> Yes, please contact me at this address and heiko@colossus.escape.de.
> That is the most reliable way to get my attention.
Sorry. I am not at my home machine, so I didn't use the alias I normally
use for you. I will try to be more carefull...
>
> >will go through the cdu31a driver and make sure it complies with the docs
> >you sent me...
>
> Maybe I messed something up with the driver.
> I could reinvestigate the diff and try to break that
> into manageable pieces...
>
> Greetings Heiko
>
I don't think you did anything wrong. You made a legitimate workaround
for a limitation in Uniform. I will be fixing Uniform this weekend, so
the workaround can then be done the Right Way(tm).
I have a list of outstanding cdrom driver bugs on my box at home. Going
from memory, I think the outstanding bugs for cdu31a include:
1) When playing audio CDs it trys to play audio past the end of the CD.
2) When converting to the uniform cdrom driver, you implemented a workaround
for the spin-up code that fakes the drive status. Uniform needs to be
modified to add pre/post ioctl callbacks. The spin-up code needs to be
moved to the pre-ioctl function. The workaround that fakes the drive
status can then be removed.
3) cdu31a (and all other ported-to-uniform cdrom drivers) can give a kernel
Oops if they are removed while in use. To fix this, Uniform needs to be
modified to add pre/post ioctl callbacks, and the pre-ioctl function needs
to increment the module use count, and the post-ioctl function needs
to decrement the module use count.
4) I have had reports of "Ughs" from cdu31a. It apprears to be accessing
user_space with interrupts disabled. Since it is accessing user_space,
this means the Ugh is probably from within a call to scd_dev_ioctl() with
a CDROMREADAUDIO ioctl, since low-level cdrom drivers should only touch
user-space for device-specific ioctls.
I hope you don't mind me copying this part of our discussion to linux-kernel,
as I think this is of more general interest...
-Erik
--
Erik B. Andersen Web: http://www.inconnect.com/~andersen/
email: andersee@debian.org
--This message was written using 73% post-consumer electrons--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu