On Mon, 24 Jan 2000, Eric Werme USG wrote:
>
> The unmount frees all the cached pages, but at substantial cost in
> benchmark knowledge, privileges, structure, and cooperation with other
> processes trying to the the FS at the same time.
Yes, and needs superuser privileges (which in many cases is not acceptable
for a benchmark).
I could certainly add a "flush cache" thing, I just do not think that
MS_INVALIDATE is the right thing to do (let's say that you have an
application that uses MS_INVALIDATE for the cache coherency reason, then
it's a good thing to make it as quick as the OS can).
We have "flush cache" ioctl's for block device accesses, where due to
partition table changes our caches can become invalid. It's commonly used
for benchmarking too, and I don't think it would be wrong to extend it to
normal files..
Linus
-
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 : Mon Jan 31 2000 - 21:00:13 EST