Re: [PATCH 1/4] fix not-and/or errors

[Roel Kluin <12o3l@tiscali.nl>]

Al Viro wrote:
> On Wed, Oct 17, 2007 at 03:46:43PM +0200, Roel Kluin wrote:
>> +++ b/drivers/misc/asus-laptop.c
>> @@ -322,7 +322,7 @@ static void write_status(acpi_handle handle, int out, int mask)
>>  
>>  	switch (mask) {
>>  	case MLED_ON:
>> -		out = !out & 0x1;
>> +		out = !(out & 0x1);
> 
> Not sure if that's what had been intended.

It seems to me if I look at the code, that it's intended to make a bool out of 'out'. That's
nonsense because of the precedence the ! will turn it into a boolean before the & 0x1.

x = !x & y behaves like x = !x for y != 0. 
for y = 1 the behavior is even the same for x = !(x & y)
so it does not matter in this case, except for clarity. I'll make it out = !out.

> 
>> @@ -2882,7 +2882,7 @@ static int st_int_ioctl(struct scsi_tape *STp, unsigned int cmd_in, unsigned lon
>>  			    !(STp->use_pf & PF_TESTED)) {
>>  				/* Try the other possible state of Page Format if not
>>  				   already tried */
>> -				STp->use_pf = !STp->use_pf | PF_TESTED;
>> +				STp->use_pf = !(STp->use_pf | PF_TESTED);
> 
> Wrong.  This code, ugly as it is, happens to be correct.  Replacement
> isn't.  I would rewrite it as ^= PF_TESTED | USE_PF; /* remove USE_PF, set *
> 						      * PF_TESTED */
> 
> The rest is covered by Alexey's patch and one I'd posted as followup.


ok, thanks, I'll correct and omit these in my follow up patch.

Roel