On Sat, 15 Apr 2000, Jens Axboe wrote:
> On Sat, Apr 15 2000, Gérard Roudier wrote:
> > > Trying to do a DVD_AUTH ioctl I get a SCSI command timeout and after
> > > that a bus reset (see below) beginning with kernel 2.3.99-pre1 through
> > > -pre5 . It works without a problem in kernel 2.2.14 and 2.3.51.
> >
> > What has changed in recent linux SCSI (and PCI related) is that the kernel
> > allows applications to provide the data direction and requires low-level
> > drivers to check that data is transferred in the right direction. The
> > ncr53c8xx driver may lock up when actual data direction is wrong, but the
> > sym53c8xx will gracefully eat the wrong data, throw them away, and report
> > an error.
>
> The command issues is REPORT_KEY and the data direction has been set
> by the Uniform CD-ROM layer. So I don't think that is the problem.
Some direction has been set but the source says it is set to
CGC_DIRECTION_WRITE and in the situation that seems to match the command
that failed (DVD_LU_SEND_ASF) the code seems to read information from some
returned data. This looks pretty inconsistent to me.
On the other hand, I donnot seem to understand how the adaptec driver
ensures that data will not go in the wrong direction. But may-be I am just
also incompetent for this code.:)
Gérard.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Apr 15 2000 - 21:00:26 EST