Re: designing fast, streaming disk i/o with mmap: help wanted

From: Roger Larsson (roger.larsson@norran.net)
Date: Sat Apr 01 2000 - 10:51:47 EST


Hi again,

Now I have poked around some more in the current read-ahead code.
Both normal and mmaped.

This is how I think it best can be used for this specific application:

* Let the butler thread read big bursts from alternating channels, i.e.
  not trying to keep all channels with equal read-ahead at all times.

Suppose there are n channels.

Init: let the butler read at least BUTLER_READ_AHEAD from all channels.

Cycle channels (i for all n): read BUTLER_READ_AHEAD from channel[i]

Master reads cached from all channels every cycle (Tmaster ms).

  BUTLER_READ_AHEAD * Tbutler >= MASTER_READ * n * Tmaster

/RogerL

-
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 : Fri Apr 07 2000 - 21:00:07 EST