI methods
> > > > > call Jul 19 11:12:34 fly kernel: [drm] Initialized radeon 2.40.0
> > > > > 20080528 Jul 19 11:12:34 fly kernel: md3: Malloc disk
> > > > > Jul 19 11:12:34 fly kernel: md4: Malloc disk
> > > > > Jul 19 11:12:35 fly kernel: md5: Malloc disk
> > > > > Jul 19 11:12:35 fly kernel: md6: Malloc disk
> > > > > Jul 19 11:12:35 fly kernel: md7: Malloc disk
> > > > > Jul 19 11:12:40 fly kernel: md8: Malloc disk
> > > > > Jul 19 11:12:50 fly kernel: md9: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md10: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md11: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md12: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md13: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md14: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md15: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md16: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md17: Malloc disk
> > > > > Jul 19 11:12:51 fly kernel: md18: Malloc disk
> > > > > Jul 19 11:12:57 fly kernel: pid 1305 (at-spi-bus-launc), uid 201:
> > > > > exited on signal 11
> > > > >
> > > > > ..what those inflationary appearing malloc disks are good for:
> > > >
> > > > The "md" device uses an "almost cloning" method, i.e., when /dev/md{N}
> > > > is opened, a /dev/md{N+1} device is auto created.
> > > >
> > > > I think there is something wrong with the md probe... but I don't have
> > > > more ideas here.
> > >
> > > This is a known behavior which originates from FreeBSD.
> > > https://github.com/freebsd/freebsd/commit/1624e20f39888e846831b8a8457c919d0cfa54c9
> > >
> > > I doubt anyone or any userspace program needs this behavior.
> >
> > Hmm.
> >
> > May be..but I'm using FreeBSD since before 2.0 and I never ever saw an
> > behavior like this.
> >
> > Yes I'm used malloc disks on this Notebook once, but I never created more
> > as two of them and I've deallocated them cleanly, there is no need to
> > create additional ones every time...
>
> Yes, I think you're having a problem with "md" here, but the information you
> provided here is not enough for me and other devs to identify the problem.

This behavior is just generally bothering and unneeded.

We do have code that works around this too.
https://github.com/DragonFlyBSD/DragonFlyBSD/blob/master/etc/autofs/special_media#L12

Above workaround code is actually just creating another bug, by not
supporting /dev/mdX, only to avoid creating a new /dev/md(X+1) for no
reason.
I may someday fix this, so that I can remove this XXX stuff which only
exists in DragonFly autofs implementation.

>
> Could you describe in more details how you used "md" here?  When the above
> "md" messages and errors happen?  etc.
>
>
> Cheers,
> --
> Aaron



More information about the Users mailing list