On 3 Dec 2002, Robert Collins wrote:
> This appears to be inconsistent across fs types. Sigh. I'll just let
> everything complete, at a worst case of what, 4Kb, it's not a great
> overhead, and it leaves everything simpler while I get rid of the huge
> wads of duplicate code.
Actually it is quite consistent at the moment.. initiated reads are
allowed to complete and is then cancelled before copying the data into the
supplied buffer
The aufs store has lots of code to deal with aborted reads but most is
actually disabled at the moment.
Regards
Henrik
Received on Tue Dec 03 2002 - 00:03:19 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:19:00 MST