Subject: Re: 64-bit daddr_t problems with libsa
To: Charles M. Hannum <abuse@spamalicious.com>
From: Simon Burge <simonb@wasabisystems.com>
List: tech-kern
Date: 01/30/2003 12:46:38
"Charles M. Hannum" wrote:

> 
> I think, more importantly, that these operations could easily be
> converted to shifts and masks, so there is really no need to remove
> the indirect block support.

Using shifts and masks instead of multiplies and divides still ends up
too large.  Also note that it's double indirect block support that I was
suggesting we (optionally) remove, not all indirect block support.

Here's sizes for alpha.

Current unmodified bootxx_ffs:
	-rw-r--r--  1 simonb  wsrc  8096 Jan 30 12:42 bootxx_ffs
	checking sizes for bootxx_ffs/bootxx_ffs.sym...
	MAXIMUM LOAD SIZE EXCEEDED (7800 > 7680)

bootxx_ffs using shifts and masks:
	-rw-r--r--  1 simonb  wsrc  7800 Jan 30 12:40 bootxx_ffs
	checking sizes for bootxx_ffs/bootxx_ffs.sym...
	MAXIMUM LOAD SIZE EXCEEDED (8096 > 7680)

bootxx_ffs with no double indirect block support
	-rw-r--r--  1 simonb  wsrc  7504 Jan 30 12:44 bootxx_ffs
	checking sizes for bootxx_ffs/bootxx_ffs.sym... OK

bootxx_ffs using 32bit daddr_t
	-rw-r--r--  1 simonb  wsrc  7440 Jan 30 12:42 bootxx_ffs
	checking sizes for bootxx_ffs/bootxx_ffs.sym... OK

bootxx_ffs using 32bit daddr_t and no double indirect block support
	-rw-r--r--  1 simonb  wsrc  7184 Jan 30 12:35 bootxx_ffs
	checking sizes for bootxx_ffs/bootxx_ffs.sym... OK

Simon.
--
Simon Burge                            <simonb@wasabisystems.com>
NetBSD Support and Service:         http://www.wasabisystems.com/