[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20191031102229.GA2794@krava>
Date: Thu, 31 Oct 2019 11:22:29 +0100
From: Jiri Olsa <jolsa@...hat.com>
To: Alexei Starovoitov <ast@...nel.org>
Cc: davem@...emloft.net, daniel@...earbox.net, netdev@...r.kernel.org,
bpf@...r.kernel.org, kernel-team@...com
Subject: Re: [PATCH bpf-next] bpf: Fix bpf jit kallsym access
On Wed, Oct 30, 2019 at 04:30:19PM -0700, Alexei Starovoitov wrote:
> Jiri reported crash when JIT is on, but net.core.bpf_jit_kallsyms is off.
> bpf_prog_kallsyms_find() was skipping addr->bpf_prog resolution
> logic in oops and stack traces. That's incorrect.
> It should only skip addr->name resolution for 'cat /proc/kallsyms'.
> That's what bpf_jit_kallsyms and bpf_jit_harden protect.
>
> Reported-by: Jiri Olsa <jolsa@...hat.com>
> Signed-off-by: Alexei Starovoitov <ast@...nel.org>
> Fixes: 3dec541b2e63 ("bpf: Add support for BTF pointers to x86 JIT")
it fixes the crash for me, thanks for quick fix
jirka
> ---
> kernel/bpf/core.c | 3 ---
> 1 file changed, 3 deletions(-)
>
> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> index 673f5d40a93e..8d3fbc86ca5e 100644
> --- a/kernel/bpf/core.c
> +++ b/kernel/bpf/core.c
> @@ -668,9 +668,6 @@ static struct bpf_prog *bpf_prog_kallsyms_find(unsigned long addr)
> {
> struct latch_tree_node *n;
>
> - if (!bpf_jit_kallsyms_enabled())
> - return NULL;
> -
> n = latch_tree_find((void *)addr, &bpf_tree, &bpf_tree_ops);
> return n ?
> container_of(n, struct bpf_prog_aux, ksym_tnode)->prog :
> --
> 2.17.1
>
Powered by blists - more mailing lists