[PATCH bpf v2 0/3] Restrict bpf_probe_read{,str}() and bpf_trace_printk()'s %s

[PATCH bpf v2 0/3] Restrict bpf_probe_read{,str}() and bpf_trace_printk()'s %s

[Daniel Borkmann]

Small set of fixes in order to restrict BPF helpers for tracing which are
broken on archs with overlapping address ranges as per discussion in [0].
I've targetted this for -bpf tree so they can be routed as fixes. Thanks!

v1 -> v2:
  - switch to reusable %pks, %pus format specifiers (Yonghong)
  - fixate %s on kernel_ds probing for archs with overlapping addr space

  [0] https://lore.kernel.org/bpf/CAHk-=wjJKo0GVixYLmqPn-Q22WFu0xHaBSjKEo7e7Yw72y5SPQ@mail.gmail.com/T/

Daniel Borkmann (3):
  bpf: restrict bpf_probe_read{,str}() only to archs where they work
  bpf: add bpf_probe_read_{user, kernel}_str() to do_refine_retval_range
  bpf: restrict bpf_trace_printk()'s %s usage and add %pks, %pus specifier

 Documentation/core-api/printk-formats.rst |  14 +++
 arch/arm/Kconfig                          |   1 +
 arch/arm64/Kconfig                        |   1 +
 arch/x86/Kconfig                          |   1 +
 init/Kconfig                              |   3 +
 kernel/bpf/verifier.c                     |   4 +-
 kernel/trace/bpf_trace.c                  | 100 ++++++++++++++--------
 lib/vsprintf.c                            |  12 +++
 8 files changed, 101 insertions(+), 35 deletions(-)

-- 
2.21.0

[PATCH bpf v2 1/3] bpf: restrict bpf_probe_read{,str}() only to archs where they work

[Daniel Borkmann]

Given the legacy bpf_probe_read{,str}() BPF helpers are broken on archs
with overlapping address ranges, we should really take the next step to
disable them from BPF use there.

To generally fix the situation, we've recently added new helper variants
bpf_probe_read_{user,kernel}() and bpf_probe_read_{user,kernel}_str().
For details on them, see 6ae08ae3dea2 ("bpf: Add probe_read_{user, kernel}
and probe_read_{user,kernel}_str helpers").

Given bpf_probe_read{,str}() have been around for ~5 years by now, there
are plenty of users at least on x86 still relying on them today, so we
cannot remove them entirely w/o breaking the BPF tracing ecosystem.

However, their use should be restricted to archs with non-overlapping
address ranges where they are working in their current form. Therefore,
move this behind a CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE and
have x86, arm64, arm select it (other archs supporting it can follow-up
on it as well).

For the remaining archs, they can workaround easily by relying on the
feature probe from bpftool which spills out defines that can be used out
of BPF C code to implement the drop-in replacement for old/new kernels
via: bpftool feature probe macro

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
---
 arch/arm/Kconfig         | 1 +
 arch/arm64/Kconfig       | 1 +
 arch/x86/Kconfig         | 1 +
 init/Kconfig             | 3 +++
 kernel/trace/bpf_trace.c | 6 ++++--
 5 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 66a04f6f4775..c77c93c485a0 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -12,6 +12,7 @@ config ARM
 	select ARCH_HAS_KEEPINITRD
 	select ARCH_HAS_KCOV
 	select ARCH_HAS_MEMBARRIER_SYNC_CORE
+	select ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
 	select ARCH_HAS_PTE_SPECIAL if ARM_LPAE
 	select ARCH_HAS_PHYS_TO_DMA
 	select ARCH_HAS_SETUP_DMA_OPS
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 40fb05d96c60..5d513f461957 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -20,6 +20,7 @@ config ARM64
 	select ARCH_HAS_KCOV
 	select ARCH_HAS_KEEPINITRD
 	select ARCH_HAS_MEMBARRIER_SYNC_CORE
+	select ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
 	select ARCH_HAS_PTE_DEVMAP
 	select ARCH_HAS_PTE_SPECIAL
 	select ARCH_HAS_SETUP_DMA_OPS
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 1197b5596d5a..2d3f963fd6f1 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -68,6 +68,7 @@ config X86
 	select ARCH_HAS_KCOV			if X86_64
 	select ARCH_HAS_MEM_ENCRYPT
 	select ARCH_HAS_MEMBARRIER_SYNC_CORE
+	select ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
 	select ARCH_HAS_PMEM_API		if X86_64
 	select ARCH_HAS_PTE_DEVMAP		if X86_64
 	select ARCH_HAS_PTE_SPECIAL
diff --git a/init/Kconfig b/init/Kconfig
index 9e22ee8fbd75..6fd13a051342 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -2279,6 +2279,9 @@ config ASN1
 
 source "kernel/Kconfig.locks"
 
+config ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
+	bool
+
 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
 	bool
 
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index ca1796747a77..b83bdaa31c7b 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -825,14 +825,16 @@ bpf_tracing_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_probe_read_user_proto;
 	case BPF_FUNC_probe_read_kernel:
 		return &bpf_probe_read_kernel_proto;
-	case BPF_FUNC_probe_read:
-		return &bpf_probe_read_compat_proto;
 	case BPF_FUNC_probe_read_user_str:
 		return &bpf_probe_read_user_str_proto;
 	case BPF_FUNC_probe_read_kernel_str:
 		return &bpf_probe_read_kernel_str_proto;
+#ifdef CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
+	case BPF_FUNC_probe_read:
+		return &bpf_probe_read_compat_proto;
 	case BPF_FUNC_probe_read_str:
 		return &bpf_probe_read_compat_str_proto;
+#endif
 #ifdef CONFIG_CGROUPS
 	case BPF_FUNC_get_current_cgroup_id:
 		return &bpf_get_current_cgroup_id_proto;
-- 
2.21.0

[PATCH bpf v2 2/3] bpf: add bpf_probe_read_{user, kernel}_str() to do_refine_retval_range

[Daniel Borkmann]

Given bpf_probe_read{,str}() BPF helpers are now only available under
CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE, we need to add the drop-in
replacements of bpf_probe_read_{kernel,user}_str() to do_refine_retval_range()
as well to avoid hitting the same issue as in 849fa50662fbc ("bpf/verifier:
refine retval R0 state for bpf_get_stack helper").

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Yonghong Song <yhs@fb.com>
---
 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 fa1d8245b925..ac922b0d3b00 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -4340,7 +4340,9 @@ static void do_refine_retval_range(struct bpf_reg_state *regs, int ret_type,
 
 	if (ret_type != RET_INTEGER ||
 	    (func_id != BPF_FUNC_get_stack &&
-	     func_id != BPF_FUNC_probe_read_str))
+	     func_id != BPF_FUNC_probe_read_str &&
+	     func_id != BPF_FUNC_probe_read_kernel_str &&
+	     func_id != BPF_FUNC_probe_read_user_str))
 		return;
 
 	ret_reg->smax_value = meta->msize_max_value;
-- 
2.21.0

[PATCH bpf v2 3/3] bpf: restrict bpf_trace_printk()'s %s usage and add %pks, %pus specifier

[Daniel Borkmann]

Usage of plain %s conversion specifier in bpf_trace_printk() suffers from the
very same issue as bpf_probe_read{,str}() helpers, that is, it is broken on
archs with overlapping address ranges.

While the helpers have been addressed through work in 6ae08ae3dea2 ("bpf: Add
probe_read_{user, kernel} and probe_read_{user, kernel}_str helpers"), we need
an option for bpf_trace_printk() as well to fix it.

Similarly as with the helpers, force users to make an explicit choice by adding
%pks and %pus specifier to bpf_trace_printk() which will then pick the corresponding
strncpy_from_unsafe*() variant to perform the access under KERNEL_DS or USER_DS.
The %pk* (kernel specifier) and %pu* (user specifier) can later also be extended
for other objects aside strings that are probed and printed under tracing, and
reused out of other facilities like bpf_seq_printf() or BTF based type printing.

Existing behavior of %s for current users is still kept working for archs where it
is not broken and therefore gated through CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE.
For archs not having this property we fall-back to pick probing under KERNEL_DS as
a sensible default.

Fixes: 8d3b7dce8622 ("bpf: add support for %s specifier to bpf_trace_printk()")
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Reported-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
---
 Documentation/core-api/printk-formats.rst | 14 ++++
 kernel/trace/bpf_trace.c                  | 94 +++++++++++++++--------
 lib/vsprintf.c                            | 12 +++
 3 files changed, 88 insertions(+), 32 deletions(-)

diff --git a/Documentation/core-api/printk-formats.rst b/Documentation/core-api/printk-formats.rst
index 8ebe46b1af39..5dfcc4592b23 100644
--- a/Documentation/core-api/printk-formats.rst
+++ b/Documentation/core-api/printk-formats.rst
@@ -112,6 +112,20 @@ used when printing stack backtraces. The specifier takes into
 consideration the effect of compiler optimisations which may occur
 when tail-calls are used and marked with the noreturn GCC attribute.
 
+Probed Pointers from BPF / tracing
+----------------------------------
+
+::
+
+	%pks	kernel string
+	%pus	user string
+
+The ``k`` and ``u`` specifiers are used for printing prior probed memory from
+either kernel memory (k) or user memory (u). The subsequent ``s`` specifier
+results in printing a string. For direct use in regular vsnprintf() the (k)
+and (u) annotation is ignored, however, when used out of BPF's bpf_trace_printk(),
+for example, it reads the memory it is pointing to without faulting.
+
 Kernel Pointers
 ---------------
 
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index b83bdaa31c7b..a010edc37ee0 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -323,17 +323,15 @@ static const struct bpf_func_proto *bpf_get_probe_write_proto(void)
 
 /*
  * Only limited trace_printk() conversion specifiers allowed:
- * %d %i %u %x %ld %li %lu %lx %lld %lli %llu %llx %p %s
+ * %d %i %u %x %ld %li %lu %lx %lld %lli %llu %llx %p %pks %pus %s
  */
 BPF_CALL_5(bpf_trace_printk, char *, fmt, u32, fmt_size, u64, arg1,
 	   u64, arg2, u64, arg3)
 {
+	int i, mod[3] = {}, fmt_cnt = 0;
+	char buf[64], fmt_ptype;
+	void *unsafe_ptr = NULL;
 	bool str_seen = false;
-	int mod[3] = {};
-	int fmt_cnt = 0;
-	u64 unsafe_addr;
-	char buf[64];
-	int i;
 
 	/*
 	 * bpf_check()->check_func_arg()->check_stack_boundary()
@@ -359,40 +357,71 @@ BPF_CALL_5(bpf_trace_printk, char *, fmt, u32, fmt_size, u64, arg1,
 		if (fmt[i] == 'l') {
 			mod[fmt_cnt]++;
 			i++;
-		} else if (fmt[i] == 'p' || fmt[i] == 's') {
+		} else if (fmt[i] == 'p') {
 			mod[fmt_cnt]++;
+			if ((fmt[i + 1] == 'k' ||
+			     fmt[i + 1] == 'u') &&
+			    fmt[i + 2] == 's') {
+				fmt_ptype = fmt[i + 1];
+				i += 2;
+				goto fmt_str;
+			}
+
 			/* disallow any further format extensions */
 			if (fmt[i + 1] != 0 &&
 			    !isspace(fmt[i + 1]) &&
 			    !ispunct(fmt[i + 1]))
 				return -EINVAL;
-			fmt_cnt++;
-			if (fmt[i] == 's') {
-				if (str_seen)
-					/* allow only one '%s' per fmt string */
-					return -EINVAL;
-				str_seen = true;
-
-				switch (fmt_cnt) {
-				case 1:
-					unsafe_addr = arg1;
-					arg1 = (long) buf;
-					break;
-				case 2:
-					unsafe_addr = arg2;
-					arg2 = (long) buf;
-					break;
-				case 3:
-					unsafe_addr = arg3;
-					arg3 = (long) buf;
-					break;
-				}
-				buf[0] = 0;
-				strncpy_from_unsafe(buf,
-						    (void *) (long) unsafe_addr,
+
+			goto fmt_next;
+		} else if (fmt[i] == 's') {
+			mod[fmt_cnt]++;
+			fmt_ptype = fmt[i];
+fmt_str:
+			if (str_seen)
+				/* allow only one '%s' per fmt string */
+				return -EINVAL;
+			str_seen = true;
+
+			if (fmt[i + 1] != 0 &&
+			    !isspace(fmt[i + 1]) &&
+			    !ispunct(fmt[i + 1]))
+				return -EINVAL;
+
+			switch (fmt_cnt) {
+			case 0:
+				unsafe_ptr = (void *)(long)arg1;
+				arg1 = (long)buf;
+				break;
+			case 1:
+				unsafe_ptr = (void *)(long)arg2;
+				arg2 = (long)buf;
+				break;
+			case 2:
+				unsafe_ptr = (void *)(long)arg3;
+				arg3 = (long)buf;
+				break;
+			}
+
+			buf[0] = 0;
+			switch (fmt_ptype) {
+			case 's':
+#ifdef CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
+				strncpy_from_unsafe(buf, unsafe_ptr,
 						    sizeof(buf));
+				break;
+#endif
+			case 'k':
+				strncpy_from_unsafe_strict(buf, unsafe_ptr,
+							   sizeof(buf));
+				break;
+			case 'u':
+				strncpy_from_unsafe_user(buf,
+					(__force void __user *)unsafe_ptr,
+							 sizeof(buf));
+				break;
 			}
-			continue;
+			goto fmt_next;
 		}
 
 		if (fmt[i] == 'l') {
@@ -403,6 +432,7 @@ BPF_CALL_5(bpf_trace_printk, char *, fmt, u32, fmt_size, u64, arg1,
 		if (fmt[i] != 'i' && fmt[i] != 'd' &&
 		    fmt[i] != 'u' && fmt[i] != 'x')
 			return -EINVAL;
+fmt_next:
 		fmt_cnt++;
 	}
 
diff --git a/lib/vsprintf.c b/lib/vsprintf.c
index 7c488a1ce318..532b6606a18a 100644
--- a/lib/vsprintf.c
+++ b/lib/vsprintf.c
@@ -2168,6 +2168,10 @@ char *fwnode_string(char *buf, char *end, struct fwnode_handle *fwnode,
  *		f full name
  *		P node name, including a possible unit address
  * - 'x' For printing the address. Equivalent to "%lx".
+ * - '[ku]s' For a BPF/tracing related format specifier, e.g. used out of
+ *           bpf_trace_printk() where [ku] prefix specifies either kernel (k)
+ *           or user (u) memory to probe, and:
+ *              s a string, equivalent to "%s" on direct vsnprintf() use
  *
  * ** When making changes please also update:
  *	Documentation/core-api/printk-formats.rst
@@ -2251,6 +2255,14 @@ char *pointer(const char *fmt, char *buf, char *end, void *ptr,
 		if (!IS_ERR(ptr))
 			break;
 		return err_ptr(buf, end, ptr, spec);
+	case 'u':
+	case 'k':
+		switch (fmt[1]) {
+		case 's':
+			return string(buf, end, ptr, spec);
+		default:
+			return error_string(buf, end, "(einval)", spec);
+		}
 	}
 
 	/* default is to _not_ leak addresses, hash before printing */
-- 
2.21.0

Re: [PATCH bpf v2 0/3] Restrict bpf_probe_read{,str}() and bpf_trace_printk()'s %s

[Alexei Starovoitov]

On Fri, May 15, 2020 at 3:11 AM Daniel Borkmann <daniel@iogearbox.net> wrote:
>
> Small set of fixes in order to restrict BPF helpers for tracing which are
> broken on archs with overlapping address ranges as per discussion in [0].
> I've targetted this for -bpf tree so they can be routed as fixes. Thanks!
>
> v1 -> v2:
>   - switch to reusable %pks, %pus format specifiers (Yonghong)
>   - fixate %s on kernel_ds probing for archs with overlapping addr space
>
>   [0] https://lore.kernel.org/bpf/CAHk-=wjJKo0GVixYLmqPn-Q22WFu0xHaBSjKEo7e7Yw72y5SPQ@mail.gmail.com/T/

Applied to bpf tree. Thanks
We'll send pr to Dave soon to hopefully get this and other fixes
into 5.7 before Sunday rc.