On Sun, Nov 10, 2024 at 9:51 PM Sergey Senozhatsky <senozhatsky@xxxxxxxxxxxx> wrote: > > Hi, > > On (24/08/29 10:42), Andrii Nakryiko wrote: > > Now that build ID related internals in kernel/bpf/stackmap.c can be used > > both in sleepable and non-sleepable contexts, we need to add additional > > rcu_read_lock()/rcu_read_unlock() protection around fetching > > perf_callchain_entry, but with the refactoring in previous commit it's > > now pretty straightforward. We make sure to do rcu_read_unlock (in > > sleepable mode only) right before stack_map_get_build_id_offset() call > > which can sleep. By that time we don't have any more use of > > perf_callchain_entry. > > Shouldn't this be backported to stable kernels? It seems that those still > do suspicious-RCU deference: > > __bpf_get_stack() > get_perf_callchain() > perf_callchain_user() > perf_get_guest_cbs() Do you see this issue in practice or have some repro? __bpf_get_stack() shouldn't be callable from sleepable BPF programs until my patch set, so I don't think there is anything to be backported. But maybe I'm missing something, which is why I'm asking whether this is a conclusion drawn from source code analysis, or there was actually a report somewhere.
