Just for clarification. The behavior I’m seeing does not appear on
w2k. Also, I’ve never observed the PC attempting to become root on its
own. I need the PC to become root because I need root to by cycle
master capable and currently my device is not cycle master capable.
IMHO, the PC drivers “should” be at least isochronous
resource manager capable in which case they should ensure that root is
cycle master capable. However, I have never observed the PC doing
this, even in w2k where it would set the contender bit in the PC’s
self-id packet (indicating that it was isochronous resource master
capable).
If there is no other solution, I don’t think that it is too big a job
to make my
device cycle master capable but this is only a partial solution. Our
device is sometimes run with a combination of hubs and repeaters which
are also not cycle master capable and, left to themselves, tend to
become root as they tend to be in the middle of the bus.
On the subject of the PC setting the contender bit in the self-id
packet and taking on the role of IRM. I have recently noticed that our
2 laptops (different makes) both set the contender bit even though
they are running XP-SP2. All other installations of XP-SP2 that I have
observed do not set the contender bit. I haven’t yet checked to see if
the laptops are taking on the IRM role but I suspect not.
Robert,
I have seen this multiple bus reset behavior of Windows since w2k. That
happened at least with all the devices I have developed drivers for. Like
you mentioned, I remember it is at attempt by the PC node to forcefully
become the root node. You are right when you say it is not a side-effect of
your driver. Why do you want to become root? For testing purposes?
Mat
Robert Newton
I think you must receive 2 bus reset notifications - one is “reset
started”, another is “reset complete”.
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
----- Original Message -----
From: “Robert Newton”
To: “Windows System Software Devs Interest List”
Sent: Monday, December 06, 2004 7:46 PM
Subject: [ntdev] 1394 bus reset and XP-SP2`
> Calling all 1394 gurus,
>
> I’m having difficulty with our 1394a device and XP-SP2. I’ve noticed
> some very bizarre bus reset behavior. Every bus reset (whether
> hardware or software generated) causes a second bus reset. This echo
> is extremely annoying as it tends to defeat my attempts to make the
> host node root.
>
> I observe this behavior even after cutting all the interesting code
> out of my bus reset notification routine so I am very sure it is not
> due to anything my driver is doing. I have never observed this kind of
> behavior with win2k. I don’t think there is a problem with XP < sp2.
> I haven’t directly verified this but I wasn’t experiencing problems
> until sp2 (this was not the only issue).
>
> Has anyone else observed this behavior? Does anyone have any idea what
> could be causing it? Could my device be doing something in it’s
> responses to queries from the OS that could cause this behavior? All
> ideas gratefully recieved.
>
> Robert Newton
> VX Technologies
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@storagecraft.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
From what I know on Windows 1394 stack (note: my knowledge was about
Win98SE and w2k betas) - PC is always IRM capable.
If PC is elected an IRM - then all IRM register accesses from the bus side
are automatically executed by 1394bus, so are the PC’s software accesses to the
same bandwidth-channel registers.
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
----- Original Message -----
From: “Robert Newton”
To: “Windows System Software Devs Interest List”
Sent: Monday, December 06, 2004 9:55 PM
Subject: Re[2]: [ntdev] 1394 bus reset and XP-SP2`
> Just for clarification. The behavior I’m seeing does not appear on
> w2k. Also, I’ve never observed the PC attempting to become root on its
> own. I need the PC to become root because I need root to by cycle
> master capable and currently my device is not cycle master capable.
>
> IMHO, the PC drivers “should” be at least isochronous
> resource manager capable in which case they should ensure that root is
> cycle master capable. However, I have never observed the PC doing
> this, even in w2k where it would set the contender bit in the PC’s
> self-id packet (indicating that it was isochronous resource master
> capable).
>
> If there is no other solution, I don’t think that it is too big a job
> to make my
> device cycle master capable but this is only a partial solution. Our
> device is sometimes run with a combination of hubs and repeaters which
> are also not cycle master capable and, left to themselves, tend to
> become root as they tend to be in the middle of the bus.
>
> On the subject of the PC setting the contender bit in the self-id
> packet and taking on the role of IRM. I have recently noticed that our
> 2 laptops (different makes) both set the contender bit even though
> they are running XP-SP2. All other installations of XP-SP2 that I have
> observed do not set the contender bit. I haven’t yet checked to see if
> the laptops are taking on the IRM role but I suspect not.
>
> > Robert,
> >
> > I have seen this multiple bus reset behavior of Windows since w2k. That
> > happened at least with all the devices I have developed drivers for. Like
> > you mentioned, I remember it is at attempt by the PC node to forcefully
> > become the root node. You are right when you say it is not a side-effect
of
> > your driver. Why do you want to become root? For testing purposes?
> >
> > Mat
>
>
> Robert Newton
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@storagecraft.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
Maxim S. Shatskih wrote:
From what I know on Windows 1394 stack (note: my knowledge was about
Win98SE and w2k betas) - PC is always IRM capable.
XPSP2 (Bug?): But the PC does not set the Contender bit.
If I connect for example my video camcorder, it becomes IRM. But windows
still ignores all IRM registers. It still uses it’s own internal registers.
Therefore I believe that XP does not care about IRMs at all.
Uwe
If PC is elected an IRM - then all IRM register accesses from the bus side
are automatically executed by 1394bus, so are the PC’s software accesses to the
same bandwidth-channel registers.
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
> XPSP2 (Bug?): But the PC does not set the Contender bit.
Maybe a bug, it worked since Win98.
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com