Hello. Patch time (drivers/block/loop.c)

[Eric Windisch <ericw@grokthis.net>]

[-- Attachment #1: Type: text/plain, Size: 891 bytes --]


Hello. I would like to formally introduce myself to this list, as this is my first post. Of course, I would not come to the table empty handed.. I bring to you a patch ;)

I was digging around the loopback block device driver and it wasn't long before I realized that it assumes you are writing if you are not reading.. which I assume could be hazardous if there is an overflow, bad memory, or someone modifies some other code without knowledge :)

Patch attached.

I may be looking at this and the multi-device drivers in the near future as I have a need for support for partitions on these devices; I think it would be really neat if the software-raid could eventually become compatable with hardware raid.. although much slower, it may be a nice thing to have for testing purposes or a last resort for those who fry their hardware raid-controllers ;)

--
Eric Windisch
http://bwbohh.net

[-- Attachment #2: loop-badmem.diff --]
[-- Type: text/plain, Size: 636 bytes --]

--- linux/drivers/block/loop.c	Fri Dec 21 12:41:53 2001
+++ linux-2.4.17-ericw/drivers/block/loop.c	Sat Dec 22 02:35:17 2001
@@ -90,8 +90,10 @@
 	if (raw_buf != loop_buf) {
 		if (cmd == READ)
 			memcpy(loop_buf, raw_buf, size);
-		else
+		else if (cmd == WRITE) 
 			memcpy(raw_buf, loop_buf, size);
+		else
+			printk(KERN_ERR "loop: Illegal command, %d", cmd);
 	}
 
 	return 0;
@@ -106,9 +108,11 @@
 	if (cmd == READ) {
 		in = raw_buf;
 		out = loop_buf;
-	} else {
+	} else if (cmd == WRITE) {
 		in = loop_buf;
 		out = raw_buf;
+	} else {
+		printk(KERN_ERR "loop: Illegal command, %d", cmd);
 	}
 
 	key = lo->lo_encrypt_key;