#Hashtab x64 archive
#Hashtab x64 Patch

+static int check_attach_modify_return(struct bpf_prog *prog, unsigned long addr) static int check_attach_modify_return(struct bpf_verifier_env *env) Kernel/bpf/verifier.c | 24 +++++++++++-ġ file changed, 11 insertions(+), 13 deletions(-)ĭiff -git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c Otherwise prog freeing will try to put trampoline via prog->aux->trampoline,īut it may not point to a valid trampoline.įixes: 6ba43b761c41 ('bpf: Attachment verification for BPF_MODIFY_RETURN') Prog->aux->trampoline = tr should be set only when prog is valid. Workqueue: events bpf_prog_free_deferred Write of size 4 at addr ffff8881ef38b8a0 by task kworker/3:5/2243 BUG: KASAN: use-after-free in bpf_trampoline_put+0x39/0x2a0 ` (2 subsequent siblings) 3 siblings, 0 replies 11+ messages in thread 4:38 ` bpf: Introduce sleepable BPF programs Alexei Starovoitov 4:38 bpf: Introduce minimal support for sleepable progs Alexei 4:38 ` Alexei Starovoitov * bpf: Fix use-after-free in fmod_ret check selftests/bpf/progs/trigger_bench.c | 7 ++ġ4 files changed, 232 insertions(+), 37 deletions(-) Tools/testing/selftests/bpf/progs/lsm.c | 9 +. selftests/bpf/benchs/bench_trigger.c | 17 ++++ Tools/testing/selftests/bpf/bench.c | 2 + Support for perf_event_array and bpf_perf_event_output() will follow as well.īpf: Fix use-after-free in fmod_ret checkĪrch/x86/net/bpf_jit_comp.c | 36 +++++. Here is 'perf report' difference of sleepable vs non-sleepable:Ġ.92% bench bpf_prog_740d4210cdcd99a3_bench_trigger_fentry_sleepĠ.26% bench _bpf_prog_exit_sleepableĠ.21% bench _bpf_prog_enter_sleepableĠ.88% bench bpf_prog_740d4210cdcd99a3_bench_trigger_fentryĬlearly sleepable vs non-sleepable program invocation overhead is high, but itĬould be acceptable in certain cases until rcu_trace is available.īpf_copy_from_user(), bpf_msleep() will be coming in the future patches. Only array and pre-allocated hash and lru maps allowed.

#Hashtab x64 Patch

In this patch only fentry/fexit/fmod_ret and lsm progs can be sleepable. This patch set introduces the minimal viable support for sleepable bpf programs. To: davem +Cc: daniel, netdev, bpf, kernel-team ` (3 more replies) 0 siblings, 4 replies 11+ messages in threadįrom: Alexei Starovoitov 4:38 UTC ( / raw) 4:38 ` bpf: Fix use-after-free in fmod_ret check Alexei Starovoitov

#Hashtab x64 archive

Bpf: Introduce minimal support for sleepable progs Netdev Archive on help / color / mirror / Atom feed * bpf: Introduce minimal support for sleepable progs 4:38 Alexei Starovoitov