[OpenWrt-Devel] [BUG] lantiq: net: wrong operator

[Kevin 'ldir' Darbyshire-Bryant]

> On 17 Feb 2019, at 08:57, Mathias Kresin <dev at kresin.me> wrote:
> 
> 08/02/2019 09:23, Petr Cvek:
>> Hello,
>> There is a wrong code in 0025-NET-MIPS-lantiq-adds-xrx200-net.patch [1], the original code:
>> +	link->rx_flow = !!(xrx200sw_read_x(XRX200_MAC_CTRL_0_FCON, port) && 0x0010);
>> +	link->tx_flow = !!(xrx200sw_read_x(XRX200_MAC_CTRL_0_FCON, port) && 0x0020);
>> wants to mask the register value and is using a logical AND and not a bitwise AND.
>> The fix should be:
>> +	link->rx_flow = !!(xrx200sw_read_x(XRX200_MAC_CTRL_0_FCON, port) & 0x0010);
>> +	link->tx_flow = !!(xrx200sw_read_x(XRX200_MAC_CTRL_0_FCON, port) & 0x0020);
>> References:
>> [1] https://github.com/openwrt/openwrt/blob/master/target/linux/lantiq/patches-4.14/0025-NET-MIPS-lantiq-adds-xrx200-net.patch#L937
>> 
> 
> Hey Petr,
> 
> it looks indeed wrong. And looking more closer at the lines, I don't get why we need the double negation.

Having seen this code structure before, it’s a non-obvious way of turning a masked and hence value based true/false into a binary 0 or 1 result:  Exhibit A moronic demo c programme:

#include <stdio.h>
  
int main()
{

	printf("%d\n", 5 & 4);
	printf("%d\n", (5 & 4));
	printf("%d\n", !(5 & 4));
	printf("%d\n", !!(5 & 4));

}

Results in:

4
4
0
1


Cheers,

Kevin D-B

012C ACB2 28C6 C53E 9775  9123 B3A2 389B 9DE2 334A

_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel