win2k hangup

HI, all

When I executed(? *^*, sorry) a DMA transferring(system memory -> local
memory), my win2k became hang up even including mouse and softice. Does
anybody can give me some advice about the windows’ hangup.

The PCI card was MPC8540, The driver had to map a PC’s PCI memory
space(Common buffer) to the local PCI memory space dynamically using the
8540’s . when the transferring size was less than 256byte, the transferring
can be finished correctly. but if the transferring size was larger than
256byte, windows became hangup.

Thanks.

This is not a trivial question to answer. The simple answer to your system
hanging is most likely a PCI hang, which is caused by the system ‘getting
confused’ on the PCI bus. This can be caused by a number of different
factors, and the most reasonable way to solve it is probably a PCI bus
analyzer (a machine that connects to the PCI bus and records all things
that happen on the PCI bus). At this point, it’s probably also good to have
a VERY THOROUGH understanding of the PCI device that you’re trying to talk
to.

The reason I think that it’s a PCI bus hang is that you say the mouse and
softice are locked up. Normally, if you have a system hang in a software
hang, the mouse and keyboard are usually working. However, on a PCI hang,
the next PCI access done by the processor will cause the processor to wait
for it to complete. At that time, nothing else will EVER happen in the
system, since the CPU is now stopped, and all peripherals that communicate
via the PCI bus are stopped, so nothing much CAN happen.


Mats

-------- Notice --------
The information in this message is confidential and may be legally
privileged. It is intended solely for the addressee. Access to this
message by anyone else is unauthorized. If you are not the intended
recipient, any disclosure, copying or distribution of the message, or any
action taken by you in reliance on it, is prohibited and may be unlawful.
If you have received this message in error, please delete it and contact
the sender immediately. Thank you.

xxxxx@lists.osr.com wrote on 12/15/2004 12:52:37 PM:

HI, all

When I executed(? *^*, sorry) a DMA transferring(system memory ->
local memory), my win2k became hang up even including mouse and
softice. Does anybody can give me some advice about the windows’ hangup.

The PCI card was MPC8540, The driver had to map a PC’s PCI memory
space(Common buffer) to the local PCI memory space dynamically using
the 8540’s . when the transferring size was less than 256byte, the
transferring can be finished correctly. but if the transferring
size was larger than 256byte, windows became hangup.

Thanks.

Questions? First check the Kernel Driver FAQ at http://www.
osronline.com/article.cfm?id=256

You are currently subscribed to ntdev as: unknown lmsubst tag argument:
‘’
To unsubscribe send a blank email to xxxxx@lists.osr.com
ForwardSourceID:NT00009836

It is not Windows per se, it is or PCI bus waiting for ready that is not
coming, or your driver sitting in endless loop at high IRQL.

-----Original Message-----
From: Ryu [mailto:xxxxx@dcc.cpg.sony.co.jp]
Sent: Wednesday, December 15, 2004 2:53 PM
To: Windows System Software Devs Interest List
Subject: [ntdev] win2k hangup

HI, all

When I executed(? *^*, sorry) a DMA transferring(system memory -> local
memory), my win2k became hang up even including mouse and softice. Does
anybody can give me some advice about the windows’ hangup.

The PCI card was MPC8540, The driver had to map a PC’s PCI memory
space(Common buffer) to the local PCI memory space dynamically using the
8540’s . when the transferring size was less than 256byte, the transferring
can be finished correctly. but if the transferring size was larger than
256byte, windows became hangup.

Thanks.

Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256

You are currently subscribed to ntdev as: unknown lmsubst tag argument: ‘’
To unsubscribe send a blank email to xxxxx@lists.osr.com

Alexander Krol wrote:

It is not Windows per se, it is or PCI bus waiting for ready that is not
coming, or your driver sitting in endless loop at high IRQL.

or a million other possibilities (an errant DMA transfer springs
immediately to mind).

I suppose checking the system state via an ICE or bus analyzer isn’t
possible (or you would have told us this information already)?

Do you have the ability to crash your system during the hang (this
requires either special hardware support or the willingness/stupidity to
short two specific pins on the PCI bus together… If you have to ask
what I’m talking about, the answer would be “no”) and find-out what’s
going on via a crash dump? [I’m writing here for the archive, not necessarily just for the OP]

During a hard hang, such hardware-based remedies are usually the only
alternative.

Peter
OSR

Mats

Thank you very much.

I can transfer data from MPC8540DDR to system memory, but can not transfer
data from system memory to local DDR. Maybe the system’ memory is faster
than local DDR. So I can read data from the device, but cannot write data to
it.

Do you think so?

----- Original Message -----
From: “Mats PETERSSON”
To: “Windows System Software Devs Interest List”
Sent: Wednesday, December 15, 2004 10:23 PM
Subject: Re: [ntdev] win2k hangup

>
> This is not a trivial question to answer. The simple answer to your system
> hanging is most likely a PCI hang, which is caused by the system ‘getting
> confused’ on the PCI bus. This can be caused by a number of different
> factors, and the most reasonable way to solve it is probably a PCI bus
> analyzer (a machine that connects to the PCI bus and records all things
> that happen on the PCI bus). At this point, it’s probably also good to
have
> a VERY THOROUGH understanding of the PCI device that you’re trying to talk
> to.
>
> The reason I think that it’s a PCI bus hang is that you say the mouse and
> softice are locked up. Normally, if you have a system hang in a software
> hang, the mouse and keyboard are usually working. However, on a PCI hang,
> the next PCI access done by the processor will cause the processor to wait
> for it to complete. At that time, nothing else will EVER happen in the
> system, since the CPU is now stopped, and all peripherals that communicate
> via the PCI bus are stopped, so nothing much CAN happen.
>
> –
> Mats
>
> -------- Notice --------
> The information in this message is confidential and may be legally
> privileged. It is intended solely for the addressee. Access to this
> message by anyone else is unauthorized. If you are not the intended
> recipient, any disclosure, copying or distribution of the message, or any
> action taken by you in reliance on it, is prohibited and may be unlawful.
> If you have received this message in error, please delete it and contact
> the sender immediately. Thank you.
>
>
> xxxxx@lists.osr.com wrote on 12/15/2004 12:52:37 PM:
>
> > HI, all
> >
> > When I executed(? ^, sorry) a DMA transferring(system memory ->
> > local memory), my win2k became hang up even including mouse and
> > softice. Does anybody can give me some advice about the windows’ hangup.
> >
> > The PCI card was MPC8540, The driver had to map a PC’s PCI memory
> > space(Common buffer) to the local PCI memory space dynamically using
> > the 8540’s . when the transferring size was less than 256byte, the
> > transferring can be finished correctly. but if the transferring
> > size was larger than 256byte, windows became hangup.
> >
> > Thanks.
> > —
> > Questions? First check the Kernel Driver FAQ at http://www.
> > osronline.com/article.cfm?id=256
> >
> > You are currently subscribed to ntdev as: unknown lmsubst tag argument:
> ‘’
> > To unsubscribe send a blank email to xxxxx@lists.osr.com
> > ForwardSourceID:NT00009836
>
>
> —
> Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@dcc.cpg.sony.co.jp
> To unsubscribe send a blank email to xxxxx@lists.osr.com

I changed the buffer size of DMA controller(the register is not introduced
in manual) from 256byte into 64 byte. Then the DMA transferring can be done.
But I don’t know the reason.

Thanks all.

From now on , I will t some reason
----- Original Message -----
From: “Ryu”
To: “Windows System Software Devs Interest List”
Sent: Friday, December 17, 2004 4:44 PM
Subject: Re: [ntdev] win2k hangup

> Mats
>
> Thank you very much.
>
> I can transfer data from MPC8540DDR to system memory, but can not transfer
> data from system memory to local DDR. Maybe the system’ memory is faster
> than local DDR. So I can read data from the device, but cannot write data
to
> it.
>
> Do you think so?
>
> ----- Original Message -----
> From: “Mats PETERSSON”
> To: “Windows System Software Devs Interest List”
> Sent: Wednesday, December 15, 2004 10:23 PM
> Subject: Re: [ntdev] win2k hangup
>
>
>
> >
> > This is not a trivial question to answer. The simple answer to your
system
> > hanging is most likely a PCI hang, which is caused by the system
‘getting
> > confused’ on the PCI bus. This can be caused by a number of different
> > factors, and the most reasonable way to solve it is probably a PCI bus
> > analyzer (a machine that connects to the PCI bus and records all things
> > that happen on the PCI bus). At this point, it’s probably also good to
> have
> > a VERY THOROUGH understanding of the PCI device that you’re trying to
talk
> > to.
> >
> > The reason I think that it’s a PCI bus hang is that you say the mouse
and
> > softice are locked up. Normally, if you have a system hang in a software
> > hang, the mouse and keyboard are usually working. However, on a PCI
hang,
> > the next PCI access done by the processor will cause the processor to
wait
> > for it to complete. At that time, nothing else will EVER happen in the
> > system, since the CPU is now stopped, and all peripherals that
communicate
> > via the PCI bus are stopped, so nothing much CAN happen.
> >
> > –
> > Mats
> >
> > -------- Notice --------
> > The information in this message is confidential and may be legally
> > privileged. It is intended solely for the addressee. Access to this
> > message by anyone else is unauthorized. If you are not the intended
> > recipient, any disclosure, copying or distribution of the message, or
any
> > action taken by you in reliance on it, is prohibited and may be
unlawful.
> > If you have received this message in error, please delete it and contact
> > the sender immediately. Thank you.
> >
> >
> > xxxxx@lists.osr.com wrote on 12/15/2004 12:52:37 PM:
> >
> > > HI, all
> > >
> > > When I executed(? ^, sorry) a DMA transferring(system memory ->
> > > local memory), my win2k became hang up even including mouse and
> > > softice. Does anybody can give me some advice about the windows’
hangup.
> > >
> > > The PCI card was MPC8540, The driver had to map a PC’s PCI memory
> > > space(Common buffer) to the local PCI memory space dynamically using
> > > the 8540’s . when the transferring size was less than 256byte, the
> > > transferring can be finished correctly. but if the transferring
> > > size was larger than 256byte, windows became hangup.
> > >
> > > Thanks.
> > > —
> > > Questions? First check the Kernel Driver FAQ at http://www.
> > > osronline.com/article.cfm?id=256
> > >
> > > You are currently subscribed to ntdev as: unknown lmsubst tag
argument:
> > ‘’
> > > To unsubscribe send a blank email to xxxxx@lists.osr.com
> > > ForwardSourceID:NT00009836
> >
> >
> > —
> > Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
> >
> > You are currently subscribed to ntdev as: xxxxx@dcc.cpg.sony.co.jp
> > To unsubscribe send a blank email to xxxxx@lists.osr.com
>
>
> —
> Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@dcc.cpg.sony.co.jp
> To unsubscribe send a blank email to xxxxx@lists.osr.com

Not sure what the reason for your problems would be, but I can guarantee
that the speed of memory on either side of the PCI bus has NOTHING to do
with your problems (at least not directly). It’s technically possible for
the PCI bus to operate correctly with memory that has several microseconds
latency. It just won’t do it very quickly…

One thing I wonder about tho’, is the fact that it’s the magical 256 bytes
per transfer. Is it possible that the 256 bytes becomes a zero-byte
transfer for some reason, and thus causes the PCI device to be “confused”
and not actually transfer things the way it should. If the register for the
size of transfer is 8 bit, this would happen. Or if the write to the
register is done with an 8-bit operation.

Other than that, I haven’t really got any ideas.


Mats

I changed the buffer size of DMA controller(the register is not
introduced
in manual) from 256byte into 64 byte. Then the DMA transferring can be
done.
But I don’t know the reason.

Thanks all.

From now on , I will t some reason
----- Original Message -----
From: “Ryu”
> To: “Windows System Software Devs Interest List”
> Sent: Friday, December 17, 2004 4:44 PM
> Subject: Re: [ntdev] win2k hangup
>
>
> > Mats
> >
> > Thank you very much.
> >
> > I can transfer data from MPC8540DDR to system memory, but can not
transfer
> > data from system memory to local DDR. Maybe the system’ memory is
faster
> > than local DDR. So I can read data from the device, but cannot write
data
> to
> > it.
> >
> > Do you think so?
> >
> > ----- Original Message -----
> > From: “Mats PETERSSON”
> > To: “Windows System Software Devs Interest List”
> > Sent: Wednesday, December 15, 2004 10:23 PM
> > Subject: Re: [ntdev] win2k hangup
> >
> >
> >
> > >
> > > This is not a trivial question to answer. The simple answer to your
> system
> > > hanging is most likely a PCI hang, which is caused by the system
> ‘getting
> > > confused’ on the PCI bus. This can be caused by a number of different
> > > factors, and the most reasonable way to solve it is probably a PCI
bus
> > > analyzer (a machine that connects to the PCI bus and records all
things
> > > that happen on the PCI bus). At this point, it’s probably also good
to
> > have
> > > a VERY THOROUGH understanding of the PCI device that you’re trying to
> talk
> > > to.
> > >
> > > The reason I think that it’s a PCI bus hang is that you say the mouse
> and
> > > softice are locked up. Normally, if you have a system hang in a
software
> > > hang, the mouse and keyboard are usually working. However, on a PCI
> hang,
> > > the next PCI access done by the processor will cause the processor to
> wait
> > > for it to complete. At that time, nothing else will EVER happen in
the
> > > system, since the CPU is now stopped, and all peripherals that
> communicate
> > > via the PCI bus are stopped, so nothing much CAN happen.
> > >
> > > –
> > > Mats
> > >
> > > -------- Notice --------
> > > The information in this message is confidential and may be legally
> > > privileged. It is intended solely for the addressee. Access to this
> > > message by anyone else is unauthorized. If you are not the intended
> > > recipient, any disclosure, copying or distribution of the message, or
> any
> > > action taken by you in reliance on it, is prohibited and may be
> unlawful.
> > > If you have received this message in error, please delete it and
contact
> > > the sender immediately. Thank you.
> > >
> > >
> > > xxxxx@lists.osr.com wrote on 12/15/2004 12:52:37 PM:
> > >
> > > > HI, all
> > > >
> > > > When I executed(? ^, sorry) a DMA transferring(system memory ->
> > > > local memory), my win2k became hang up even including mouse and
> > > > softice. Does anybody can give me some advice about the windows’
> hangup.
> > > >
> > > > The PCI card was MPC8540, The driver had to map a PC’s PCI memory
> > > > space(Common buffer) to the local PCI memory space dynamically
using
> > > > the 8540’s . when the transferring size was less than 256byte, the
> > > > transferring can be finished correctly. but if the transferring
> > > > size was larger than 256byte, windows became hangup.
> > > >
> > > > Thanks.
> > > > —
> > > > Questions? First check the Kernel Driver FAQ at http://www.
> > > > osronline.com/article.cfm?id=256
> > > >
> > > > You are currently subscribed to ntdev as: unknown lmsubst tag
> argument:
> > > ‘’
> > > > To unsubscribe send a blank email to
xxxxx@lists.osr.com
> > > > ForwardSourceID:NT00009836
> > >
> > >
> > > —
> > > Questions? First check the Kernel Driver FAQ at
> > http://www.osronline.com/article.cfm?id=256
> > >
> > > You are currently subscribed to ntdev as: xxxxx@dcc.cpg.sony.co.jp
> > > To unsubscribe send a blank email to xxxxx@lists.osr.com
> >
> >
> > —
> > Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
> >
> > You are currently subscribed to ntdev as: xxxxx@dcc.cpg.sony.co.jp
> > To unsubscribe send a blank email to xxxxx@lists.osr.com
>
>
> —
> Questions? First check the Kernel Driver FAQ at http://www.
> osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@3dlabs.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com

> ForwardSourceID:NT00009A8E