PERR/SERR

NTDEV
1

Using a PCI Bus Analyser, a PERR/SERR ( both ) is getting triggered on a
specific address ( belongs to memory space of a specific adapter ). NMI
BSOD was seen in the target machine when this occurred.

Does a device driver has the potential to cause this kind of PERR/SERR
?? How are these issues resolved generally ?? Any expert comments ??

Thanks,
BJ

2

Assuming this Specific address is actually valid for the device, then the
driver is not at fault.

SERR/PERR are indications that the hardware thinks something has gone
wrong. PERR, I beleive means that the parity has gone wrong on the PCI bus
itself. This shouldn’t happen, and is an indication that the hardware is
faulty (stuck pin?) or that there is a general problem with for instance
noise/signal shape/clock stability in the hardware.

The driver will have a hard time getting into this state, unless it’s
actually accessing something that the hardware dislikes. Of course, it’s
technically possible to design hardware that pulls PERR and/or SERR for
some other reason than the “standard” reason, say for instance if you try
to write a one to a reserved bitfield in register X, the device may decide
that this is so bad that we want to blow the system up. It’s not a nice way
to indicate that the device isn’t happy, but it would be possible to design
it in such a way.

–
Mats

xxxxx@lists.osr.com wrote on 10/06/2004 05:04:10 PM:

Using a PCI Bus Analyser, a PERR/SERR ( both ) is getting triggered
on a specific address ( belongs to memory space of a specific
adapter ). NMI BSOD was seen in the target machine when this occurred.

Does a device driver has the potential to cause this kind of
PERR/SERR ?? How are these issues resolved generally ?? Any expert
comments ??

Thanks,
BJ

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:NT00004B96