Re: [PATCH v3 3/6] nfs-utils: Implement srcaddr binding in rpc_socket

[Ben Greear <greearb@candelatech.com>]

On 06/10/2011 03:06 PM, Chuck Lever wrote:
>
> On Jun 10, 2011, at 5:08 PM, greearb@candelatech.com wrote:
>
>> From: Ben Greear<greearb@candelatech.com>
>>
>> This implements the actual binding, if we are passed
>> a non-null local_ip structure.
>
> Why not _always_ pass a valid local_ip structure, and simply set .addr to an appropriate ANYADDR by default?  Then .is_set wouldn't be necessary, would it?  It would also simplify the logic in nfs_validate_options().

I like it as is because almost none of the new code is actually
used unless users pass in the srcaddr= option.  So, if I *did*
introduce any bugs, hopefully they would be limited to users
of the new option, and not a real regression.

Maybe after the srcaddr= code is used a bit I could go back and
do that cleanup.

But, I don't feel strongly about it, so if you think it's
worth the bother, I'll try changing the code as you suggest.

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com