Reply: [net] bonding: use return instead of goto
> On Fri, Feb 05, 2016 at 09:42:24AM +0800, 张胜举 wrote: > > > On Wed, Feb 03, 2016 at 06:15:22AM +, Zhang Shengju wrote: > > > > Replace 'goto' with 'return' to remove unnecessary check at label: > > > > err_undo_flags. > > > > > > I think you're going to have to explain how you came to the > > > conclusion that the check isn't necessary. > ... > > The reason is that 'err_undo_flags' do two things for the first slave > > device: > > 1. revert bond mac address if it is set by the slave device. > > 2. revert bond device type if it's not ARPHRD_ETHER. > > > > I think it's not necessary for the three places, they changed neither > > bond mac address nor type. > > it's straightforward to return directly. > > I see what you're saying, and that does look to be true if you're only adding > a > singular first device right now. But looking at the enslave and release > paths, I > don't see anything preventing concurrent slave adds and removes, which > could mean there are situations where those checks really are necessary. I > don't actually know. > > -- > Jarod Wilson > ja...@redhat.com All of three places don't change any attributes of bond device, just for sanity check before enslaving. If it's necessary for these three places to goto 'err_undo_flags', it will mean that all sanity checks before these three places need to do so, which is not right clearly. Even there is concurrent situation which make the check necessary, which I haven't figure out yet, it's not right place to do this. Thanks, Shengju
Re: Reply: [net] bonding: use return instead of goto
On Fri, Feb 05, 2016 at 09:42:24AM +0800, 张胜举 wrote: > > On Wed, Feb 03, 2016 at 06:15:22AM +, Zhang Shengju wrote: > > > Replace 'goto' with 'return' to remove unnecessary check at label: > > > err_undo_flags. > > > > I think you're going to have to explain how you came to the conclusion > > that the check isn't necessary. ... > The reason is that 'err_undo_flags' do two things for the first slave > device: > 1. revert bond mac address if it is set by the slave device. > 2. revert bond device type if it's not ARPHRD_ETHER. > > I think it's not necessary for the three places, they changed neither bond > mac address nor type. > it's straightforward to return directly. I see what you're saying, and that does look to be true if you're only adding a singular first device right now. But looking at the enslave and release paths, I don't see anything preventing concurrent slave adds and removes, which could mean there are situations where those checks really are necessary. I don't actually know. -- Jarod Wilson ja...@redhat.com
Reply: [net] bonding: use return instead of goto
> On Wed, Feb 03, 2016 at 06:15:22AM +, Zhang Shengju wrote: > > Replace 'goto' with 'return' to remove unnecessary check at label: > > err_undo_flags. > > I think you're going to have to explain how you came to the conclusion that > the check isn't necessary. > > -- > Jarod Wilson > ja...@redhat.com Hi Jarod, The reason is that 'err_undo_flags' do two things for the first slave device: 1. revert bond mac address if it is set by the slave device. 2. revert bond device type if it's not ARPHRD_ETHER. I think it's not necessary for the three places, they changed neither bond mac address nor type. it's straightforward to return directly. Thanks, Shengju