Alan Cox wrote:
>
> ...
> o Cache more group descriptors on ext2/ext3 (Arjan van de Ven)
[ See below for the rant ]
I'd be interested in knowing what testing was performed, what
workload this addresses, what improvements were observed, etc.
Did anyone consider and/or test removal of the cache altogether,
and adding a touch_buffer() to the backing buffer_head, to allow the
VM to perform the LRU management instead?
Did anyone try leaving the LRU at its current size and seeing if
a touch_buffer() improves the problematic testcase?
If there is indeed a workload which is improved by this patch,
I would like know what it is, please - I would like to see if allowing
the VM to manage the LRU also fixes that workload, because that's
surely better than pinning down 256 kilobytes of memory per mounted
filesystem.
Thanks.
The rant:
Not singling out Arjan; certainly this is not the most egregious
case lately. But. Will people please stop sending kernel
patches straight to tree owners without copying the appropriate
mailing list?
There is no benefit in keeping all the other kernel developers
in the dark. If a patch is not security-related and is not
trivially boring, tree-owners should consider just dropping
the thing if it has not been seen by the other developers.
It's not as if this mailing list is overwhelmed with technical
content, is it? If people are shy, or are reluctant to disrupt
the social and political ambiance of linux-kernel, there are
other lists, including
linux-fsdevel@vger.kernel.org
linux-mm@kvack.org
netdev@oss.sgi.com
ext2-devel@lists.sourceforge.net
and of course many others.
-
-
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/
This archive was generated by hypermail 2b29 : Sun Apr 07 2002 - 22:00:15 EST