From patchwork Fri Apr 30 14:20:21 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Greg Kroah-Hartman X-Patchwork-Id: 12233521 X-Patchwork-Delegate: bpf@iogearbox.net Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-19.0 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 5AB6FC43619 for ; Fri, 30 Apr 2021 14:20:47 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 374B761474 for ; Fri, 30 Apr 2021 14:20:47 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232951AbhD3OVe (ORCPT ); Fri, 30 Apr 2021 10:21:34 -0400 Received: from mail.kernel.org ([198.145.29.99]:57802 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232910AbhD3OVc (ORCPT ); Fri, 30 Apr 2021 10:21:32 -0400 Received: by mail.kernel.org (Postfix) with ESMTPSA id E8A5361459; Fri, 30 Apr 2021 14:20:43 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1619792444; bh=/BSsFORZ6JEtWqYp2Ieks3xs7+yAxpn2KnmUZ42HNiQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=c/EX57ZwbdwmfcA45JV2AWYj/G876Kb2AlrFZZ0lLy6Wyixxpvnt0heYqVPcRg60x OIDpXRNmNsdImacP2DoW3wZo7+tJwNZGWcVL11BAdyLK7Y2AxOGoT4YymDymYEGEFR HM0pRAOUYbX71Aly6CFeYht8O1CE38FukBTvnVxA= From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org, stable@vger.kernel.org Cc: Greg Kroah-Hartman , bpf@vger.kernel.org, Piotr Krysiuk , Benedict Schlueter , Daniel Borkmann , John Fastabend , Alexei Starovoitov , Frank van der Linden Subject: [PATCH 5.4 7/8] bpf: Tighten speculative pointer arithmetic mask Date: Fri, 30 Apr 2021 16:20:21 +0200 Message-Id: <20210430141911.396953187@linuxfoundation.org> X-Mailer: git-send-email 2.31.1 In-Reply-To: <20210430141911.137473863@linuxfoundation.org> References: <20210430141911.137473863@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: bpf@vger.kernel.org X-Patchwork-Delegate: bpf@iogearbox.net From: Daniel Borkmann commit 7fedb63a8307dda0ec3b8969a3b233a1dd7ea8e0 upstream. This work tightens the offset mask we use for unprivileged pointer arithmetic in order to mitigate a corner case reported by Piotr and Benedict where in the speculative domain it is possible to advance, for example, the map value pointer by up to value_size-1 out-of-bounds in order to leak kernel memory via side-channel to user space. Before this change, the computed ptr_limit for retrieve_ptr_limit() helper represents largest valid distance when moving pointer to the right or left which is then fed as aux->alu_limit to generate masking instructions against the offset register. After the change, the derived aux->alu_limit represents the largest potential value of the offset register which we mask against which is just a narrower subset of the former limit. For minimal complexity, we call sanitize_ptr_alu() from 2 observation points in adjust_ptr_min_max_vals(), that is, before and after the simulated alu operation. In the first step, we retieve the alu_state and alu_limit before the operation as well as we branch-off a verifier path and push it to the verification stack as we did before which checks the dst_reg under truncation, in other words, when the speculative domain would attempt to move the pointer out-of-bounds. In the second step, we retrieve the new alu_limit and calculate the absolute distance between both. Moreover, we commit the alu_state and final alu_limit via update_alu_sanitation_state() to the env's instruction aux data, and bail out from there if there is a mismatch due to coming from different verification paths with different states. Reported-by: Piotr Krysiuk Reported-by: Benedict Schlueter Signed-off-by: Daniel Borkmann Reviewed-by: John Fastabend Acked-by: Alexei Starovoitov Tested-by: Benedict Schlueter [fllinden@amazon.com: backported to 5.4] Signed-off-by: Frank van der Linden Signed-off-by: Greg Kroah-Hartman --- kernel/bpf/verifier.c | 73 ++++++++++++++++++++++++++++++-------------------- 1 file changed, 44 insertions(+), 29 deletions(-) --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -4278,7 +4278,7 @@ static int retrieve_ptr_limit(const stru bool off_is_neg = off_reg->smin_value < 0; bool mask_to_left = (opcode == BPF_ADD && off_is_neg) || (opcode == BPF_SUB && !off_is_neg); - u32 off, max = 0, ptr_limit = 0; + u32 max = 0, ptr_limit = 0; if (!tnum_is_const(off_reg->var_off) && (off_reg->smin_value < 0) != (off_reg->smax_value < 0)) @@ -4287,26 +4287,18 @@ static int retrieve_ptr_limit(const stru switch (ptr_reg->type) { case PTR_TO_STACK: /* Offset 0 is out-of-bounds, but acceptable start for the - * left direction, see BPF_REG_FP. + * left direction, see BPF_REG_FP. Also, unknown scalar + * offset where we would need to deal with min/max bounds is + * currently prohibited for unprivileged. */ max = MAX_BPF_STACK + mask_to_left; - /* Indirect variable offset stack access is prohibited in - * unprivileged mode so it's not handled here. - */ - off = ptr_reg->off + ptr_reg->var_off.value; - if (mask_to_left) - ptr_limit = MAX_BPF_STACK + off; - else - ptr_limit = -off - 1; + ptr_limit = -(ptr_reg->var_off.value + ptr_reg->off); break; case PTR_TO_MAP_VALUE: max = ptr_reg->map_ptr->value_size; - if (mask_to_left) { - ptr_limit = ptr_reg->umax_value + ptr_reg->off; - } else { - off = ptr_reg->smin_value + ptr_reg->off; - ptr_limit = ptr_reg->map_ptr->value_size - off - 1; - } + ptr_limit = (mask_to_left ? + ptr_reg->smin_value : + ptr_reg->umax_value) + ptr_reg->off; break; default: return REASON_TYPE; @@ -4361,10 +4353,12 @@ static int sanitize_ptr_alu(struct bpf_v struct bpf_insn *insn, const struct bpf_reg_state *ptr_reg, const struct bpf_reg_state *off_reg, - struct bpf_reg_state *dst_reg) + struct bpf_reg_state *dst_reg, + struct bpf_insn_aux_data *tmp_aux, + const bool commit_window) { + struct bpf_insn_aux_data *aux = commit_window ? cur_aux(env) : tmp_aux; struct bpf_verifier_state *vstate = env->cur_state; - struct bpf_insn_aux_data *aux = cur_aux(env); bool off_is_neg = off_reg->smin_value < 0; bool ptr_is_dst_reg = ptr_reg == dst_reg; u8 opcode = BPF_OP(insn->code); @@ -4383,18 +4377,33 @@ static int sanitize_ptr_alu(struct bpf_v if (vstate->speculative) goto do_sim; - alu_state = off_is_neg ? BPF_ALU_NEG_VALUE : 0; - alu_state |= ptr_is_dst_reg ? - BPF_ALU_SANITIZE_SRC : BPF_ALU_SANITIZE_DST; - err = retrieve_ptr_limit(ptr_reg, off_reg, &alu_limit, opcode); if (err < 0) return err; + if (commit_window) { + /* In commit phase we narrow the masking window based on + * the observed pointer move after the simulated operation. + */ + alu_state = tmp_aux->alu_state; + alu_limit = abs(tmp_aux->alu_limit - alu_limit); + } else { + alu_state = off_is_neg ? BPF_ALU_NEG_VALUE : 0; + alu_state |= ptr_is_dst_reg ? + BPF_ALU_SANITIZE_SRC : BPF_ALU_SANITIZE_DST; + } + err = update_alu_sanitation_state(aux, alu_state, alu_limit); if (err < 0) return err; do_sim: + /* If we're in commit phase, we're done here given we already + * pushed the truncated dst_reg into the speculative verification + * stack. + */ + if (commit_window) + return 0; + /* Simulate and find potential out-of-bounds access under * speculative execution from truncation as a result of * masking when off was not within expected range. If off @@ -4506,6 +4515,7 @@ static int adjust_ptr_min_max_vals(struc smin_ptr = ptr_reg->smin_value, smax_ptr = ptr_reg->smax_value; u64 umin_val = off_reg->umin_value, umax_val = off_reg->umax_value, umin_ptr = ptr_reg->umin_value, umax_ptr = ptr_reg->umax_value; + struct bpf_insn_aux_data tmp_aux = {}; u8 opcode = BPF_OP(insn->code); u32 dst = insn->dst_reg; int ret; @@ -4564,12 +4574,15 @@ static int adjust_ptr_min_max_vals(struc !check_reg_sane_offset(env, ptr_reg, ptr_reg->type)) return -EINVAL; - switch (opcode) { - case BPF_ADD: - ret = sanitize_ptr_alu(env, insn, ptr_reg, off_reg, dst_reg); + if (sanitize_needed(opcode)) { + ret = sanitize_ptr_alu(env, insn, ptr_reg, off_reg, dst_reg, + &tmp_aux, false); if (ret < 0) return sanitize_err(env, insn, ret, off_reg, dst_reg); + } + switch (opcode) { + case BPF_ADD: /* We can take a fixed offset as long as it doesn't overflow * the s32 'off' field */ @@ -4620,10 +4633,6 @@ static int adjust_ptr_min_max_vals(struc } break; case BPF_SUB: - ret = sanitize_ptr_alu(env, insn, ptr_reg, off_reg, dst_reg); - if (ret < 0) - return sanitize_err(env, insn, ret, off_reg, dst_reg); - if (dst_reg == off_reg) { /* scalar -= pointer. Creates an unknown scalar */ verbose(env, "R%d tried to subtract pointer from scalar\n", @@ -4706,6 +4715,12 @@ static int adjust_ptr_min_max_vals(struc if (sanitize_check_bounds(env, insn, dst_reg) < 0) return -EACCES; + if (sanitize_needed(opcode)) { + ret = sanitize_ptr_alu(env, insn, dst_reg, off_reg, dst_reg, + &tmp_aux, true); + if (ret < 0) + return sanitize_err(env, insn, ret, off_reg, dst_reg); + } return 0; }