Not asked of me, but as always, I do have an opinion:
I think the real reason for the very large disk caches is that the
cost of a track buffer for simple read-ahead is about the same as the
1 MB "cache" on cheap modern drives. And with very simple logic they
can "cache" several physical tracks, say the ones that contain the inode
and the last few sectors of the most recently accessed file. Sometimes
this saves you a rotational delay time reading or writing a sector span,
so it can do better than the OS then (I admit, that doesn't happen often).
And the cost/benefit tradeoff is worth it, because the cost is so little.
[Someone who really knows may correct me, however.]
--Charles
/"\ |
\ / ASCII Ribbon Campaign |
X Against HTML Mail |--Charles Marslett
/ \ | www.wordmark.org
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/