Re: linux-kernel-digest V1 #4192

Paul Barton-Davis (pbd@Op.Net)
Sat, 24 Jul 1999 00:37:58 -0400


>> disk logging and the gui displaying progress charts, and the tcp are all
>> logically disconnected systems, the only necessary shared resources are
>> CPU time, and irqs, so they shouldn't have to affect each other, but they
>> are too closely coupled under Linux right now.
>
>
>Au contraire. Those systems are closely coupled on Linux because the
>micro-kernel experiment showed the consequences of separating them.

au contraire again. microkernels don't show the effect of decoupling
them, they show the effect of distinct implementations of each
subsystem. again, your example:

>Process 1 Process 2 Process 3
>read data from a RT pipe and read from the file Gnuplot
>write to disk on the FS do computation
> output to standard out
>
>The closely coupled OS will correctly note the shared data between
>the three processes for a major performance win.

relies on the assumption that the file is a data storage device. the
framebuffer is not a data storage device, and neither is the network
interface, and there is *no* reason for there to be coupling between
i/o to them and the disk writes. Here's a different example:

Task 1 Task 2 Task 3
generates audio samples and copies data to disk on FS does FFT and
copies the data to a DAC displays result

there is no reason for the output of task 1, task 2 or task 3 to be
coupled (by locks and other kernel synchronization mechanisms) other
than the fact that will all copy data from the same page(s). i'm
starting to sound like a broken record here.

--p

-
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/