> Yes. Essentially, we shouldn't cache meta-data more aggressively, but we
> should be _extremely_ aggressive at throwing out pages that we used for
> data writes..
>
> I don't think we really want buffer aging, I think we need to just have
> some way of getting rid of data buffers..
Perhaps it would help to clear the touched bit for a buffer after updating the
page cache following a write? Assuming that the b_count of a data buffer goes to
0 quickly following the write, clearing the touched bit would make the page seem
unreferenced the next time shrink_mmap looked at it, and it would then be
reclaimed.
Regards,
Bill
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu