>Yes. If you see how to do it - patches are welcome. One of the possible
>ways is to create a structure when you register driver, remove it upon
>rmmod, replace ->b_dev and friends with pointers to that structure and
>make the removal of this beast invalidate the buffers.
Wouldn't it be easyer to add a "has_safe_mediachange:" method to
block_device_operations, which would reply 1 for devices that can
safely hang on to their cached data between lastcloses, and 0 for
those where it is problematic. And if the method is not defined at
all, assume 0 and play it safe (thus removing the need to update all
drivers all at once).
That way, slow devices that do not have partitions, and do have
reliable disk change detection (or a fallback mechanism a la
FD_BROKEN_DCL) are not penalized.
Regards,
Alain
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Oct 15 2000 - 21:00:27 EST