On Sat, Jun 16, 2007 at 05:36:11PM -0400, Alan Stern wrote: > On Sat, 16 Jun 2007, Matthew Dharm wrote: > > > I've cut the log down to the parts I find interesting.... > > > > On Sat, Jun 16, 2007 at 05:08:04PM +0200, Miernik wrote: > > > scsi0: PCI error Interrupt at seqaddr = 0x7 > > > scsi0: Data Parity Error Detected during address or write data phase > > > > Here's the interesting part. The PCI error is being reported by the > > Adaptec controller, not the USB interface. The error is clearly coming > > from scsi0, which is the Adaptec. > > > > I have no idea why that is happening. But, it suggests to me that there is > > some sort of hardware problem at work. The USB code shouldn't be able to > > affect a real SCSI controller in any way. > > I suspect the PCI error is unrelated to the other ext3 problems. It > doesn't show up in the second dmesg log.
Maybe not, but it's still suspicious. PCI Data Parity Errors should,
basically, never happen in a properly functioning system.
For those unfamiliar with PCI, whoever is driving the bus electrically
(initiator for a write, target for a read) also drives a single bit of
parity (actually, 2 bits for 64-bit transfers). The 'reciever' of the data
is responsible for checking this and deciding what to do; 'ignore' is a
valid option, but not recommended.
I've seen maybe a half-dozen of this sort of error over the last 10 years,
where I spend most of my time working with prototype hardware that is prone
to errors. Most commonly, I see this sort of thing when I take hardware to
operating extremes, where the silicon I/O drive gates start to fail to
perform properly.
In the embedded world (where my day job is), a system would be yanked from
service for such an error, even if it otherwise appeared to be functioning.
It's a sign of something Very Bad(tm) happening.
It makes me very suspicious that something else could be wrong with that
box, even if it's unrelated to any USB failures.
Matt
--
Matthew Dharm Home: [EMAIL PROTECTED]
Maintainer, Linux USB Mass Storage Driver
Now payink attention, please. This is mouse. Click-click. Easy to
use, da? Now you try...
-- Pitr to Miranda
User Friendly, 10/11/1998
pgp2jaBUcuNNT.pgp
Description: PGP signature
------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/
_______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
