In ll_rw_block(), make_request(), there is a section od code that does
this:
case WRITERAW:
case WRITE:
... refile_buffer(bh);
This code does and will break anyone performing Asynch IO with their on
buffer heads because it will take your local buffer head request, and
try to file it in the buffer cache internal lists, and corrupt system
memory. Using WRITERAW does work, however, there was no documentation
indicating that this had changed, and that the buffer cache was now
totally incestuous with the drivers (excuse me -- living in sin).
This either needs to be documented or fixed so folks who are calling
ll_rw_blk() with their own buffer heads don't spend two days chasing
down bugs introduced into Linux 2.4. We need to have a simple and clean
AIO interface that is not totally incestuous with the buffer cache.
Jeff
-
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 - 21:00:25 EST