netdev - Re: [PATCH] bpf: fix a missing check of return value

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20181220195855.akmg7dcg2njz25gs@ast-mbp.dhcp.thefacebook.com>
Date:   Thu, 20 Dec 2018 11:58:57 -0800
From:   Alexei Starovoitov <alexei.starovoitov@...il.com>
To:     Kangjie Lu <kjlu@....edu>
Cc:     pakki001@....edu, Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] bpf: fix a missing check of return value

On Thu, Dec 20, 2018 at 01:45:56PM -0600, Kangjie Lu wrote:
> check_reg_arg() may fail and not mark correct data in "env".  This
> fix inserts a check that ensures check_reg_arg() is successful, and
> if it is not, the fix stops further operations and returns an error
> upstream.
> 
> Signed-off-by: Kangjie Lu <kjlu@....edu>
> ---
>  kernel/bpf/verifier.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 51ba84d4d34a..ced8cc6470b1 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -2619,7 +2619,9 @@ static int check_func_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
>  	/* after the call registers r0 - r5 were scratched */
>  	for (i = 0; i < CALLER_SAVED_REGS; i++) {
>  		mark_reg_not_init(env, caller->regs, caller_saved[i]);
> -		check_reg_arg(env, caller_saved[i], DST_OP_NO_MARK);
> +		err = check_reg_arg(env, caller_saved[i], DST_OP_NO_MARK);
> +		if (err)
> +			return err;

it cannot fail here.
we have the same pattern in few other places where we know it cannot fail.
I prefer to leave the code as-is.