Hi,
On Thu, 17 Feb 2000 08:42:58 -0800 (PST), John Alvord <jalvo@mbay.net>
said:
>> shrink_mmap() is the obvious place. The biggest problem is that we will
>> need to export enough functionality to allow other filesystems to
>> hook their own data structures onto the page cache LRU list, so that
>> filesystem-specific caches can be reclaimed in the same shrink_mmap()
>> loop.
> You are going to run into some actual scheduling conflicts. A relatively
> low CPU/memory program can generate I/O which can flood an I/O channel. To
> avoid the flood, that program will have to be suspended... and that might
> result in a high idle time. The schedular must know not to run that
> program and that the idle time is for the "best".
That's not an issue. Whenever we suspend processes because of lack of
free pages or due to write throttling, those processes are put into
sleep state and are not considered for running by the scheduler. Waking
these processes up when the resources become available is the job of
whatever is managing that resource --- it's not a task scheduling
problem.
--Stephen
-
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 : Wed Feb 23 2000 - 21:00:21 EST