From: Carsten Schiers <carsten@schiers.de>
To: mdlabriola@yahoo.com, xen-devel <xen-devel@lists.xensource.com>
Subject: AW: AW: 3rd party drivers in domU or dom0
Date: Wed, 15 Apr 2009 15:37:48 +0200 [thread overview]
Message-ID: <4824957.131239802668173.JavaMail.root@uhura> (raw)
Good. The patch is reverting some functionality that I only understood quite
briefly. I think it's on devices that provide more than one function (e.g.
video 01:01.0 audio 01:01.1) and that in order to do a reset on one of them.
Somehow there is a restriction that all PCI devices that are behind one
bridge need to be assigned to the same domain. For consumer boards, nearly
all of them have all PCI slots behind one bridge. I am also not sure, whether
this is only for HVM, whereas we speak about PV here.
So I think it's a step back and I use the mentioned patch. All this PCI stuff
is quite nifty (and a major difference compared to other solutions), but I
haven't seen that much details on it from a pure user perspective (and for
REALY understanding hte code, I am too stupid).
One open question I have: is it realy necessary that PCI devices sharing a GSI
have to be co-assigned? (Just in case one of the gurus knows and will enlighten
us here ;o) )
BR,
Carsten-
----- Originalnachricht -----
Von: mdlabriola@yahoo.com
Gesendet: Mit, 15.4.2009 14:30
An: Carsten Schiers <carsten@schiers.de> ; xen-devel <xen-devel@lists.xensource.com>
Betreff: Re: AW: [Xen-devel] 3rd party drivers in domU or dom0
Yeah, that's the problem. I tried adding all the other mentioned PCI devices to the domU as well, but then I got some other error message (I'm away from my notebook right now). It was something about memory boundaries not aligning, if memory serves...
-Mike
------Original Message------
From: Carsten Schiers
To: Michael Labriola
To: xen-devel
Subject: AW: [Xen-devel] 3rd party drivers in domU or dom0
Sent: Apr 15, 2009 08:19
Despite what Keir said on your timing requirements, I guess you get the famous
co-assignment message. See at the end of this:
http://www.nabble.com/xen-3.3.0-pv-pci-passthrough-co-assigned-problem-td20008460.html
BR,
Carsten.
----- Originalnachricht -----
Von: mdlabriola@yahoo.com
Gesendet: Mit, 15.4.2009 13:56
An: xen-devel <xen-devel@lists.xensource.com>
Betreff: [Xen-devel] 3rd party drivers in domU or dom0
I'm trying to use Xen 3.3.1 w/ the xen-linux mucurial kernel along with the Sabtech Swift PCI NTDS driver. It's a 3rd party module that I have the source for, although I can't remember what the license says offhand (taints kernel as proprietary).
Basically I need either my dom0 or one of my domUs to be outputting NTDS data via 2 of these cards at 50 Hz each. Initial tests have proven that the driver still works when compiled for dom0, but it's not quite meeting our timing requirements. Most data is on time but we have a couple spikes over a 10 min run where 1 second's worth of data is taking around 1.03 seconds, if I remember correctly, and our spec only allows us to be .008 seconds late.
I tried to hide the PCI cards from dom0, to see what the behavior is like in a domU, but I can't get that to work either. The pciback new_slot/bind stuff appears to work (I don't get errors), but when I try to create my domU, I get errors listing other devices that must also be sent to domU (looks like everything on that PCI bus)
Do you guys think there's a better chance of getting these boards working on domU or dom0?
-Mike
---
Michael D Labriola
21 Rip Van Winkle Cir
Warwick, RI 02886
401-316-9844_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
---
Michael D Labriola
21 Rip Van Winkle Cir
Warwick, RI 02886
401-316-9844
reply other threads:[~2009-04-15 13:37 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=4824957.131239802668173.JavaMail.root@uhura \
--to=carsten@schiers.de \
--cc=mdlabriola@yahoo.com \
--cc=xen-devel@lists.xensource.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.