From: Halil Pasic <pasic@linux.ibm.com>
To: Cornelia Huck <cohuck@redhat.com>
Cc: "Jason J . Herne" <jjherne@linux.ibm.com>,
linux-s390@vger.kernel.org, Eric Farman <farman@linux.ibm.com>,
Alex Williamson <alex.williamson@redhat.com>,
kvm@vger.kernel.org, Pierre Morel <pmorel@linux.ibm.com>,
Farhan Ali <alifm@linux.ibm.com>,
qemu-devel@nongnu.org, qemu-s390x@nongnu.org
Subject: Re: [PATCH 0/3] vfio-ccw: support hsch/csch (kernel part)
Date: Fri, 7 Dec 2018 16:49:09 +0100 [thread overview]
Message-ID: <20181207164909.66abc005@oc2783563651> (raw)
In-Reply-To: <20181207110529.3b293124.cohuck@redhat.com>
On Fri, 7 Dec 2018 11:05:29 +0100
Cornelia Huck <cohuck@redhat.com> wrote:
[..]
> > To clarify my concern let me quote from the PoP
> > (SA22-7832-10 page 14-9):
> >
> > """
> > If a device presents unsolicited status while the
> > associated subchannel is disabled, that status is
> > discarded by the channel subsystem without
> > generating an I/O-interruption condition. How-
> > ever, if the status presented contains unit check,
> > the channel subsystem issues the clear signal for
> > the associated subchannel and does not gener-
> > ate an I/O-interruption condition. This should be
> > taken into account when the program uses MOD-
> > IFY SUBCHANNEL to enable a subchannel. For
> > example, the medium on the associated device
> > that was present when the subchannel became
> > disabled may have been replaced, and, there-
> > fore, the program should verify the integrity of
> > that medium.
> > """
>
> Hm, so is your concern that we might have a status (unit check) if we
> have an enabled subchannel that might not be present if the subchannel
> had been disabled all the time? Is that a problem in practice?
>
No idea if it is a problem in practice.
> > > >
> > > > I think Jason has discovered some problems related to this while
> > > > doing his DASD IPL with vfio-ccw work, but I don't quite remember
> > > > any more.
> > >
> > > cc:ing Jason, in case he remembers :)
>
> Like in that case. Couldn't a unit check status also arrive just when
> the subchannel has been enabled, and the code therefore has to deal
> with it anyway?
>
I assumed that programming note is there for a reason. Of course if
it can not been proven it ain't cheating. I don't remember exactly this
interacts with the rest of the architecture. In fact I asked my question,
because my feeling was that tying the virtual an the backing subchannel
together is simpler, than proving that we are fine without doing it.
> > >
> > > > IMHO it may be possible to emulate enable/disable, but it seems way
> > > > more error prone and complicated, than letting the guest
> > > > enable/disable the host subchannel.
> > > >
> > > > I have no idea what was the reason for going with the initial design.
> > > > I would appreciate any hints or explanations, but I'm well aware
> > > > that it was a long time ago.
> > >
> > > I don't really remember either, and any non-public mails from that time
> > > are inaccessible to me :(
> > >
> > > It *might* be an artifact of the original design (which operated at the
> > > ccw_device rather than the subchannel level), though.
> > >
> >
> > Interesting.
> >
> > > > > Parameters (like for channel measurements) are a different game.
> > > > > It is something we should look into, but it will need a different
> > > > > region.
> > > >
> > > > Yes emulation only channel measurements seem even less likely than
> > > > proper enable/disable. And 'that would need a different' region
> > > > helps me understanding the scope of async_cmd_region. Maybe we
> > > > should reconsider the comment '+ * @cmd_region: MMIO region for
> > > > asynchronous I/O commands other than START'.
> > >
> > > What do you think is wrong with that comment?
> > >
> >
> > Well msch is also an async I/O command other than START. If msch does not
> > belong here but needs it's own region, then this description seems too
> > generic.
>
> Why do you consider msch to be async? ssch, hsch, csch all have the
> potential to cause the execution of an asynchronous (start/halt/clear)
> function, while msch just (possibly) modifies the subchannel and is
> done.
>
Right, my bad. Got confused by my Z channel io is async superstition. I
did not quite understand what async means in this context.
Regards,
Halil
next prev parent reply other threads:[~2018-12-07 15:49 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-22 16:54 [PATCH 0/3] vfio-ccw: support hsch/csch (kernel part) Cornelia Huck
2018-11-22 16:54 ` [PATCH 1/3] vfio-ccw: add capabilities chain Cornelia Huck
2018-11-23 12:28 ` Pierre Morel
2018-11-23 12:45 ` Cornelia Huck
2018-11-23 13:26 ` Pierre Morel
2018-11-27 19:04 ` Farhan Ali
2018-11-28 9:05 ` Cornelia Huck
2018-12-17 21:53 ` Eric Farman
2018-12-18 17:24 ` Cornelia Huck
2018-12-18 17:56 ` Eric Farman
2018-12-19 16:28 ` Alex Williamson
2018-12-21 11:12 ` Cornelia Huck
2018-11-22 16:54 ` [PATCH 2/3] s390/cio: export hsch to modules Cornelia Huck
2018-11-23 12:30 ` Pierre Morel
2018-11-22 16:54 ` [PATCH 3/3] vfio-ccw: add handling for asnyc channel instructions Cornelia Huck
2018-11-23 13:08 ` Pierre Morel
2018-11-26 9:47 ` Cornelia Huck
2018-11-27 19:09 ` Farhan Ali
2018-11-28 9:02 ` Cornelia Huck
2018-11-28 14:31 ` Farhan Ali
2018-11-28 14:52 ` Cornelia Huck
2018-11-28 15:00 ` Farhan Ali
2018-11-28 15:35 ` Cornelia Huck
2018-11-28 15:55 ` Farhan Ali
2019-01-18 13:53 ` Cornelia Huck
2018-11-27 19:57 ` Farhan Ali
2018-11-28 8:41 ` Cornelia Huck
2018-11-28 16:36 ` [qemu-s390x] " Halil Pasic
2018-11-29 16:52 ` Cornelia Huck
2018-11-29 17:24 ` Halil Pasic
2018-12-17 21:54 ` Eric Farman
2018-12-18 16:45 ` Cornelia Huck
2018-11-24 21:07 ` [qemu-s390x] [PATCH 0/3] vfio-ccw: support hsch/csch (kernel part) Halil Pasic
2018-11-26 9:26 ` Cornelia Huck
2018-11-26 18:57 ` Farhan Ali
2018-11-26 19:00 ` Cornelia Huck
2018-12-04 12:38 ` Halil Pasic
2018-12-04 13:11 ` Cornelia Huck
2018-12-04 15:02 ` Halil Pasic
2018-12-05 12:54 ` Cornelia Huck
2018-12-05 18:34 ` Farhan Ali
2018-12-06 14:39 ` Cornelia Huck
2018-12-06 15:26 ` Farhan Ali
2018-12-06 16:21 ` Cornelia Huck
2018-12-06 17:50 ` Farhan Ali
2018-12-07 9:34 ` Cornelia Huck
2018-12-06 18:47 ` Halil Pasic
2018-12-07 10:05 ` Cornelia Huck
2018-12-07 15:49 ` Halil Pasic [this message]
2018-12-07 16:54 ` Halil Pasic
2018-12-19 11:54 ` Cornelia Huck
2018-12-19 14:17 ` Halil Pasic
2018-12-21 11:23 ` Cornelia Huck
2018-12-21 12:42 ` Halil Pasic
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20181207164909.66abc005@oc2783563651 \
--to=pasic@linux.ibm.com \
--cc=alex.williamson@redhat.com \
--cc=alifm@linux.ibm.com \
--cc=cohuck@redhat.com \
--cc=farman@linux.ibm.com \
--cc=jjherne@linux.ibm.com \
--cc=kvm@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=pmorel@linux.ibm.com \
--cc=qemu-devel@nongnu.org \
--cc=qemu-s390x@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).