All of lore.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org,
	syzbot+3d0b2441dbb71751615e@syzkaller.appspotmail.com,
	syzbot+c8504affd4fdd0c1b626@syzkaller.appspotmail.com,
	syzbot+e5190cb881d8660fb1a3@syzkaller.appspotmail.com,
	syzbot+efae31b384d5badbd620@syzkaller.appspotmail.com,
	Daniel Borkmann <daniel@iogearbox.net>,
	Alexei Starovoitov <ast@kernel.org>, Yonghong Song <yhs@fb.com>,
	Edward Cree <ecree@solarflare.com>
Subject: [PATCH 4.14 40/62] bpf: reject passing modified ctx to helper functions
Date: Sat, 11 Jan 2020 10:50:22 +0100	[thread overview]
Message-ID: <20200111094848.163822410@linuxfoundation.org> (raw)
In-Reply-To: <20200111094837.425430968@linuxfoundation.org>

From: Daniel Borkmann <daniel@iogearbox.net>

commit 58990d1ff3f7896ee341030e9a7c2e4002570683 upstream.

As commit 28e33f9d78ee ("bpf: disallow arithmetic operations on
context pointer") already describes, f1174f77b50c ("bpf/verifier:
rework value tracking") removed the specific white-listed cases
we had previously where we would allow for pointer arithmetic in
order to further generalize it, and allow e.g. context access via
modified registers. While the dereferencing of modified context
pointers had been forbidden through 28e33f9d78ee, syzkaller did
recently manage to trigger several KASAN splats for slab out of
bounds access and use after frees by simply passing a modified
context pointer to a helper function which would then do the bad
access since verifier allowed it in adjust_ptr_min_max_vals().

Rejecting arithmetic on ctx pointer in adjust_ptr_min_max_vals()
generally could break existing programs as there's a valid use
case in tracing in combination with passing the ctx to helpers as
bpf_probe_read(), where the register then becomes unknown at
verification time due to adding a non-constant offset to it. An
access sequence may look like the following:

  offset = args->filename;  /* field __data_loc filename */
  bpf_probe_read(&dst, len, (char *)args + offset); // args is ctx

There are two options: i) we could special case the ctx and as
soon as we add a constant or bounded offset to it (hence ctx type
wouldn't change) we could turn the ctx into an unknown scalar, or
ii) we generalize the sanity test for ctx member access into a
small helper and assert it on the ctx register that was passed
as a function argument. Fwiw, latter is more obvious and less
complex at the same time, and one case that may potentially be
legitimate in future for ctx member access at least would be for
ctx to carry a const offset. Therefore, fix follows approach
from ii) and adds test cases to BPF kselftests.

Fixes: f1174f77b50c ("bpf/verifier: rework value tracking")
Reported-by: syzbot+3d0b2441dbb71751615e@syzkaller.appspotmail.com
Reported-by: syzbot+c8504affd4fdd0c1b626@syzkaller.appspotmail.com
Reported-by: syzbot+e5190cb881d8660fb1a3@syzkaller.appspotmail.com
Reported-by: syzbot+efae31b384d5badbd620@syzkaller.appspotmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 kernel/bpf/verifier.c                       |   45 ++++++++++++++-------
 tools/testing/selftests/bpf/test_verifier.c |   58 +++++++++++++++++++++++++++-
 2 files changed, 87 insertions(+), 16 deletions(-)

--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -1251,6 +1251,30 @@ static int check_ptr_alignment(struct bp
 	return check_generic_ptr_alignment(reg, pointer_desc, off, size, strict);
 }
 
+static int check_ctx_reg(struct bpf_verifier_env *env,
+			 const struct bpf_reg_state *reg, int regno)
+{
+	/* Access to ctx or passing it to a helper is only allowed in
+	 * its original, unmodified form.
+	 */
+
+	if (reg->off) {
+		verbose("dereference of modified ctx ptr R%d off=%d disallowed\n",
+			regno, reg->off);
+		return -EACCES;
+	}
+
+	if (!tnum_is_const(reg->var_off) || reg->var_off.value) {
+		char tn_buf[48];
+
+		tnum_strn(tn_buf, sizeof(tn_buf), reg->var_off);
+		verbose("variable ctx access var_off=%s disallowed\n", tn_buf);
+		return -EACCES;
+	}
+
+	return 0;
+}
+
 /* truncate register to smaller size (in bytes)
  * must be called with size < BPF_REG_SIZE
  */
@@ -1320,22 +1344,10 @@ static int check_mem_access(struct bpf_v
 			verbose("R%d leaks addr into ctx\n", value_regno);
 			return -EACCES;
 		}
-		/* ctx accesses must be at a fixed offset, so that we can
-		 * determine what type of data were returned.
-		 */
-		if (reg->off) {
-			verbose("dereference of modified ctx ptr R%d off=%d+%d, ctx+const is allowed, ctx+const+const is not\n",
-				regno, reg->off, off - reg->off);
-			return -EACCES;
-		}
-		if (!tnum_is_const(reg->var_off) || reg->var_off.value) {
-			char tn_buf[48];
+		err = check_ctx_reg(env, reg, regno);
+		if (err < 0)
+			return err;
 
-			tnum_strn(tn_buf, sizeof(tn_buf), reg->var_off);
-			verbose("variable ctx access var_off=%s off=%d size=%d",
-				tn_buf, off, size);
-			return -EACCES;
-		}
 		err = check_ctx_access(env, insn_idx, off, size, t, &reg_type);
 		if (!err && t == BPF_READ && value_regno >= 0) {
 			/* ctx access returns either a scalar, or a
@@ -1573,6 +1585,9 @@ static int check_func_arg(struct bpf_ver
 		expected_type = PTR_TO_CTX;
 		if (type != expected_type)
 			goto err_type;
+		err = check_ctx_reg(env, reg, regno);
+		if (err < 0)
+			return err;
 	} else if (arg_type == ARG_PTR_TO_MEM ||
 		   arg_type == ARG_PTR_TO_UNINIT_MEM) {
 		expected_type = PTR_TO_STACK;
--- a/tools/testing/selftests/bpf/test_verifier.c
+++ b/tools/testing/selftests/bpf/test_verifier.c
@@ -7281,7 +7281,7 @@ static struct bpf_test tests[] = {
 				    offsetof(struct __sk_buff, mark)),
 			BPF_EXIT_INSN(),
 		},
-		.errstr = "dereference of modified ctx ptr R1 off=68+8, ctx+const is allowed, ctx+const+const is not",
+		.errstr = "dereference of modified ctx ptr",
 		.result = REJECT,
 		.prog_type = BPF_PROG_TYPE_SCHED_CLS,
 	},
@@ -7944,6 +7944,62 @@ static struct bpf_test tests[] = {
 		.errstr = "BPF_XADD stores into R2 packet",
 		.prog_type = BPF_PROG_TYPE_XDP,
 	},
+	{
+		"pass unmodified ctx pointer to helper",
+		.insns = {
+			BPF_MOV64_IMM(BPF_REG_2, 0),
+			BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0,
+				     BPF_FUNC_csum_update),
+			BPF_MOV64_IMM(BPF_REG_0, 0),
+			BPF_EXIT_INSN(),
+		},
+		.prog_type = BPF_PROG_TYPE_SCHED_CLS,
+		.result = ACCEPT,
+	},
+	{
+		"pass modified ctx pointer to helper, 1",
+		.insns = {
+			BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -612),
+			BPF_MOV64_IMM(BPF_REG_2, 0),
+			BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0,
+				     BPF_FUNC_csum_update),
+			BPF_MOV64_IMM(BPF_REG_0, 0),
+			BPF_EXIT_INSN(),
+		},
+		.prog_type = BPF_PROG_TYPE_SCHED_CLS,
+		.result = REJECT,
+		.errstr = "dereference of modified ctx ptr",
+	},
+	{
+		"pass modified ctx pointer to helper, 2",
+		.insns = {
+			BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -612),
+			BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0,
+				     BPF_FUNC_get_socket_cookie),
+			BPF_MOV64_IMM(BPF_REG_0, 0),
+			BPF_EXIT_INSN(),
+		},
+		.result_unpriv = REJECT,
+		.result = REJECT,
+		.errstr_unpriv = "dereference of modified ctx ptr",
+		.errstr = "dereference of modified ctx ptr",
+	},
+	{
+		"pass modified ctx pointer to helper, 3",
+		.insns = {
+			BPF_LDX_MEM(BPF_W, BPF_REG_3, BPF_REG_1, 0),
+			BPF_ALU64_IMM(BPF_AND, BPF_REG_3, 4),
+			BPF_ALU64_REG(BPF_ADD, BPF_REG_1, BPF_REG_3),
+			BPF_MOV64_IMM(BPF_REG_2, 0),
+			BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0,
+				     BPF_FUNC_csum_update),
+			BPF_MOV64_IMM(BPF_REG_0, 0),
+			BPF_EXIT_INSN(),
+		},
+		.prog_type = BPF_PROG_TYPE_SCHED_CLS,
+		.result = REJECT,
+		.errstr = "variable ctx access var_off=(0x0; 0x4)",
+	},
 };
 
 static int probe_filter_length(const struct bpf_insn *fp)



  parent reply	other threads:[~2020-01-11 10:12 UTC|newest]

Thread overview: 67+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-11  9:49 [PATCH 4.14 00/62] 4.14.164-stable review Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 01/62] USB: dummy-hcd: use usb_urb_dir_in instead of usb_pipein Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 02/62] USB: dummy-hcd: increase max number of devices to 32 Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 03/62] locking/spinlock/debug: Fix various data races Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 04/62] netfilter: ctnetlink: netns exit must wait for callbacks Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 05/62] mwifiex: Fix heap overflow in mmwifiex_process_tdls_action_frame() Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 06/62] libtraceevent: Fix lib installation with O= Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 07/62] x86/efi: Update e820 with reserved EFI boot services data to fix kexec breakage Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 08/62] efi/gop: Return EFI_NOT_FOUND if there are no usable GOPs Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 09/62] efi/gop: Return EFI_SUCCESS if a usable GOP was found Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 10/62] efi/gop: Fix memory leak in __gop_query32/64() Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 11/62] ARM: vexpress: Set-up shared OPP table instead of individual for each CPU Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 12/62] netfilter: uapi: Avoid undefined left-shift in xt_sctp.h Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 13/62] netfilter: nf_tables: validate NFT_SET_ELEM_INTERVAL_END Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 14/62] ARM: dts: Cygnus: Fix MDIO node address/size cells Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 15/62] spi: spi-cavium-thunderx: Add missing pci_release_regions() Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 16/62] ASoC: topology: Check return value for soc_tplg_pcm_create() Greg Kroah-Hartman
2020-01-11  9:49 ` [PATCH 4.14 17/62] ARM: dts: bcm283x: Fix critical trip point Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 18/62] bpf, mips: Limit to 33 tail calls Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 19/62] ARM: dts: am437x-gp/epos-evm: fix panel compatible Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 20/62] samples: bpf: Replace symbol compare of trace_event Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 21/62] samples: bpf: fix syscall_tp due to unused syscall Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 22/62] powerpc: Ensure that swiotlb buffer is allocated from low memory Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 23/62] bnx2x: Do not handle requests from VFs after parity Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 24/62] bnx2x: Fix logic to get total no. of PFs per engine Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 25/62] net: usb: lan78xx: Fix error message format specifier Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 26/62] rfkill: Fix incorrect check to avoid NULL pointer dereference Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 27/62] ASoC: wm8962: fix lambda value Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 28/62] regulator: rn5t618: fix module aliases Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 29/62] kconfig: dont crash on NULL expressions in expr_eq() Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 30/62] perf/x86/intel: Fix PT PMI handling Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 31/62] fs: avoid softlockups in s_inodes iterators Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 32/62] net: stmmac: Do not accept invalid MTU values Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 33/62] net: stmmac: RX buffer size must be 16 byte aligned Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 34/62] s390/dasd/cio: Interpret ccw_device_get_mdc return value correctly Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 35/62] s390/dasd: fix memleak in path handling error case Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 36/62] block: fix memleak when __blk_rq_map_user_iov() is failed Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 37/62] parisc: Fix compiler warnings in debug_core.c Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 38/62] llc2: Fix return statement of llc_stat_ev_rx_null_dsap_xid_c (and _test_c) Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 39/62] hv_netvsc: Fix unwanted rx_table reset Greg Kroah-Hartman
2020-01-11  9:50 ` Greg Kroah-Hartman [this message]
2020-01-11  9:50 ` [PATCH 4.14 41/62] bpf: Fix passing modified ctx to ld/abs/ind instruction Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 42/62] PCI/switchtec: Read all 64 bits of part_event_bitmap Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 43/62] mmc: block: Convert RPMB to a character device Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 44/62] mmc: block: Delete mmc_access_rpmb() Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 45/62] mmc: block: Fix bug when removing RPMB chardev Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 46/62] mmc: core: Prevent bus reference leak in mmc_blk_init() Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 47/62] mmc: block: propagate correct returned value in mmc_rpmb_ioctl Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 48/62] gtp: fix bad unlock balance in gtp_encap_enable_socket Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 49/62] macvlan: do not assume mac_header is set in macvlan_broadcast() Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 50/62] net: dsa: mv88e6xxx: Preserve priority when setting CPU port Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 51/62] net: stmmac: dwmac-sun8i: Allow all RGMII modes Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 52/62] net: stmmac: dwmac-sunxi: " Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 53/62] net: usb: lan78xx: fix possible skb leak Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 54/62] pkt_sched: fq: do not accept silly TCA_FQ_QUANTUM Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 55/62] USB: core: fix check for duplicate endpoints Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 56/62] USB: serial: option: add Telit ME910G1 0x110a composition Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 57/62] sctp: free cmd->obj.chunk for the unprocessed SCTP_CMD_REPLY Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 58/62] tcp: fix "old stuff" D-SACK causing SACK to be treated as D-SACK Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 59/62] vxlan: fix tos value before xmit Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 60/62] vlan: vlan_changelink() should propagate errors Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 61/62] net: sch_prio: When ungrafting, replace with FIFO Greg Kroah-Hartman
2020-01-11  9:50 ` [PATCH 4.14 62/62] vlan: fix memory leak in vlan_dev_set_egress_priority Greg Kroah-Hartman
2020-01-11 14:55 ` [PATCH 4.14 00/62] 4.14.164-stable review Guenter Roeck
2020-01-11 18:38 ` Naresh Kamboju
2020-01-13 15:47 ` Jon Hunter
2020-01-13 15:47   ` Jon Hunter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200111094848.163822410@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=ast@kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=ecree@solarflare.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=syzbot+3d0b2441dbb71751615e@syzkaller.appspotmail.com \
    --cc=syzbot+c8504affd4fdd0c1b626@syzkaller.appspotmail.com \
    --cc=syzbot+e5190cb881d8660fb1a3@syzkaller.appspotmail.com \
    --cc=syzbot+efae31b384d5badbd620@syzkaller.appspotmail.com \
    --cc=yhs@fb.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.