[PATCH] bugfix for drivers/char/mem.c

[PATCH] bugfix for drivers/char/mem.c

[Steven Hand]

There is a bug in the function read_kmem() in drivers/char/mem.c  -
the file position is incorrectly updated after a read() from /dev/kmem.
This means that successive reads from /dev/kmem without an intervening
seek will return incorrect data. 

The bug has been around for quite a while, and is present in all the
2.2.x kernels.  The following patch (against v.2.3.12) fixes it.


S.

--- v2.3.12/linux/drivers/char/mem.c	Sun Aug  1 11:09:11 1999
+++ linux/drivers/char/mem.c	Sun Aug  1 11:25:07 1999
@@ -245,11 +245,14 @@
 		count -= read;
 	}
 
-	virtr = vread(buf, (char *)p, count);
-	if (virtr < 0)
-		return virtr;
-	*ppos += p + virtr;
-	return virtr + read;
+	if(count) {
+		if((virtr = vread(buf, (char *)p, count)) < 0)
+			return virtr;
+		read += virtr;
+	}
+
+	*ppos += read;
+	return read;
 }
 
 /*


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://humbolt.geo.uu.nl/Linux-MM/

Re: [PATCH] bugfix for drivers/char/mem.c

[Kanoj Sarcar]

> 
> There is a bug in the function read_kmem() in drivers/char/mem.c  -
> the file position is incorrectly updated after a read() from /dev/kmem.
> This means that successive reads from /dev/kmem without an intervening
> seek will return incorrect data. 
> 
> The bug has been around for quite a while, and is present in all the
> 2.2.x kernels.  The following patch (against v.2.3.12) fixes it.
> 
> 
> S.
> 
> --- v2.3.12/linux/drivers/char/mem.c	Sun Aug  1 11:09:11 1999
> +++ linux/drivers/char/mem.c	Sun Aug  1 11:25:07 1999
> @@ -245,11 +245,14 @@
>  		count -= read;
>  	}
>  
> -	virtr = vread(buf, (char *)p, count);
> -	if (virtr < 0)
> -		return virtr;
> -	*ppos += p + virtr;
> -	return virtr + read;
> +	if(count) {
> +		if((virtr = vread(buf, (char *)p, count)) < 0)
> +			return virtr;
> +		read += virtr;
> +	}
> +
> +	*ppos += read;
> +	return read;
>  }
>  
>  /*
>

I see the problem. Can I suggest a simpler fix, based on the fact that
vread() returns 0 when count == 0?

-	*ppos += p + virtr;
+	*ppos += read + virtr;

Kanoj
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://humbolt.geo.uu.nl/Linux-MM/

Re: [PATCH] bugfix for drivers/char/mem.c

[Steven Hand]

>I see the problem. Can I suggest a simpler fix, based on the fact 
>that vread() returns 0 when count == 0?
>
>-	*ppos += p + virtr;
>+	*ppos += read + virtr;

Sure if you prefer it. 

The main reason I added the if(count) { ... } was just to avoid 
calling vread() when count is zero (since in this case there's no 
entry to find on the vmlist, and hence the for loop in vread() will 
just iterate over all vmalloc()'d areas before returning zero.)


S.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://humbolt.geo.uu.nl/Linux-MM/