Re: [PATCH] wl1271: Fix overflow in wl1271_boot_upload_nvs

[Luciano Coelho <luciano.coelho@nokia.com>]

On Wed, 2010-09-22 at 09:53 +0200, ext Ido Yariv wrote:
> Due to miscalculation of nvs_len, excessive data was sent to the
> firmware.
> Fix this by first setting nvs_ptr to point to the first NVS table,
> and computing the total size of all NVS tables accordingly.
> 
> Signed-off-by: Ido Yariv <ido@wizery.com>
> ---

This looks reasonable, thanks!

But I still want to have it briefly tested before I accept it.  Our
tester will try it out today or tomorrow and, if everything is okay,
I'll ack it.


>  drivers/net/wireless/wl12xx/wl1271_boot.c |   17 ++++++-----------
>  1 files changed, 6 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/net/wireless/wl12xx/wl1271_boot.c b/drivers/net/wireless/wl12xx/wl1271_boot.c
> index fc21db8..e5a7f04 100644
> --- a/drivers/net/wireless/wl12xx/wl1271_boot.c
> +++ b/drivers/net/wireless/wl12xx/wl1271_boot.c
> @@ -274,11 +274,11 @@ static int wl1271_boot_upload_nvs(struct wl1271 *wl)

[...]

> -	nvs_aligned = kmemdup(nvs_ptr, nvs_len, GFP_KERNEL); if
> -	(!nvs_aligned) return -ENOMEM;
> +	nvs_aligned = kmemdup(nvs_ptr, nvs_len, GFP_KERNEL);
> +	if (!nvs_aligned)
> +		return -ENOMEM;

This looks pretty odd.  But I checked wireless-testing and it is looking
bad there.  It's not like that in our internal tree, but git blames me
for doing it in wireless-testing. :) Thanks for fixing.

-- 
Cheers,
Luca.