Re: [PATCH 1/2] - net/socket.c::sys_bind() cleanup.

[Jesper Juhl <juhl-lkml@dif.dk>]

James Morris wrote:
> On Wed, 17 Nov 2004, Jesper Juhl wrote:
> 
> 
>>Not exactely : 
>>
>>
>>
>>>-		if((err=move_addr_to_kernel(umyaddr,addrlen,address))>=0) {
>>
>>>+	err = move_addr_to_kernel(umyaddr, addrlen, address);
>>>+	if (err)
>>>+		goto out_put;
>>
>>
>>The original tests for err >= 0, your replacement tests if err is != 0
> 
> 
> Look at move_addr_to_kernel(), it only returns 0 or -error.
> 
> The patch looks good to me.
> 
Right, I had not looked at it in detail. I just reacted to the claim 
that "it does exactely the same" but I could see in the posted patch 
that it didn't do exactely the same and there was no explanation of why 
it was ok to have that difference.
After reading move_addr_to_kernel(), I agree that the patch looks fine.

--
Jesper Juhl