On 18 September 2012 13:51, Sergei Shtylyov <sshtylyov@xxxxxxxxxx> wrote:
> Hello.
>
>
> On 18-09-2012 13:32, Jonas Gorski wrote:
>
>> While calculating the mac address the pointer for the current octet was
>> never reset back to the least significant one after being decremented
>> because of an octet overflow. This resulted in the code continuing to
>> increment at the current octet, potentially generating duplicate or
>> invalid mac addresses.
>
>
>> As a second issue the pointer was allowed to advance up to the most
>> significant octet, modifying the OUI, and potentially changing the type
>> of mac address.
>
>
>> Rewrite the code so it resets the pointer to the least significant
>> in each outer loop step, and bails out when the least significant octet
>> of the OUI is reached.
>
>
>> Signed-off-by: Jonas Gorski <jonas.gorski@xxxxxxxxx>
>> ---
>> arch/mips/bcm63xx/boards/board_bcm963xx.c | 16 +++++++++-------
>> 1 file changed, 9 insertions(+), 7 deletions(-)
>
>
>> diff --git a/arch/mips/bcm63xx/boards/board_bcm963xx.c
>> b/arch/mips/bcm63xx/boards/board_bcm963xx.c
>> index ea4ea77..f0fcec6 100644
>> --- a/arch/mips/bcm63xx/boards/board_bcm963xx.c
>> +++ b/arch/mips/bcm63xx/boards/board_bcm963xx.c
>> @@ -720,7 +720,7 @@ const char *board_get_name(void)
>> */
>> static int board_get_mac_address(u8 *mac)
>> {
>> - u8 *p;
>> + u8 *oui;
>> int count;
>>
>> if (mac_addr_used >= nvram.mac_addr_count) {
>> @@ -729,21 +729,23 @@ static int board_get_mac_address(u8 *mac)
>> }
>>
>> memcpy(mac, nvram.mac_addr_base, ETH_ALEN);
>> - p = mac + ETH_ALEN - 1;
>> + oui = mac + ETH_ALEN/2 - 1;
>> count = mac_addr_used;
>>
>> while (count--) {
>> + p = mac + ETH_ALEN - 1;
>
>
> But didn't you remove 'p' above? Did you compile this?
Argh. Yes, but the wrong version (my "user space" one to test it). Let
me try that again ... . Thanks for catching it.
Jonas
