On 3/31/21 1:08 AM, Eric Dumazet wrote:
> On Tue, Mar 30, 2021 at 6:02 PM Taehee Yoo wrote:
>>
>> On 3/31/21 12:40 AM, Eric Dumazet wrote:
>> > This seems a serious regression compared to old code (in net tree)
>> >
>> > Have you added RTNL requirement in all this code ?
>> >
>> > We wo
On 3/31/21 12:40 AM, Eric Dumazet wrote:
> This seems a serious regression compared to old code (in net tree)
>
> Have you added RTNL requirement in all this code ?
>
> We would like to use RTNL only if strictly needed.
Yes, I agree with you.
This patchset actually relies on existed RTNL, which i
ip6_mc_msfget() should be called under RTNL because it accesses RTNL
protected data. but the caller doesn't acquire rtnl_lock().
So, data couldn't be protected.
Therefore, it adds rtnl_lock() in do_ipv6_getsockopt(),
which is the caller of ip6_mc_msfget().
Splat looks like:
===
