Re: Something broken in sys_swapon

[Simon Kirby <sim@netnation.com>]

Aha!  I then tried this patch:

--- linux/fs/block_dev.c.orig	Sun Nov  4 11:35:05 2001
+++ linux/fs/block_dev.c	Sun Nov  4 12:21:51 2001
@@ -84,6 +84,15 @@
 	}
 
 	oldsize = blksize_size[MAJOR(dev)][MINOR(dev)];
+
+	printk("Changing device %02x:%02x block size from %u to %u\n",
+		MAJOR(dev),MINOR(dev),
+		oldsize,size);
+	if (MAJOR(dev) == 0x03 && MINOR(dev) == 0x42){
+		printk("...Refused.\n");
+		return 0;
+	}
+
 	if (oldsize == size)
 		return 0;
 
...And now my system boots fine with /dev/hdb2 swap in the fstab.

In fact, I tried /dev/hdb1 after and then I couldn't read any more from
/boot (which is /dev/hda1).  So, some sort of wraparound is happening
here.  Why would blksize_size[3][2] be affected by blksize_size[3][0x42]?

Simon-

[  Stormix Technologies Inc.  ][  NetNation Communications Inc. ]
[       sim@stormix.com       ][       sim@netnation.com        ]
[ Opinions expressed are not necessarily those of my employers. ]