On Thu, 13 Apr 2000, Stephen C. Tweedie wrote:
>[..] The cost
>of having it too large on a less file-intensive workload may be less
>than the cost of having it too small if you really are using that many
>inodes. [..]
Agreed. That's the point of this patch:
I'm not claming the patch to be the right solution or the right
implementation, however the icache/dcache should be someway bound in
function of the icache/dacache hashtable sizes.
The other way is to left everything unbound as now and to assume the whole
memory can be populated only by inodes+dentries or only by dentries and 1
inode (hardlinks) and to allocate an huge hashtable for icache and dcache
wasting memory in 99% of real world usages.
The icache/dcache stealing is completly another issue and it should be
probably addressed queueing someway unused dentries and inodes into the
page-LRU.
Andrea
-
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 : Sat Apr 15 2000 - 21:00:21 EST