blob-cvs-commit

Quirk in blob/jffs2.c?

[Clive N. <cl...@ba...>]

In read_inode of blob/jffs2.c the following for loop appears

	for (page = 0; page <= size / 4096; page++) {
		/* Print a '.' every 0x40000 bytes */

Should "size / 4096" not be "(size-1) / 4096" ?

Clive

Re: Quirk in blob/jffs2.c?

[Erik M. <er...@bi...>]

On Thu, Jul 15, 2004 at 02:24:54PM +1200, Clive Nicolson wrote:
> In read_inode of blob/jffs2.c the following for loop appears
> 
> 	for (page = 0; page <= size / 4096; page++) {
> 		/* Print a '.' every 0x40000 bytes */
> 
> Should "size / 4096" not be "(size-1) / 4096" ?

I haven't looked at the file, but if it is indeed a bug, the cleaner
fix is to replace <= by <.

But I think we want size/4096, think about what happens when size = 0
or size = 1.

(again, I haven't looked at the file cause I don't have the source
available right here and my internet link is a bit flakey, so I'm just
guessing).


Erik

-- 
----  Erik Mouw  ----  www.bitwizard.nl  ----  +31 15 2600 998  ----

Re: Quirk in blob/jffs2.c?

[Russ D. <Rus...@as...>]

On Wed, 2004-07-14 at 19:24, Clive Nicolson wrote:
> In read_inode of blob/jffs2.c the following for loop appears
> 
> 	for (page = 0; page <= size / 4096; page++) {
> 		/* Print a '.' every 0x40000 bytes */

Its been a while since I have touched that code, however, I would think
that it would be very common for a file size to be a multiple of page
size, thus if it is a problem, I would think that it would have shown
up.

Have you observed this problem occuring? It should be simple to make a
test case where it would fail

-- 
Russ Dill <Rus...@as...>