Re: [PATCH net v2] ipv6: fix a BUG in rt6_get_pcpu_route() under PREEMPT_RT

On Fri, Dec 19, 2025 at 3:52 AM Jiayuan Chen [off-list ref] wrote:

On PREEMPT_RT kernels, after rt6_get_pcpu_route() returns NULL, the
current task can be preempted. Another task running on the same CPU
may then execute rt6_make_pcpu_route() and successfully install a
pcpu_rt entry. When the first task resumes execution, its cmpxchg()
in rt6_make_pcpu_route() will fail because rt6i_pcpu is no longer
NULL, triggering the BUG_ON(prev). It's easy to reproduce it by adding
mdelay() after rt6_get_pcpu_route().

Using preempt_disable/enable is not appropriate here because
ip6_rt_pcpu_alloc() may sleep.

Fix this by:
1. Removing the BUG_ON and instead handling the race gracefully by
   freeing our allocation and returning the existing pcpu_rt when
   cmpxchg() fails.
2. Keeping the BUG_ON for non-PREEMPT_RT kernels, since preemption
   should not occur in this context and a cmpxchg failure would
   indicate a real bug.

Link: https://syzkaller.appspot.com/bug?extid=9b35e9bc0951140d13e6
Fixes: 951f788a80ff ("ipv6: fix a BUG in rt6_get_pcpu_route()")

I would rather find when PREEMPT_RT was added/enabled, there is no
point blaming such an old commit
which was correct at the time, and forcing pointless backports to old
linux kernels.