linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker
@ 2014-11-14 12:50 Andrey Ryabinin
  2014-11-14 12:50 ` [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log Andrey Ryabinin
                   ` (2 more replies)
  0 siblings, 3 replies; 11+ messages in thread
From: Andrey Ryabinin @ 2014-11-14 12:50 UTC (permalink / raw)
  To: linux-kernel
  Cc: Andrey Ryabinin, Andrew Morton, Peter Zijlstra, Sasha Levin,
	Randy Dunlap, Rasmus Villemoes, Jonathan Corbet, Michal Marek,
	Thomas Gleixner, Ingo Molnar, H. Peter Anvin, Yury Gribov,
	Dmitry Vyukov, Konstantin Khlebnikov, x86, linux-doc,
	linux-kbuild

Next spin of UBSan and there are two patches from now.
"Two there should be; no more, no less. One to embody the UBSan,
the other to not deadlock it."

Patches on top of v3.18-rc4. As usual they are available in git:
	git://github.com/aryabinin/linux ubsan/v2

Changes since v1:
   - Refactoring and cleanups in lib/ubsan.c including Sasha's complains.
   - Some spelling fixes from Randy
   - Fixed possible memory corruption on 64 big endian machines, spotted by Rasmus.
   - Links to the relevant GCC documentation added into changelog (Peter).
   - Added documentation.
   - Fix deadlock caused by kernel/printk/printk.c instrumentation
        (patch "kernel: printk: specify alignment for struct printk_log").
   - Dropped useless 'Indirect call of a function through a function pointer of the wrong type'
     checker. GCC doesn't support this, and as clang manual says it's for C++ only.
   - Added checker for __builtin_unreachable() calls.
   - Removed redundant -fno-sanitize=float-cast-overflow from CFLAGS.
   - Added lock to prevent mixing reports.

Andrey Ryabinin (2):
  kernel: printk: specify alignment for struct printk_log
  UBSan: run-time undefined behavior sanity checker

 Documentation/ubsan.txt               |  69 +++++
 Makefile                              |  10 +-
 arch/x86/Kconfig                      |   1 +
 arch/x86/boot/Makefile                |   1 +
 arch/x86/boot/compressed/Makefile     |   1 +
 arch/x86/realmode/rm/Makefile         |   1 +
 arch/x86/vdso/Makefile                |   2 +
 drivers/firmware/efi/libstub/Makefile |   1 +
 include/linux/sched.h                 |   3 +
 kernel/printk/printk.c                |  10 +-
 lib/Kconfig.debug                     |  22 ++
 lib/Makefile                          |   3 +
 lib/ubsan.c                           | 567 ++++++++++++++++++++++++++++++++++
 lib/ubsan.h                           |  84 +++++
 scripts/Makefile.lib                  |   6 +
 15 files changed, 775 insertions(+), 6 deletions(-)
 create mode 100644 Documentation/ubsan.txt
 create mode 100644 lib/ubsan.c
 create mode 100644 lib/ubsan.h

-- 
2.1.3


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log
  2014-11-14 12:50 [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker Andrey Ryabinin
@ 2014-11-14 12:50 ` Andrey Ryabinin
  2014-11-14 17:22   ` Joe Perches
  2014-11-14 12:50 ` [PATCH v2 2/2] UBSan: run-time undefined behavior sanity checker Andrey Ryabinin
  2015-11-03 17:20 ` [PATCH v2 0/2] " Sasha Levin
  2 siblings, 1 reply; 11+ messages in thread
From: Andrey Ryabinin @ 2014-11-14 12:50 UTC (permalink / raw)
  To: linux-kernel
  Cc: Andrey Ryabinin, Andrew Morton, Peter Zijlstra, Sasha Levin,
	Randy Dunlap, Rasmus Villemoes, Jonathan Corbet, Michal Marek,
	Thomas Gleixner, Ingo Molnar, H. Peter Anvin, Yury Gribov,
	Dmitry Vyukov, Konstantin Khlebnikov, x86, linux-doc,
	linux-kbuild

On architectures that have support for efficient unaligned access
struct printk_log has 4-byte alignment.
Specify alignment attribute in type declaration.

The whole point of this patch is to fix deadlock which happening
when UBSan detects unaligned access in printk() thus UBSan recursively
calls printk() with logbuf_lock held by top printk() call.

Signed-off-by: Andrey Ryabinin <a.ryabinin@samsung.com>
---
 kernel/printk/printk.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index ced2b84..39be027 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -223,7 +223,11 @@ struct printk_log {
 	u8 facility;		/* syslog facility */
 	u8 flags:5;		/* internal record flags */
 	u8 level:3;		/* syslog level */
-};
+}
+#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+__packed __aligned(4)
+#endif
+;
 
 /*
  * The logbuf_lock protects kmsg buffer, indices, counters.  This can be taken
@@ -261,11 +265,7 @@ static u32 clear_idx;
 #define LOG_LINE_MAX		(1024 - PREFIX_MAX)
 
 /* record buffer */
-#if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
-#define LOG_ALIGN 4
-#else
 #define LOG_ALIGN __alignof__(struct printk_log)
-#endif
 #define __LOG_BUF_LEN (1 << CONFIG_LOG_BUF_SHIFT)
 static char __log_buf[__LOG_BUF_LEN] __aligned(LOG_ALIGN);
 static char *log_buf = __log_buf;
-- 
2.1.3


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH v2 2/2] UBSan: run-time undefined behavior sanity checker
  2014-11-14 12:50 [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker Andrey Ryabinin
  2014-11-14 12:50 ` [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log Andrey Ryabinin
@ 2014-11-14 12:50 ` Andrey Ryabinin
  2015-11-03 17:20 ` [PATCH v2 0/2] " Sasha Levin
  2 siblings, 0 replies; 11+ messages in thread
From: Andrey Ryabinin @ 2014-11-14 12:50 UTC (permalink / raw)
  To: linux-kernel
  Cc: Andrey Ryabinin, Andrew Morton, Peter Zijlstra, Sasha Levin,
	Randy Dunlap, Rasmus Villemoes, Jonathan Corbet, Michal Marek,
	Thomas Gleixner, Ingo Molnar, H. Peter Anvin, Yury Gribov,
	Dmitry Vyukov, Konstantin Khlebnikov, x86, linux-doc,
	linux-kbuild

UBSan uses compile-time instrumentation to catch undefined behavior (UB).
Compiler inserts code that perform certain kinds of checks before
operations that could cause UB. If check fails (i.e. UB detected)
__ubsan_handle_* function called to print error message.

So the most of the work is done by compiler. This patch just
implements ubsan handlers printing errors.

GCC has this capability since 4.9 [1] (see -fsanitize=undefined
option and its suboptions).
However upcoming GCC 5.0 has more checkers implemented [2].
Article [3] has a bit more details about UBSan in the GCC.

Different kinds of checks could be enabled via boot parameter:
ubsan_handle=OEAINVBSLU.
If ubsan_handle not present in cmdline default options are used: ELNVBSLU

	O - different kinds of overflows
	E - negation overflow, division overflow, division by zero.
	A - misaligned memory access.
	I - load from/store to an object with insufficient space.
	N - null argument declared with nonnull attribute,
	  returned null from function which never returns null, null ptr dereference.
	V - variable size array with non-positive length
	B - out-of-bounds accesses.
	S - shifting out-of-bounds.
	L - load of invalid value (value out of range for the enum type, loading other then 0/1 to bool type)
	U - __builtin_unreachable() was called

[1] - https://gcc.gnu.org/onlinedocs/gcc-4.9.0/gcc/Debugging-Options.html
[2] - https://gcc.gnu.org/onlinedocs/gcc/Debugging-Options.html
[3] - http://developerblog.redhat.com/2014/10/16/gcc-undefined-behavior-sanitizer-ubsan/

Signed-off-by: Andrey Ryabinin <a.ryabinin@samsung.com>
---
 Documentation/ubsan.txt               |  69 +++++
 Makefile                              |  10 +-
 arch/x86/Kconfig                      |   1 +
 arch/x86/boot/Makefile                |   1 +
 arch/x86/boot/compressed/Makefile     |   1 +
 arch/x86/realmode/rm/Makefile         |   1 +
 arch/x86/vdso/Makefile                |   2 +
 drivers/firmware/efi/libstub/Makefile |   1 +
 include/linux/sched.h                 |   3 +
 lib/Kconfig.debug                     |  22 ++
 lib/Makefile                          |   3 +
 lib/ubsan.c                           | 567 ++++++++++++++++++++++++++++++++++
 lib/ubsan.h                           |  84 +++++
 scripts/Makefile.lib                  |   6 +
 14 files changed, 770 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/ubsan.txt
 create mode 100644 lib/ubsan.c
 create mode 100644 lib/ubsan.h

diff --git a/Documentation/ubsan.txt b/Documentation/ubsan.txt
new file mode 100644
index 0000000..04d6aa6
--- /dev/null
+++ b/Documentation/ubsan.txt
@@ -0,0 +1,69 @@
+Undefined Behavior Sanitizer - UBSan
+
+Overview
+========
+
+UBSan is a runtime undefined behaviour checker.
+
+UBSan uses compile-time instrumentation to catch undefined behavior (UB).
+Compiler inserts code that perform certain kinds of checks before operations
+that may cause UB. If check fails (i.e. UB detected) __ubsan_handle_*
+function called to print error message.
+
+GCC has that feature since 4.9 [1] (see -fsanitize=undefined option and its suboptions).
+Upcoming GCC 5.0 has more checkers implemented [2].
+
+Usage
+=======
+
+To enable UBSAN configure kernel with:
+
+	  CONFIG_UBSAN=y
+
+and to check the entire kernel:
+
+        CONFIG_UBSAN_SANITIZE_ALL=y
+
+To enable instrumentation for specific files or directories, add a line
+similar to the following to the respective kernel Makefile:
+
+        For a single file (e.g. main.o):
+                UBSAN_SANITIZE_main.o := y
+
+        For all files in one directory:
+                UBSAN_SANITIZE := y
+
+To exclude files from being instrumented even if CONFIG_UBSAN_SANITIZE_ALL=y,
+use:
+
+                UBSAN_SANITIZE_main.o := n
+        and:
+                UBSAN_SANITIZE := n
+
+Various types of checkers may be enabled via boot parameter, e.g.:
+     ubsan_handle=OEAINVBSLU
+
+Possible options are:
+        O - different kinds of integer overflows
+        E - negation overflow, division overflow, division by zero.
+        A - misaligned memory access.
+        I - load from/store to an object with insufficient space.
+        N - passing NULL as argument declared with nonnull attribute,
+             returned null from function which never returns null,
+             null ptr dereference.
+        V - variable size array with non-positive length
+        B - out-of-bounds memory accesses
+        S - shifting out-of-bounds
+        L - load of invalid value (value out of range for the enum type,
+              loading other then 0/1 to bool type)
+        U - __builtin_unreachable() was called
+
+If ubsan_handle not present in cmdline the default options are used: ELNVBSLU.
+Additionally option 'A' (misaligned access checks) enabled by default if
+architecture doesn't support unaligned accesses.
+
+References
+==========
+
+[1] - https://gcc.gnu.org/onlinedocs/gcc-4.9.0/gcc/Debugging-Options.html
+[2] - https://gcc.gnu.org/onlinedocs/gcc/Debugging-Options.html
diff --git a/Makefile b/Makefile
index bc7eb6a..007e06b 100644
--- a/Makefile
+++ b/Makefile
@@ -377,6 +377,7 @@ LDFLAGS_MODULE  =
 CFLAGS_KERNEL	=
 AFLAGS_KERNEL	=
 CFLAGS_GCOV	= -fprofile-arcs -ftest-coverage
+CFLAGS_UBSAN	= $(call cc-option, -fsanitize=undefined)
 
 
 # Use USERINCLUDE when you must reference the UAPI directories only.
@@ -421,7 +422,7 @@ export MAKE AWK GENKSYMS INSTALLKERNEL PERL PYTHON UTS_MACHINE
 export HOSTCXX HOSTCXXFLAGS LDFLAGS_MODULE CHECK CHECKFLAGS
 
 export KBUILD_CPPFLAGS NOSTDINC_FLAGS LINUXINCLUDE OBJCOPYFLAGS LDFLAGS
-export KBUILD_CFLAGS CFLAGS_KERNEL CFLAGS_MODULE CFLAGS_GCOV
+export KBUILD_CFLAGS CFLAGS_KERNEL CFLAGS_MODULE CFLAGS_GCOV CFLAGS_UBSAN
 export KBUILD_AFLAGS AFLAGS_KERNEL AFLAGS_MODULE
 export KBUILD_AFLAGS_MODULE KBUILD_CFLAGS_MODULE KBUILD_LDFLAGS_MODULE
 export KBUILD_AFLAGS_KERNEL KBUILD_CFLAGS_KERNEL
@@ -670,6 +671,13 @@ endif
 endif
 KBUILD_CFLAGS += $(stackp-flag)
 
+ifdef CONFIG_UBSAN
+  ifeq ($(CFLAGS_UBSAN),)
+    $(warning Cannot use CONFIG_UBSAN: \
+             -fsanitize=undefined not supported by compiler)
+  endif
+endif
+
 ifeq ($(COMPILER),clang)
 KBUILD_CPPFLAGS += $(call cc-option,-Qunused-arguments,)
 KBUILD_CPPFLAGS += $(call cc-option,-Wno-unknown-warning-option,)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index ded8a67..359d8ba 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -55,6 +55,7 @@ config X86
 	select HAVE_FUNCTION_TRACER
 	select HAVE_FUNCTION_GRAPH_TRACER
 	select HAVE_FUNCTION_GRAPH_FP_TEST
+	select HAVE_ARCH_UBSAN_SANTIZE_ALL
 	select HAVE_SYSCALL_TRACEPOINTS
 	select SYSCTL_EXCEPTION_TRACE
 	select HAVE_KVM
diff --git a/arch/x86/boot/Makefile b/arch/x86/boot/Makefile
index 5b016e2..95bf522 100644
--- a/arch/x86/boot/Makefile
+++ b/arch/x86/boot/Makefile
@@ -57,6 +57,7 @@ endif
 KBUILD_CFLAGS	:= $(USERINCLUDE) $(REALMODE_CFLAGS) -D_SETUP
 KBUILD_AFLAGS	:= $(KBUILD_CFLAGS) -D__ASSEMBLY__
 GCOV_PROFILE := n
+UBSAN_SANITIZE := n
 
 $(obj)/bzImage: asflags-y  := $(SVGA_MODE)
 
diff --git a/arch/x86/boot/compressed/Makefile b/arch/x86/boot/compressed/Makefile
index 704f58a..16940f8 100644
--- a/arch/x86/boot/compressed/Makefile
+++ b/arch/x86/boot/compressed/Makefile
@@ -19,6 +19,7 @@ KBUILD_CFLAGS += $(call cc-option,-fno-stack-protector)
 
 KBUILD_AFLAGS  := $(KBUILD_CFLAGS) -D__ASSEMBLY__
 GCOV_PROFILE := n
+UBSAN_SANITIZE :=n
 
 LDFLAGS := -m elf_$(UTS_MACHINE)
 LDFLAGS_vmlinux := -T
diff --git a/arch/x86/realmode/rm/Makefile b/arch/x86/realmode/rm/Makefile
index 7c0d7be..5c48444 100644
--- a/arch/x86/realmode/rm/Makefile
+++ b/arch/x86/realmode/rm/Makefile
@@ -69,3 +69,4 @@ KBUILD_CFLAGS	:= $(LINUXINCLUDE) $(REALMODE_CFLAGS) -D_SETUP -D_WAKEUP \
 		   -I$(srctree)/arch/x86/boot
 KBUILD_AFLAGS	:= $(KBUILD_CFLAGS) -D__ASSEMBLY__
 GCOV_PROFILE := n
+UBSAN_SANITIZE := n
diff --git a/arch/x86/vdso/Makefile b/arch/x86/vdso/Makefile
index 5a4affe..3054473 100644
--- a/arch/x86/vdso/Makefile
+++ b/arch/x86/vdso/Makefile
@@ -4,6 +4,8 @@
 
 KBUILD_CFLAGS += $(DISABLE_LTO)
 
+UBSAN_SANITIZE := n
+
 VDSO64-$(CONFIG_X86_64)		:= y
 VDSOX32-$(CONFIG_X86_X32_ABI)	:= y
 VDSO32-$(CONFIG_X86_32)		:= y
diff --git a/drivers/firmware/efi/libstub/Makefile b/drivers/firmware/efi/libstub/Makefile
index b14bc2b..7b07624 100644
--- a/drivers/firmware/efi/libstub/Makefile
+++ b/drivers/firmware/efi/libstub/Makefile
@@ -19,6 +19,7 @@ KBUILD_CFLAGS			:= $(cflags-y) \
 				   $(call cc-option,-fno-stack-protector)
 
 GCOV_PROFILE			:= n
+UBSAN_SANITIZE			:= n
 
 lib-y				:= efi-stub-helper.o
 lib-$(CONFIG_EFI_ARMSTUB)	+= arm-stub.o fdt.o
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 5e344bb..36d0e89 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1483,6 +1483,9 @@ struct task_struct {
 	struct held_lock held_locks[MAX_LOCK_DEPTH];
 	gfp_t lockdep_reclaim_gfp;
 #endif
+#ifdef CONFIG_UBSAN
+	unsigned int in_ubsan;
+#endif
 
 /* journalling filesystem info */
 	void *journal_info;
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 4e35a5d..8e80d7c 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -646,6 +646,28 @@ config DEBUG_SHIRQ
 	  Drivers ought to be able to handle interrupts coming in at those
 	  points; some don't and need to be caught.
 
+config HAVE_ARCH_UBSAN_SANTIZE_ALL
+	bool
+
+config UBSAN
+	bool "Undefined behaviour sanity checker"
+	help
+	  This option enables undefined behaviour sanity checker
+	  Compile-time instrumentation is used to detect various undefined
+	  behaviours in runtime. Various types of checks may be enabled
+	  via boot parameter ubsan_handle (see: Documentation/ubsan.txt).
+
+config UBSAN_SANITIZE_ALL
+	bool "Enable instrumentation for the entire kernel"
+	depends on UBSAN
+	depends on HAVE_ARCH_UBSAN_SANTIZE_ALL
+	default y
+	help
+	  This option activates instrumentation for the entire kernel.
+	  If you don't enable this option, you have to explicitly specify
+	  UBSAN_SANITIZE := y for the files/directories you want to check for UB.
+
+
 menu "Debug Lockups and Hangs"
 
 config LOCKUP_DETECTOR
diff --git a/lib/Makefile b/lib/Makefile
index 7512dc9..d8c05c68 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -196,3 +196,6 @@ quiet_cmd_build_OID_registry = GEN     $@
 clean-files	+= oid_registry_data.c
 
 obj-$(CONFIG_UCS2_STRING) += ucs2_string.o
+obj-$(CONFIG_UBSAN) += ubsan.o
+
+UBSAN_SANITIZE_ubsan.o := n
diff --git a/lib/ubsan.c b/lib/ubsan.c
new file mode 100644
index 0000000..d946e79
--- /dev/null
+++ b/lib/ubsan.c
@@ -0,0 +1,567 @@
+/*
+ * UBSan error reporting functions
+ *
+ * Copyright (c) 2014 Samsung Electronics Co., Ltd.
+ * Author: Andrey Ryabinin <a.ryabinin@samsung.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+
+#include <linux/bitops.h>
+#include <linux/bug.h>
+#include <linux/ctype.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/sched.h>
+
+#include "ubsan.h"
+
+const char *type_check_kinds[] = {
+	"load of",
+	"store to",
+	"reference binding to",
+	"member access within",
+	"member call on",
+	"constructor call on",
+	"downcast of",
+	"downcast of"
+};
+
+enum {
+	SUM_OVERFLOW,
+	SUB_OVERFLOW,
+	MUL_OVERFLOW,
+	NEG_OVERFLOW,
+	DIVREM_OVERFLOW,
+	ALIGNMENT,
+	OBJECT_SIZE,
+	NONNULL_ARG,
+	NONNULL_RET,
+	NULL_PTR,
+	VLA_BOUND,
+	OUT_OF_BOUNDS,
+	SHIFT_OUT_OF_BOUNDS,
+	INVALID_LOAD,
+	UNREACHABLE,
+	HANDLERS_END,
+};
+
+/*
+ * By default enable everything except signed overflows and
+ * misaligned accesses
+ */
+static unsigned long ubsan_handle = GENMASK(HANDLERS_END, 0) &
+	~(BIT_MASK(SUM_OVERFLOW) | BIT_MASK(SUB_OVERFLOW) |
+#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+		BIT_MASK(ALIGNMENT) |
+#endif
+		BIT_MASK(MUL_OVERFLOW));
+
+#define REPORTED_BIT 31
+
+#if (BITS_PER_LONG == 64) && defined(__BIG_ENDIAN)
+#define COLUMN_MASK (~(1U << REPORTED_BIT))
+#define LINE_MASK   (~0U)
+#else
+#define COLUMN_MASK   (~0U)
+#define LINE_MASK (~(1U << REPORTED_BIT))
+#endif
+
+#define VALUE_LENGTH 40
+
+static bool was_reported(struct source_location *location)
+{
+	return test_and_set_bit(REPORTED_BIT, &location->reported);
+}
+
+static void print_source_location(const char *prefix,
+				struct source_location *loc)
+{
+	pr_err("%s %s:%d:%d\n", prefix, loc->file_name,
+		loc->line & LINE_MASK, loc->column & COLUMN_MASK);
+}
+
+static void enable_handler(unsigned int handler)
+{
+	set_bit(handler, &ubsan_handle);
+}
+
+static bool handler_enabled(unsigned int handler)
+{
+	return test_bit(handler, &ubsan_handle);
+}
+
+static bool suppress_report(unsigned int handler, struct source_location *loc)
+{
+	return current->in_ubsan || !handler_enabled(handler)
+		|| was_reported(loc);
+}
+
+static bool type_is_int(struct type_descriptor *type)
+{
+	return type->type_kind == type_kind_int;
+}
+
+static bool type_is_signed(struct type_descriptor *type)
+{
+	WARN_ON(!type_is_int(type));
+	return  type->type_info & 1;
+}
+
+static unsigned type_bit_width(struct type_descriptor *type)
+{
+	return 1 << (type->type_info >> 1);
+}
+
+static bool is_inline_int(struct type_descriptor *type)
+{
+	unsigned inline_bits = sizeof(unsigned long)*8;
+	unsigned bits = type_bit_width(type);
+
+	WARN_ON(!type_is_int(type));
+
+	return bits <= inline_bits;
+}
+
+static s_max get_signed_val(struct type_descriptor *type, unsigned long val)
+{
+	if (is_inline_int(type)) {
+		unsigned extra_bits = sizeof(s_max)*8 - type_bit_width(type);
+		return ((s_max)val) << extra_bits >> extra_bits;
+	}
+
+	if (type_bit_width(type) == 64)
+		return *(s64 *)val;
+
+	return *(s_max *)val;
+}
+
+static bool val_is_negative(struct type_descriptor *type, unsigned long val)
+{
+	return type_is_signed(type) && get_signed_val(type, val) < 0;
+}
+
+static u_max get_unsigned_val(struct type_descriptor *type, unsigned long val)
+{
+	if (is_inline_int(type))
+		return val;
+
+	if (type_bit_width(type) == 64)
+		return *(u64 *)val;
+
+	return *(u_max *)val;
+}
+
+static void val_to_string(char *str, size_t size, struct type_descriptor *type,
+	unsigned long value)
+{
+	u_max val = get_unsigned_val(type, value);
+
+	if (type_is_int(type)) {
+		if (type_bit_width(type) == 128)
+			scnprintf(str, size, "0x%08x%08x%08x%08x",
+				(u32)(val >> 96),
+				(u32)(val >> 64),
+				(u32)(val >> 32),
+				(u32)(val));
+		else if (type_is_signed(type))
+			scnprintf(str, size, "%lld",
+				(s64)get_signed_val(type, value));
+		else
+			scnprintf(str, size, "%llu",
+				(u64)get_unsigned_val(type, value));
+	}
+}
+
+static bool location_is_valid(struct source_location *loc)
+{
+	return loc->file_name != NULL;
+}
+
+static DEFINE_SPINLOCK(report_lock);
+
+static void ubsan_prologue(struct source_location *location,
+			unsigned long *flags)
+{
+	current->in_ubsan++;
+	spin_lock_irqsave(&report_lock, *flags);
+
+	pr_err("========================================"
+		"========================================\n");
+	print_source_location("UBSan: Undefined behaviour in", location);
+}
+
+static void ubsan_epilogue(unsigned long *flags)
+{
+	dump_stack();
+	pr_err("========================================"
+		"========================================\n");
+	spin_unlock_irqrestore(&report_lock, *flags);
+	current->in_ubsan--;
+}
+
+static void handle_overflow(struct overflow_data *data, unsigned long lhs,
+			unsigned long rhs, unsigned int handler, char op)
+{
+
+	struct type_descriptor *type = data->type;
+	unsigned long flags;
+	char lhs_val_str[VALUE_LENGTH];
+	char rhs_val_str[VALUE_LENGTH];
+
+	if (suppress_report(handler, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	val_to_string(lhs_val_str, sizeof(lhs_val_str), type, lhs);
+	val_to_string(rhs_val_str, sizeof(rhs_val_str), type, rhs);
+	pr_err("%s integer overflow:\n",
+		type_is_signed(type) ? "signed" : "unsigned");
+	pr_err("%s %c %s cannot be represented in type %s\n",
+		lhs_val_str,
+		op,
+		rhs_val_str,
+		type->type_name);
+
+	ubsan_epilogue(&flags);
+}
+
+void __ubsan_handle_add_overflow(struct overflow_data *data,
+				unsigned long lhs,
+				unsigned long rhs)
+{
+
+	handle_overflow(data, lhs, rhs, SUM_OVERFLOW, '+');
+}
+EXPORT_SYMBOL(__ubsan_handle_add_overflow);
+
+void __ubsan_handle_sub_overflow(struct overflow_data *data,
+				unsigned long lhs,
+				unsigned long rhs)
+{
+	handle_overflow(data, lhs, rhs, SUB_OVERFLOW, '-');
+}
+EXPORT_SYMBOL(__ubsan_handle_sub_overflow);
+
+void __ubsan_handle_mul_overflow(struct overflow_data *data,
+				unsigned long lhs,
+				unsigned long rhs)
+{
+	handle_overflow(data, lhs, rhs, MUL_OVERFLOW, '*');
+}
+EXPORT_SYMBOL(__ubsan_handle_mul_overflow);
+
+void __ubsan_handle_negate_overflow(struct overflow_data *data,
+				unsigned long old_val)
+{
+	unsigned long flags;
+	char old_val_str[VALUE_LENGTH];
+
+	if (suppress_report(NEG_OVERFLOW, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	val_to_string(old_val_str, sizeof(old_val_str), data->type, old_val);
+
+	pr_err("negation of %s cannot be represented in type %s:\n",
+		old_val_str, data->type->type_name);
+
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_negate_overflow);
+
+
+void __ubsan_handle_divrem_overflow(struct overflow_data *data,
+				unsigned long lhs,
+				unsigned long rhs)
+{
+	unsigned long flags;
+	char rhs_val_str[VALUE_LENGTH];
+
+	if (suppress_report(DIVREM_OVERFLOW, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	val_to_string(rhs_val_str, sizeof(rhs_val_str), data->type, rhs);
+
+	if (type_is_signed(data->type) && get_signed_val(data->type, rhs) == -1)
+		pr_err("division of %s by -1 cannot be represented in type %s\n",
+			rhs_val_str, data->type->type_name);
+	else
+		pr_err("division by zero\n");
+
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_divrem_overflow);
+
+static void handle_null_ptr_deref(struct type_mismatch_data *data)
+{
+	unsigned long flags;
+
+	if (suppress_report(NULL_PTR, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	pr_err("%s null pointer of type %s\n",
+		type_check_kinds[data->type_check_kind],
+		data->type->type_name);
+
+	ubsan_epilogue(&flags);
+}
+
+static void handle_missaligned_access(struct type_mismatch_data *data,
+				unsigned long ptr)
+{
+	unsigned long flags;
+
+	if (suppress_report(ALIGNMENT, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	pr_err("%s misaligned address %p for type %s\n",
+		type_check_kinds[data->type_check_kind],
+		(void *)ptr, data->type->type_name);
+	pr_err("which requires %ld byte alignment\n", data->alignment);
+
+	ubsan_epilogue(&flags);
+}
+
+static void handle_object_size_mismatch(struct type_mismatch_data *data,
+					unsigned long ptr)
+{
+	unsigned long flags;
+
+	if (suppress_report(OBJECT_SIZE, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+	pr_err("%s address %pk with insufficient space\n",
+		type_check_kinds[data->type_check_kind],
+		(void *) ptr);
+	pr_err("for an object of type %s\n", data->type->type_name);
+	ubsan_epilogue(&flags);
+}
+
+void __ubsan_handle_type_mismatch(struct type_mismatch_data *data,
+				unsigned long ptr)
+{
+
+	if (!ptr)
+		handle_null_ptr_deref(data);
+	else if (data->alignment && !IS_ALIGNED(ptr, data->alignment))
+		handle_missaligned_access(data, ptr);
+	else
+		handle_object_size_mismatch(data, ptr);
+}
+EXPORT_SYMBOL(__ubsan_handle_type_mismatch);
+
+void __ubsan_handle_nonnull_arg(struct nonnull_arg_data *data)
+{
+	unsigned long flags;
+
+	if (suppress_report(NONNULL_ARG, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	pr_err("null pointer passed as argument %d, declared with nonnull attribute\n",
+		data->arg_index);
+
+	if (location_is_valid(&data->attr_location))
+		print_source_location("nonnull attribute declared in ",
+				&data->attr_location);
+
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_nonnull_arg);
+
+void __ubsan_handle_nonnull_return(struct nonnull_return_data *data)
+{
+	unsigned long flags;
+
+	if (suppress_report(NONNULL_RET, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	pr_err("null pointer returned from function declared to never return null\n");
+
+	if (location_is_valid(&data->attr_location))
+		print_source_location("returns_nonnull attribute specified in",
+				&data->attr_location);
+
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_nonnull_return);
+
+void __ubsan_handle_vla_bound_not_positive(struct vla_bound_data *data,
+					unsigned long bound)
+{
+	unsigned long flags;
+	char bound_str[VALUE_LENGTH];
+
+	if (suppress_report(VLA_BOUND, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	val_to_string(bound_str, sizeof(bound_str), data->type, bound);
+	pr_err("variable length array bound value %s <= 0\n", bound_str);
+
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_vla_bound_not_positive);
+
+void __ubsan_handle_out_of_bounds(struct out_of_bounds_data *data,
+				unsigned long index)
+{
+	unsigned long flags;
+	char index_str[VALUE_LENGTH];
+
+	if (suppress_report(OUT_OF_BOUNDS, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	val_to_string(index_str, sizeof(index_str), data->index_type, index);
+	pr_err("index %s is out of range for type %s\n", index_str,
+		data->array_type->type_name);
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_out_of_bounds);
+
+void __ubsan_handle_shift_out_of_bounds(
+	struct shift_out_of_bounds_data *data,
+	unsigned long lhs,
+	unsigned long rhs)
+{
+	unsigned long flags;
+	struct type_descriptor *rhs_type = data->rhs_type;
+	struct type_descriptor *lhs_type = data->lhs_type;
+	char rhs_str[VALUE_LENGTH];
+	char lhs_str[VALUE_LENGTH];
+
+	if (suppress_report(SHIFT_OUT_OF_BOUNDS, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	val_to_string(rhs_str, sizeof(rhs_str), rhs_type, rhs);
+	val_to_string(lhs_str, sizeof(lhs_str), lhs_type, lhs);
+
+	if (val_is_negative(rhs_type, rhs))
+		pr_err("shift exponent %s is negative\n", rhs_str);
+
+	else if (get_unsigned_val(rhs_type, rhs) >=
+		type_bit_width(lhs_type))
+		pr_err("shift exponent %s is to large for %u-bit type %s\n",
+			rhs_str,
+			type_bit_width(lhs_type),
+			lhs_type->type_name);
+	else if (val_is_negative(lhs_type, lhs))
+		pr_err("left shift of negative value %s\n",
+			lhs_str);
+	else
+		pr_err("left shift of %s by %s places cannot be"
+			"represented in type %s\n",
+			lhs_str, rhs_str,
+			lhs_type->type_name);
+
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_shift_out_of_bounds);
+
+void __ubsan_handle_builtin_unreachable(struct unreachable_data *data)
+{
+	unsigned long flags;
+
+	if (suppress_report(UNREACHABLE, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+	pr_err("calling __builtin_unreachable()\n");
+	ubsan_epilogue(&flags);
+
+}
+EXPORT_SYMBOL(__ubsan_handle_builtin_unreachable);
+
+void __ubsan_handle_load_invalid_value(struct invalid_value_data *data,
+				unsigned long val)
+{
+	unsigned long flags;
+	char val_str[VALUE_LENGTH];
+
+	if (suppress_report(INVALID_LOAD, &data->location))
+		return;
+
+	ubsan_prologue(&data->location, &flags);
+
+	val_to_string(val_str, sizeof(val_str), data->type, val);
+
+	pr_err("load of value %s is not a valid value for type %s\n",
+		val_str, data->type->type_name);
+
+	ubsan_epilogue(&flags);
+}
+EXPORT_SYMBOL(__ubsan_handle_load_invalid_value);
+
+static int __init setup_ubsan_handlers(char *str)
+{
+	ubsan_handle = 0;
+
+	for (; *str; str++) {
+		switch (tolower(*str)) {
+		case 'o':
+			enable_handler(SUM_OVERFLOW);
+			enable_handler(SUB_OVERFLOW);
+			enable_handler(MUL_OVERFLOW);
+			break;
+		case 'e':
+			enable_handler(NEG_OVERFLOW);
+			enable_handler(DIVREM_OVERFLOW);
+			break;
+		case 'a':
+			enable_handler(ALIGNMENT);
+			break;
+		case 'i':
+			enable_handler(OBJECT_SIZE);
+			break;
+		case 'n':
+			enable_handler(NONNULL_ARG);
+			enable_handler(NONNULL_RET);
+			enable_handler(NULL_PTR);
+			break;
+		case 'v':
+			enable_handler(VLA_BOUND);
+			break;
+		case 'b':
+			enable_handler(OUT_OF_BOUNDS);
+			break;
+		case 's':
+			enable_handler(SHIFT_OUT_OF_BOUNDS);
+			break;
+		case 'l':
+			enable_handler(INVALID_LOAD);
+			break;
+		case 'u':
+			enable_handler(UNREACHABLE);
+			break;
+		default:
+			pr_err("skipping unknown option '%c'\n", *str);
+			break;
+		}
+	}
+
+	return 0;
+}
+
+early_param("ubsan_handle", setup_ubsan_handlers);
diff --git a/lib/ubsan.h b/lib/ubsan.h
new file mode 100644
index 0000000..7407127
--- /dev/null
+++ b/lib/ubsan.h
@@ -0,0 +1,84 @@
+#ifndef _LIB_UBSAN_H
+#define _LIB_UBSAN_H
+
+enum {
+	type_kind_int = 0,
+	type_kind_float = 1,
+	type_unknown = 0xffff
+};
+
+struct type_descriptor {
+	u16 type_kind;
+	u16 type_info;
+	char type_name[1];
+};
+
+struct source_location {
+	const char *file_name;
+	union {
+		unsigned long reported;
+		struct {
+			u32 line;
+			u32 column;
+		};
+	};
+};
+
+struct overflow_data {
+	struct source_location location;
+	struct type_descriptor *type;
+};
+
+struct type_mismatch_data {
+	struct source_location location;
+	struct type_descriptor *type;
+	unsigned long alignment;
+	unsigned char type_check_kind;
+};
+
+struct nonnull_arg_data {
+	struct source_location location;
+	struct source_location attr_location;
+	int arg_index;
+};
+
+struct nonnull_return_data {
+	struct source_location location;
+	struct source_location attr_location;
+};
+
+struct vla_bound_data {
+	struct source_location location;
+	struct type_descriptor *type;
+};
+
+struct out_of_bounds_data {
+	struct source_location location;
+	struct type_descriptor *array_type;
+	struct type_descriptor *index_type;
+};
+
+struct shift_out_of_bounds_data {
+	struct source_location location;
+	struct type_descriptor *lhs_type;
+	struct type_descriptor *rhs_type;
+};
+
+struct unreachable_data {
+	struct source_location location;
+};
+
+struct invalid_value_data {
+	struct source_location location;
+	struct type_descriptor *type;
+};
+
+#ifdef CONFIG_ARCH_SUPPORTS_INT128
+typedef __int128 s_max;
+typedef unsigned __int128 u_max;
+#else
+typedef s64 s_max;
+typedef u64 u_max;
+#endif
+
+#endif
diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib
index 54be19a..3f631b9 100644
--- a/scripts/Makefile.lib
+++ b/scripts/Makefile.lib
@@ -119,6 +119,12 @@ _c_flags += $(if $(patsubst n%,, \
 		$(CFLAGS_GCOV))
 endif
 
+ifeq ($(CONFIG_UBSAN),y)
+_c_flags += $(if $(patsubst n%,, \
+		$(UBSAN_SANITIZE_$(basetarget).o)$(UBSAN_SANITIZE)$(CONFIG_UBSAN_SANITIZE_ALL)), \
+		$(CFLAGS_UBSAN))
+endif
+
 # If building the kernel in a separate objtree expand all occurrences
 # of -Idir to -I$(srctree)/dir except for absolute paths (starting with '/').
 
-- 
2.1.3


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log
  2014-11-14 12:50 ` [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log Andrey Ryabinin
@ 2014-11-14 17:22   ` Joe Perches
  2014-11-14 20:54     ` Andrey Ryabinin
  0 siblings, 1 reply; 11+ messages in thread
From: Joe Perches @ 2014-11-14 17:22 UTC (permalink / raw)
  To: Andrey Ryabinin
  Cc: linux-kernel, Andrew Morton, Peter Zijlstra, Sasha Levin,
	Randy Dunlap, Rasmus Villemoes, Jonathan Corbet, Michal Marek,
	Thomas Gleixner, Ingo Molnar, H. Peter Anvin, Yury Gribov,
	Dmitry Vyukov, Konstantin Khlebnikov, x86, linux-doc,
	linux-kbuild

On Fri, 2014-11-14 at 15:50 +0300, Andrey Ryabinin wrote:
> On architectures that have support for efficient unaligned access
> struct printk_log has 4-byte alignment.
> Specify alignment attribute in type declaration.
> 
> The whole point of this patch is to fix deadlock which happening
> when UBSan detects unaligned access in printk() thus UBSan recursively
> calls printk() with logbuf_lock held by top printk() call.
[]
> diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
[]
> @@ -223,7 +223,11 @@ struct printk_log {
>  	u8 facility;		/* syslog facility */
>  	u8 flags:5;		/* internal record flags */
>  	u8 level:3;		/* syslog level */
> -};
> +}
> +#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
> +__packed __aligned(4)
> +#endif

Why is adding __packed useful or __aligned(4) useful?

The struct is naturally aligned on u64 and should be
the size of 2 u64s.

struct printk_log {
	u64 ts_nsec;		/* timestamp in nanoseconds */
	u16 len;		/* length of entire record */
	u16 text_len;		/* length of text buffer */
	u16 dict_len;		/* length of dictionary buffer */
	u8 facility;		/* syslog facility */
	u8 flags:5;		/* internal record flags */
	u8 level:3;		/* syslog level */
};

Is there any case when it's not sizeof(u64) * 2?



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log
  2014-11-14 17:22   ` Joe Perches
@ 2014-11-14 20:54     ` Andrey Ryabinin
  0 siblings, 0 replies; 11+ messages in thread
From: Andrey Ryabinin @ 2014-11-14 20:54 UTC (permalink / raw)
  To: Joe Perches
  Cc: Andrey Ryabinin, LKML, Andrew Morton, Peter Zijlstra,
	Sasha Levin, Randy Dunlap, Rasmus Villemoes, Jonathan Corbet,
	Michal Marek, Thomas Gleixner, Ingo Molnar, H. Peter Anvin,
	Yury Gribov, Dmitry Vyukov, Konstantin Khlebnikov, x86,
	linux-doc, linux-kbuild

2014-11-14 20:22 GMT+03:00 Joe Perches <joe@perches.com>:
> On Fri, 2014-11-14 at 15:50 +0300, Andrey Ryabinin wrote:
>> On architectures that have support for efficient unaligned access
>> struct printk_log has 4-byte alignment.
>> Specify alignment attribute in type declaration.
>>
>> The whole point of this patch is to fix deadlock which happening
>> when UBSan detects unaligned access in printk() thus UBSan recursively
>> calls printk() with logbuf_lock held by top printk() call.
> []
>> diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
> []
>> @@ -223,7 +223,11 @@ struct printk_log {
>>       u8 facility;            /* syslog facility */
>>       u8 flags:5;             /* internal record flags */
>>       u8 level:3;             /* syslog level */
>> -};
>> +}
>> +#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
>> +__packed __aligned(4)
>> +#endif
>
> Why is adding __packed useful or __aligned(4) useful?
>

On x86_64, from compiler's point of view, 'struct printk_log' should
be 8-byte aligned.
But it stored in log buffer only with 4-byte alignment (LOG_ALIGN define).

This inconsistency makes UBSAN unhappy.  UBSAN uses printk to report errors,
thus any access to 4-byte aligned 'struct printk_log' causing
recursive printk() call from ubsan handler.
And if logbuf_lock was taken we end up with deadlock.
Specifying alignment removes this inconsistency. Now compiler knows
that 'printk_log'
actually aligned on 4, and it makes UBSAN happy and silent.

Attribute 'aligned' without attribute 'packed' can only increase alignment.
So __packed is used here because we need to decrease alignment from 8 to 4.


> The struct is naturally aligned on u64 and should be

Not, always. On 32-bits it will be 4-bytes aligned, on 64-bits - 8-bytes aligned

> the size of 2 u64s.
>
> struct printk_log {
>         u64 ts_nsec;            /* timestamp in nanoseconds */
>         u16 len;                /* length of entire record */
>         u16 text_len;           /* length of text buffer */
>         u16 dict_len;           /* length of dictionary buffer */
>         u8 facility;            /* syslog facility */
>         u8 flags:5;             /* internal record flags */
>         u8 level:3;             /* syslog level */
> };
>
> Is there any case when it's not sizeof(u64) * 2?
>

I think no. As I said __packed used for decreasing alignment only.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker
  2014-11-14 12:50 [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker Andrey Ryabinin
  2014-11-14 12:50 ` [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log Andrey Ryabinin
  2014-11-14 12:50 ` [PATCH v2 2/2] UBSan: run-time undefined behavior sanity checker Andrey Ryabinin
@ 2015-11-03 17:20 ` Sasha Levin
  2015-11-03 17:35   ` Dmitry Vyukov
  2015-11-03 18:15   ` Andrey Ryabinin
  2 siblings, 2 replies; 11+ messages in thread
From: Sasha Levin @ 2015-11-03 17:20 UTC (permalink / raw)
  To: Andrey Ryabinin, linux-kernel
  Cc: Andrew Morton, Peter Zijlstra, Randy Dunlap, Rasmus Villemoes,
	Jonathan Corbet, Michal Marek, Thomas Gleixner, Ingo Molnar,
	H. Peter Anvin, Yury Gribov, Dmitry Vyukov,
	Konstantin Khlebnikov, x86, linux-doc, linux-kbuild

Anyone knows why wasn't this merged a year ago? I didn't see any review comments, and it
ended up finding real bugs.

On 11/14/2014 07:50 AM, Andrey Ryabinin wrote:
> Next spin of UBSan and there are two patches from now.
> "Two there should be; no more, no less. One to embody the UBSan,
> the other to not deadlock it."
> 
> Patches on top of v3.18-rc4. As usual they are available in git:
> 	git://github.com/aryabinin/linux ubsan/v2
> 
> Changes since v1:
>    - Refactoring and cleanups in lib/ubsan.c including Sasha's complains.
>    - Some spelling fixes from Randy
>    - Fixed possible memory corruption on 64 big endian machines, spotted by Rasmus.
>    - Links to the relevant GCC documentation added into changelog (Peter).
>    - Added documentation.
>    - Fix deadlock caused by kernel/printk/printk.c instrumentation
>         (patch "kernel: printk: specify alignment for struct printk_log").
>    - Dropped useless 'Indirect call of a function through a function pointer of the wrong type'
>      checker. GCC doesn't support this, and as clang manual says it's for C++ only.
>    - Added checker for __builtin_unreachable() calls.
>    - Removed redundant -fno-sanitize=float-cast-overflow from CFLAGS.
>    - Added lock to prevent mixing reports.
> 
> Andrey Ryabinin (2):
>   kernel: printk: specify alignment for struct printk_log
>   UBSan: run-time undefined behavior sanity checker
> 
>  Documentation/ubsan.txt               |  69 +++++
>  Makefile                              |  10 +-
>  arch/x86/Kconfig                      |   1 +
>  arch/x86/boot/Makefile                |   1 +
>  arch/x86/boot/compressed/Makefile     |   1 +
>  arch/x86/realmode/rm/Makefile         |   1 +
>  arch/x86/vdso/Makefile                |   2 +
>  drivers/firmware/efi/libstub/Makefile |   1 +
>  include/linux/sched.h                 |   3 +
>  kernel/printk/printk.c                |  10 +-
>  lib/Kconfig.debug                     |  22 ++
>  lib/Makefile                          |   3 +
>  lib/ubsan.c                           | 567 ++++++++++++++++++++++++++++++++++
>  lib/ubsan.h                           |  84 +++++
>  scripts/Makefile.lib                  |   6 +
>  15 files changed, 775 insertions(+), 6 deletions(-)
>  create mode 100644 Documentation/ubsan.txt
>  create mode 100644 lib/ubsan.c
>  create mode 100644 lib/ubsan.h
> 


^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker
  2015-11-03 17:20 ` [PATCH v2 0/2] " Sasha Levin
@ 2015-11-03 17:35   ` Dmitry Vyukov
  2015-11-03 17:45     ` Dmitry Vyukov
  2015-11-03 18:15   ` Andrey Ryabinin
  1 sibling, 1 reply; 11+ messages in thread
From: Dmitry Vyukov @ 2015-11-03 17:35 UTC (permalink / raw)
  To: Sasha Levin
  Cc: Andrey Ryabinin, LKML, Andrew Morton, Peter Zijlstra,
	Randy Dunlap, Rasmus Villemoes, Jonathan Corbet, Michal Marek,
	Thomas Gleixner, Ingo Molnar, H. Peter Anvin, Yury Gribov,
	Konstantin Khlebnikov, x86, linux-doc,
	open list:KERNEL BUILD + fi...

I guess just nobody reviewed this. I would like to see this merged as well.



On Tue, Nov 3, 2015 at 6:20 PM, Sasha Levin <sasha.levin@oracle.com> wrote:
> Anyone knows why wasn't this merged a year ago? I didn't see any review comments, and it
> ended up finding real bugs.
>
> On 11/14/2014 07:50 AM, Andrey Ryabinin wrote:
>> Next spin of UBSan and there are two patches from now.
>> "Two there should be; no more, no less. One to embody the UBSan,
>> the other to not deadlock it."
>>
>> Patches on top of v3.18-rc4. As usual they are available in git:
>>       git://github.com/aryabinin/linux ubsan/v2
>>
>> Changes since v1:
>>    - Refactoring and cleanups in lib/ubsan.c including Sasha's complains.
>>    - Some spelling fixes from Randy
>>    - Fixed possible memory corruption on 64 big endian machines, spotted by Rasmus.
>>    - Links to the relevant GCC documentation added into changelog (Peter).
>>    - Added documentation.
>>    - Fix deadlock caused by kernel/printk/printk.c instrumentation
>>         (patch "kernel: printk: specify alignment for struct printk_log").
>>    - Dropped useless 'Indirect call of a function through a function pointer of the wrong type'
>>      checker. GCC doesn't support this, and as clang manual says it's for C++ only.
>>    - Added checker for __builtin_unreachable() calls.
>>    - Removed redundant -fno-sanitize=float-cast-overflow from CFLAGS.
>>    - Added lock to prevent mixing reports.
>>
>> Andrey Ryabinin (2):
>>   kernel: printk: specify alignment for struct printk_log
>>   UBSan: run-time undefined behavior sanity checker
>>
>>  Documentation/ubsan.txt               |  69 +++++
>>  Makefile                              |  10 +-
>>  arch/x86/Kconfig                      |   1 +
>>  arch/x86/boot/Makefile                |   1 +
>>  arch/x86/boot/compressed/Makefile     |   1 +
>>  arch/x86/realmode/rm/Makefile         |   1 +
>>  arch/x86/vdso/Makefile                |   2 +
>>  drivers/firmware/efi/libstub/Makefile |   1 +
>>  include/linux/sched.h                 |   3 +
>>  kernel/printk/printk.c                |  10 +-
>>  lib/Kconfig.debug                     |  22 ++
>>  lib/Makefile                          |   3 +
>>  lib/ubsan.c                           | 567 ++++++++++++++++++++++++++++++++++
>>  lib/ubsan.h                           |  84 +++++
>>  scripts/Makefile.lib                  |   6 +
>>  15 files changed, 775 insertions(+), 6 deletions(-)
>>  create mode 100644 Documentation/ubsan.txt
>>  create mode 100644 lib/ubsan.c
>>  create mode 100644 lib/ubsan.h
>>
>

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker
  2015-11-03 17:35   ` Dmitry Vyukov
@ 2015-11-03 17:45     ` Dmitry Vyukov
  2015-11-03 17:50       ` Sasha Levin
  0 siblings, 1 reply; 11+ messages in thread
From: Dmitry Vyukov @ 2015-11-03 17:45 UTC (permalink / raw)
  To: Sasha Levin, Andrey Ryabinin
  Cc: LKML, Andrew Morton, Peter Zijlstra, Randy Dunlap,
	Rasmus Villemoes, Jonathan Corbet, Michal Marek, Thomas Gleixner,
	Ingo Molnar, H. Peter Anvin, Yury Gribov, Konstantin Khlebnikov,
	x86, linux-doc, open list:KERNEL BUILD + fi...

+Andrey

On Tue, Nov 3, 2015 at 6:35 PM, Dmitry Vyukov <dvyukov@google.com> wrote:
> I guess just nobody reviewed this. I would like to see this merged as well.
>
>
>
> On Tue, Nov 3, 2015 at 6:20 PM, Sasha Levin <sasha.levin@oracle.com> wrote:
>> Anyone knows why wasn't this merged a year ago? I didn't see any review comments, and it
>> ended up finding real bugs.
>>
>> On 11/14/2014 07:50 AM, Andrey Ryabinin wrote:
>>> Next spin of UBSan and there are two patches from now.
>>> "Two there should be; no more, no less. One to embody the UBSan,
>>> the other to not deadlock it."
>>>
>>> Patches on top of v3.18-rc4. As usual they are available in git:
>>>       git://github.com/aryabinin/linux ubsan/v2
>>>
>>> Changes since v1:
>>>    - Refactoring and cleanups in lib/ubsan.c including Sasha's complains.
>>>    - Some spelling fixes from Randy
>>>    - Fixed possible memory corruption on 64 big endian machines, spotted by Rasmus.
>>>    - Links to the relevant GCC documentation added into changelog (Peter).
>>>    - Added documentation.
>>>    - Fix deadlock caused by kernel/printk/printk.c instrumentation
>>>         (patch "kernel: printk: specify alignment for struct printk_log").
>>>    - Dropped useless 'Indirect call of a function through a function pointer of the wrong type'
>>>      checker. GCC doesn't support this, and as clang manual says it's for C++ only.
>>>    - Added checker for __builtin_unreachable() calls.
>>>    - Removed redundant -fno-sanitize=float-cast-overflow from CFLAGS.
>>>    - Added lock to prevent mixing reports.
>>>
>>> Andrey Ryabinin (2):
>>>   kernel: printk: specify alignment for struct printk_log
>>>   UBSan: run-time undefined behavior sanity checker
>>>
>>>  Documentation/ubsan.txt               |  69 +++++
>>>  Makefile                              |  10 +-
>>>  arch/x86/Kconfig                      |   1 +
>>>  arch/x86/boot/Makefile                |   1 +
>>>  arch/x86/boot/compressed/Makefile     |   1 +
>>>  arch/x86/realmode/rm/Makefile         |   1 +
>>>  arch/x86/vdso/Makefile                |   2 +
>>>  drivers/firmware/efi/libstub/Makefile |   1 +
>>>  include/linux/sched.h                 |   3 +
>>>  kernel/printk/printk.c                |  10 +-
>>>  lib/Kconfig.debug                     |  22 ++
>>>  lib/Makefile                          |   3 +
>>>  lib/ubsan.c                           | 567 ++++++++++++++++++++++++++++++++++
>>>  lib/ubsan.h                           |  84 +++++
>>>  scripts/Makefile.lib                  |   6 +
>>>  15 files changed, 775 insertions(+), 6 deletions(-)
>>>  create mode 100644 Documentation/ubsan.txt
>>>  create mode 100644 lib/ubsan.c
>>>  create mode 100644 lib/ubsan.h
>>>
>>

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker
  2015-11-03 17:45     ` Dmitry Vyukov
@ 2015-11-03 17:50       ` Sasha Levin
  0 siblings, 0 replies; 11+ messages in thread
From: Sasha Levin @ 2015-11-03 17:50 UTC (permalink / raw)
  To: Dmitry Vyukov, Andrey Ryabinin
  Cc: LKML, Andrew Morton, Peter Zijlstra, Randy Dunlap,
	Rasmus Villemoes, Jonathan Corbet, Michal Marek, Thomas Gleixner,
	Ingo Molnar, H. Peter Anvin, Yury Gribov, Konstantin Khlebnikov,
	x86, linux-doc, open list:KERNEL BUILD + fi...

On 11/03/2015 12:45 PM, Dmitry Vyukov wrote:
> On Tue, Nov 3, 2015 at 6:35 PM, Dmitry Vyukov <dvyukov@google.com> wrote:
>> > I guess just nobody reviewed this. I would like to see this merged as well.

Hm, at v2 it was working pretty well and the feeling I got was that there are
no further review comments (rather than no one reviewed it).


Thanks,
Sasha

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker
  2015-11-03 17:20 ` [PATCH v2 0/2] " Sasha Levin
  2015-11-03 17:35   ` Dmitry Vyukov
@ 2015-11-03 18:15   ` Andrey Ryabinin
  2015-11-03 18:46     ` Sasha Levin
  1 sibling, 1 reply; 11+ messages in thread
From: Andrey Ryabinin @ 2015-11-03 18:15 UTC (permalink / raw)
  To: Sasha Levin
  Cc: LKML, Andrew Morton, Peter Zijlstra, Randy Dunlap,
	Rasmus Villemoes, Jonathan Corbet, Michal Marek, Thomas Gleixner,
	Ingo Molnar, H. Peter Anvin, Yury Gribov, Dmitry Vyukov,
	Konstantin Khlebnikov, x86, open list:DOCUMENTATION,
	open list:KERNEL BUILD + fi...

2015-11-03 20:20 GMT+03:00 Sasha Levin <sasha.levin@oracle.com>:
> Anyone knows why wasn't this merged a year ago? I didn't see any review comments, and it
> ended up finding real bugs.
>

Probably because I didn't ping so it just lost.
I'll send v3 after the merge window.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker
  2015-11-03 18:15   ` Andrey Ryabinin
@ 2015-11-03 18:46     ` Sasha Levin
  0 siblings, 0 replies; 11+ messages in thread
From: Sasha Levin @ 2015-11-03 18:46 UTC (permalink / raw)
  To: Andrey Ryabinin
  Cc: LKML, Andrew Morton, Peter Zijlstra, Randy Dunlap,
	Rasmus Villemoes, Jonathan Corbet, Michal Marek, Thomas Gleixner,
	Ingo Molnar, H. Peter Anvin, Yury Gribov, Dmitry Vyukov,
	Konstantin Khlebnikov, x86, open list:DOCUMENTATION,
	open list:KERNEL BUILD + fi...

On 11/03/2015 01:15 PM, Andrey Ryabinin wrote:
> 2015-11-03 20:20 GMT+03:00 Sasha Levin <sasha.levin@oracle.com>:
>> Anyone knows why wasn't this merged a year ago? I didn't see any review comments, and it
>> ended up finding real bugs.
>>
> 
> Probably because I didn't ping so it just lost.
> I'll send v3 after the merge window.
> 

Thanks Andrey!

I've just tried v2 on -next, and that causes a hang during boot (seems like we're
hitting undefined behavior while already handling undefined behaviour.


Thanks,
Sasha

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2015-11-03 18:47 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-11-14 12:50 [PATCH v2 0/2] UBSan: run-time undefined behavior sanity checker Andrey Ryabinin
2014-11-14 12:50 ` [PATCH v2 1/2] kernel: printk: specify alignment for struct printk_log Andrey Ryabinin
2014-11-14 17:22   ` Joe Perches
2014-11-14 20:54     ` Andrey Ryabinin
2014-11-14 12:50 ` [PATCH v2 2/2] UBSan: run-time undefined behavior sanity checker Andrey Ryabinin
2015-11-03 17:20 ` [PATCH v2 0/2] " Sasha Levin
2015-11-03 17:35   ` Dmitry Vyukov
2015-11-03 17:45     ` Dmitry Vyukov
2015-11-03 17:50       ` Sasha Levin
2015-11-03 18:15   ` Andrey Ryabinin
2015-11-03 18:46     ` Sasha Levin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).