Subject: Should LFS support block sizes larger than 16k?
To: None <current-users@netbsd.org>
From: Blair Sadewitz <blair.sadewitz@gmail.com>
List: current-users
Date: 08/13/2007 13:17:41 
Using 4.99.28, the kernel panics in pool_get() (in lfs_vfsops.c, IIRC)
 upon mounting an LFS filesystem with a block size (and frag size
calculated according to the default shift) >= 16k.

I suspect that using a block size of MAXPHYS with a larger frag size
(I'm thinking 8k) might ameliorate the slowdown which occurs, for
example, at the end of untarring pkgsrc.

Can anyone reproduce this?  I'm not sure why it's happening [yet].

Regards,

--Blair