From f8a53573a05fbcc863593d61b22c7cf86fb1c416 Mon Sep 17 00:00:00 2001
From: Chris Down <chris@chrisdown.name>
Date: Tue, 15 Jun 2021 17:52:53 +0100
Subject: [PATCH] printk: Userspace format indexing support
References: bnc#1189998 (PREEMPT_RT prerequisite backports)
Patch-mainline: v5.15-rc1
Git-commit: 337015573718b161891a3473d25f59273f2e626b
We have a number of systems industry-wide that have a subset of their
functionality that works as follows:
1. Receive a message from local kmsg, serial console, or netconsole;
2. Apply a set of rules to classify the message;
3. Do something based on this classification (like scheduling a
remediation for the machine), rinse, and repeat.
As a couple of examples of places we have this implemented just inside
Facebook, although this isn't a Facebook-specific problem, we have this
inside our netconsole processing (for alarm classification), and as part
of our machine health checking. We use these messages to determine
fairly important metrics around production health, and it's important
that we get them right.
While for some kinds of issues we have counters, tracepoints, or metrics
with a stable interface which can reliably indicate the issue, in order
to react to production issues quickly we need to work with the interface
which most kernel developers naturally use when developing: printk.
Most production issues come from unexpected phenomena, and as such
usually the code in question doesn't have easily usable tracepoints or
other counters available for the specific problem being mitigated. We
have a number of lines of monitoring defence against problems in
production (host metrics, process metrics, service metrics, etc), and
where it's not feasible to reliably monitor at another level, this kind
of pragmatic netconsole monitoring is essential.
As one would expect, monitoring using printk is rather brittle for a
number of reasons -- most notably that the message might disappear
entirely in a new version of the kernel, or that the message may change
in some way that the regex or other classification methods start to
silently fail.
One factor that makes this even harder is that, under normal operation,
many of these messages are never expected to be hit. For example, there
may be a rare hardware bug which one wants to detect if it was to ever
happen again, but its recurrence is not likely or anticipated. This
precludes using something like checking whether the printk in question
was printed somewhere fleetwide recently to determine whether the
message in question is still present or not, since we don't anticipate
that it should be printed anywhere, but still need to monitor for its
future presence in the long-term.
This class of issue has happened on a number of occasions, causing
unhealthy machines with hardware issues to remain in production for
longer than ideal. As a recent example, some monitoring around
blk_update_request fell out of date and caused semi-broken machines to
remain in production for longer than would be desirable.
Searching through the codebase to find the message is also extremely
fragile, because many of the messages are further constructed beyond
their callsite (eg. btrfs_printk and other module-specific wrappers,
each with their own functionality). Even if they aren't, guessing the
format and formulation of the underlying message based on the aesthetics
of the message emitted is not a recipe for success at scale, and our
previous issues with fleetwide machine health checking demonstrate as
much.
This provides a solution to the issue of silently changed or deleted
printks: we record pointers to all printk format strings known at
compile time into a new .printk_index section, both in vmlinux and
modules. At runtime, this can then be iterated by looking at
<debugfs>/printk/index/<module>, which emits the following format, both
readable by humans and able to be parsed by machines:
$ head -1 vmlinux; shuf -n 5 vmlinux
# <level[,flags]> filename:line function "format"
<5> block/blk-settings.c:661 disk_stack_limits "%s: Warning: Device %s is misaligned\n"
<4> kernel/trace/trace.c:8296 trace_create_file "Could not create tracefs '%s' entry\n"
<6> arch/x86/kernel/hpet.c:144 _hpet_print_config "hpet: %s(%d):\n"
<6> init/do_mounts.c:605 prepare_namespace "Waiting for root device %s...\n"
<6> drivers/acpi/osl.c:1410 acpi_no_auto_serialize_setup "ACPI: auto-serialization disabled\n"
This mitigates the majority of cases where we have a highly-specific
printk which we want to match on, as we can now enumerate and check
whether the format changed or the printk callsite disappeared entirely
in userspace. This allows us to catch changes to printks we monitor
earlier and decide what to do about it before it becomes problematic.
There is no additional runtime cost for printk callers or printk itself,
and the assembly generated is exactly the same.
Signed-off-by: Chris Down <chris@chrisdown.name>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Jessica Yu <jeyu@kernel.org>
Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: John Ogness <john.ogness@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Kees Cook <keescook@chromium.org>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Tested-by: Petr Mladek <pmladek@suse.com>
Reported-by: kernel test robot <lkp@intel.com>
Acked-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Jessica Yu <jeyu@kernel.org> # for module.{c,h}
Signed-off-by: Petr Mladek <pmladek@suse.com>
Link: https://lore.kernel.org/r/e42070983637ac5e384f17fbdbe86d19c7b212a5.1623775748.git.chris@chrisdown.name
Signed-off-by: Mel Gorman <mgorman@suse.de>
---
MAINTAINERS | 5 ++
arch/arm/kernel/entry-v7m.S | 2 +-
arch/arm/lib/backtrace-clang.S | 2 +-
arch/arm/lib/backtrace.S | 2 +-
arch/arm/mach-rpc/io-acorn.S | 2 +-
arch/arm/vfp/vfphw.S | 6 +--
arch/ia64/include/uapi/asm/cmpxchg.h | 4 +-
arch/openrisc/kernel/entry.S | 6 +--
arch/powerpc/kernel/head_fsl_booke.S | 2 +-
arch/um/include/shared/user.h | 3 +-
arch/x86/kernel/head_32.S | 2 +-
include/asm-generic/vmlinux.lds.h | 13 +++++
include/linux/module.h | 5 ++
include/linux/printk.h | 95 ++++++++++++++++++++++++++++++++++--
init/Kconfig | 14 ++++++
kernel/module.c | 5 ++
kernel/printk/Makefile | 1 +
kernel/printk/printk.c | 13 +++--
18 files changed, 158 insertions(+), 24 deletions(-)
diff --git a/MAINTAINERS b/MAINTAINERS
index 0ea6e299439d..80be7cf64e4e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -14963,6 +14963,11 @@ S: Maintained
F: include/linux/printk.h
F: kernel/printk/
+PRINTK INDEXING
+R: Chris Down <chris@chrisdown.name>
+S: Maintained
+F: kernel/printk/index.c
+
PRISM54 WIRELESS DRIVER
M: Luis Chamberlain <mcgrof@kernel.org>
L: linux-wireless@vger.kernel.org
diff --git a/arch/arm/kernel/entry-v7m.S b/arch/arm/kernel/entry-v7m.S
index d0e898608d30..7bde93c10962 100644
--- a/arch/arm/kernel/entry-v7m.S
+++ b/arch/arm/kernel/entry-v7m.S
@@ -23,7 +23,7 @@
adr r0, strerr
mrs r1, ipsr
mov r2, lr
- bl printk
+ bl _printk
#endif
mov r0, sp
bl show_regs
diff --git a/arch/arm/lib/backtrace-clang.S b/arch/arm/lib/backtrace-clang.S
index 6174c45f53a5..5b2cdb1003e3 100644
--- a/arch/arm/lib/backtrace-clang.S
+++ b/arch/arm/lib/backtrace-clang.S
@@ -202,7 +202,7 @@ for_each_frame: tst frame, mask @ Check for address exceptions
1006: adr r0, .Lbad
mov r1, loglvl
mov r2, frame
- bl printk
+ bl _printk
no_frame: ldmfd sp!, {r4 - r9, fp, pc}
ENDPROC(c_backtrace)
.pushsection __ex_table,"a"
diff --git a/arch/arm/lib/backtrace.S b/arch/arm/lib/backtrace.S
index 872f658638d9..e8408f22d4dc 100644
--- a/arch/arm/lib/backtrace.S
+++ b/arch/arm/lib/backtrace.S
@@ -103,7 +103,7 @@ for_each_frame: tst frame, mask @ Check for address exceptions
1006: adr r0, .Lbad
mov r1, loglvl
mov r2, frame
- bl printk
+ bl _printk
no_frame: ldmfd sp!, {r4 - r9, pc}
ENDPROC(c_backtrace)
diff --git a/arch/arm/mach-rpc/io-acorn.S b/arch/arm/mach-rpc/io-acorn.S
index b9082a2a2a01..aa9bf0d771c0 100644
--- a/arch/arm/mach-rpc/io-acorn.S
+++ b/arch/arm/mach-rpc/io-acorn.S
@@ -25,4 +25,4 @@ ENTRY(insl)
ENTRY(outsl)
adr r0, .Liosl_warning
mov r1, lr
- b printk
+ b _printk
diff --git a/arch/arm/vfp/vfphw.S b/arch/arm/vfp/vfphw.S
index d5837bf05a9a..6f7926c9c179 100644
--- a/arch/arm/vfp/vfphw.S
+++ b/arch/arm/vfp/vfphw.S
@@ -23,7 +23,7 @@
#ifdef DEBUG
stmfd sp!, {r0-r3, ip, lr}
ldr r0, =1f
- bl printk
+ bl _printk
ldmfd sp!, {r0-r3, ip, lr}
.pushsection .rodata, "a"
@@ -38,7 +38,7 @@
stmfd sp!, {r0-r3, ip, lr}
mov r1, \arg
ldr r0, =1f
- bl printk
+ bl _printk
ldmfd sp!, {r0-r3, ip, lr}
.pushsection .rodata, "a"
@@ -55,7 +55,7 @@
mov r2, \arg2
mov r1, \arg1
ldr r0, =1f
- bl printk
+ bl _printk
ldmfd sp!, {r0-r3, ip, lr}
.pushsection .rodata, "a"
diff --git a/arch/ia64/include/uapi/asm/cmpxchg.h b/arch/ia64/include/uapi/asm/cmpxchg.h
index 926c6cb1e029..2c2f3cfeaa77 100644
--- a/arch/ia64/include/uapi/asm/cmpxchg.h
+++ b/arch/ia64/include/uapi/asm/cmpxchg.h
@@ -143,9 +143,9 @@ extern long ia64_cmpxchg_called_with_bad_pointer(void);
do { \
if (_cmpxchg_bugcheck_count-- <= 0) { \
void *ip; \
- extern int printk(const char *fmt, ...); \
+ extern int _printk(const char *fmt, ...); \
ip = (void *) ia64_getreg(_IA64_REG_IP); \
- printk("CMPXCHG_BUGCHECK: stuck at %p on word %p\n", ip, (v));\
+ _printk("CMPXCHG_BUGCHECK: stuck at %p on word %p\n", ip, (v));\
break; \
} \
} while (0)
diff --git a/arch/openrisc/kernel/entry.S b/arch/openrisc/kernel/entry.S
index 98e4f97db515..4edabac39773 100644
--- a/arch/openrisc/kernel/entry.S
+++ b/arch/openrisc/kernel/entry.S
@@ -552,7 +552,7 @@ EXCEPTION_ENTRY(_external_irq_handler)
l.movhi r3,hi(42f)
l.ori r3,r3,lo(42f)
l.sw 0x0(r1),r3
- l.jal printk
+ l.jal _printk
l.sw 0x4(r1),r4
l.addi r1,r1,0x8
@@ -683,8 +683,8 @@ ENTRY(_sys_call_handler)
l.sw -4(r1),r27
l.sw -8(r1),r11
l.addi r1,r1,-8
- l.movhi r27,hi(printk)
- l.ori r27,r27,lo(printk)
+ l.movhi r27,hi(_printk)
+ l.ori r27,r27,lo(_printk)
l.jalr r27
l.nop
l.addi r1,r1,8
diff --git a/arch/powerpc/kernel/head_fsl_booke.S b/arch/powerpc/kernel/head_fsl_booke.S
index 0f9642f36b49..9a2f4265e6d2 100644
--- a/arch/powerpc/kernel/head_fsl_booke.S
+++ b/arch/powerpc/kernel/head_fsl_booke.S
@@ -858,7 +858,7 @@ _GLOBAL(load_up_spe)
ori r3,r3,87f@l
mr r4,r2 /* current */
lwz r5,_NIP(r1)
- bl printk
+ bl _printk
#endif
b interrupt_return
#ifdef CONFIG_PRINTK
diff --git a/arch/um/include/shared/user.h b/arch/um/include/shared/user.h
index e793e4212f0a..dd4badffdeb3 100644
--- a/arch/um/include/shared/user.h
+++ b/arch/um/include/shared/user.h
@@ -38,7 +38,8 @@ extern void panic(const char *fmt, ...)
#define UM_KERN_CONT KERN_CONT
#ifdef UML_CONFIG_PRINTK
-extern int printk(const char *fmt, ...)
+#define printk(...) _printk(__VA_ARGS__)
+extern int _printk(const char *fmt, ...)
__attribute__ ((format (printf, 1, 2)));
#else
static inline int printk(const char *fmt, ...)
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index 67f590425d90..d8c64dab0efe 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -432,7 +432,7 @@ SYM_FUNC_START(early_ignore_irq)
pushl 32(%esp)
pushl 40(%esp)
pushl $int_msg
- call printk
+ call _printk
call dump_stack
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index 62669b36a772..aa50bf2959fe 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -483,6 +483,8 @@
\
TRACEDATA \
\
+ PRINTK_INDEX \
+ \
/* Kernel symbol table: Normal symbols */ \
__ksymtab : AT(ADDR(__ksymtab) - LOAD_OFFSET) { \
__start___ksymtab = .; \
@@ -894,6 +896,17 @@
#define TRACEDATA
#endif
+#ifdef CONFIG_PRINTK_INDEX
+#define PRINTK_INDEX \
+ .printk_index : AT(ADDR(.printk_index) - LOAD_OFFSET) { \
+ __start_printk_index = .; \
+ *(.printk_index) \
+ __stop_printk_index = .; \
+ }
+#else
+#define PRINTK_INDEX
+#endif
+
#define NOTES \
.notes : AT(ADDR(.notes) - LOAD_OFFSET) { \
__start_notes = .; \
diff --git a/include/linux/module.h b/include/linux/module.h
index 3d80a9d908f5..29e6ae076e4e 100644
--- a/include/linux/module.h
+++ b/include/linux/module.h
@@ -511,6 +511,11 @@ struct module {
struct klp_modinfo *klp_info;
#endif
+#ifdef CONFIG_PRINTK_INDEX
+ unsigned int printk_index_size;
+ struct pi_entry **printk_index_start;
+#endif
+
#ifdef CONFIG_MODULE_UNLOAD
/* What modules depend on me? */
struct list_head source_list;
diff --git a/include/linux/printk.h b/include/linux/printk.h
index b394fb65646c..0cc0d186439e 100644
--- a/include/linux/printk.h
+++ b/include/linux/printk.h
@@ -162,12 +162,12 @@ asmlinkage __printf(1, 0)
int vprintk(const char *fmt, va_list args);
asmlinkage __printf(1, 2) __cold
-int printk(const char *fmt, ...);
+int _printk(const char *fmt, ...);
/*
* Special printk facility for scheduler/timekeeping use only, _DO_NOT_USE_ !
*/
-__printf(1, 2) __cold int printk_deferred(const char *fmt, ...);
+__printf(1, 2) __cold int _printk_deferred(const char *fmt, ...);
extern void __printk_safe_enter(void);
extern void __printk_safe_exit(void);
@@ -217,12 +217,12 @@ int vprintk(const char *s, va_list args)
return 0;
}
static inline __printf(1, 2) __cold
-int printk(const char *s, ...)
+int _printk(const char *s, ...)
{
return 0;
}
static inline __printf(1, 2) __cold
-int printk_deferred(const char *s, ...)
+int _printk_deferred(const char *s, ...)
{
return 0;
}
@@ -355,6 +355,93 @@ extern int kptr_restrict;
#define pr_fmt(fmt) fmt
#endif
+struct module;
+
+#ifdef CONFIG_PRINTK_INDEX
+struct pi_entry {
+ const char *fmt;
+ const char *func;
+ const char *file;
+ unsigned int line;
+
+ /*
+ * While printk and pr_* have the level stored in the string at compile
+ * time, some subsystems dynamically add it at runtime through the
+ * format string. For these dynamic cases, we allow the subsystem to
+ * tell us the level at compile time.
+ *
+ * NULL indicates that the level, if any, is stored in fmt.
+ */
+ const char *level;
+
+ /*
+ * The format string used by various subsystem specific printk()
+ * wrappers to prefix the message.
+ *
+ * Note that the static prefix defined by the pr_fmt() macro is stored
+ * directly in the message format (@fmt), not here.
+ */
+ const char *subsys_fmt_prefix;
+} __packed;
+
+#define __printk_index_emit(_fmt, _level, _subsys_fmt_prefix) \
+ do { \
+ if (__builtin_constant_p(_fmt) && __builtin_constant_p(_level)) { \
+ /*
+ * We check __builtin_constant_p multiple times here
+ * for the same input because GCC will produce an error
+ * if we try to assign a static variable to fmt if it
+ * is not a constant, even with the outer if statement.
+ */ \
+ static const struct pi_entry _entry \
+ __used = { \
+ .fmt = __builtin_constant_p(_fmt) ? (_fmt) : NULL, \
+ .func = __func__, \
+ .file = __FILE__, \
+ .line = __LINE__, \
+ .level = __builtin_constant_p(_level) ? (_level) : NULL, \
+ .subsys_fmt_prefix = _subsys_fmt_prefix,\
+ }; \
+ static const struct pi_entry *_entry_ptr \
+ __used __section(".printk_index") = &_entry; \
+ } \
+ } while (0)
+
+#else /* !CONFIG_PRINTK_INDEX */
+#define __printk_index_emit(...) do {} while (0)
+#endif /* CONFIG_PRINTK_INDEX */
+
+/*
+ * Some subsystems have their own custom printk that applies a va_format to a
+ * generic format, for example, to include a device number or other metadata
+ * alongside the format supplied by the caller.
+ *
+ * In order to store these in the way they would be emitted by the printk
+ * infrastructure, the subsystem provides us with the start, fixed string, and
+ * any subsequent text in the format string.
+ *
+ * We take a variable argument list as pr_fmt/dev_fmt/etc are sometimes passed
+ * as multiple arguments (eg: `"%s: ", "blah"`), and we must only take the
+ * first one.
+ *
+ * subsys_fmt_prefix must be known at compile time, or compilation will fail
+ * (since this is a mistake). If fmt or level is not known at compile time, no
+ * index entry will be made (since this can legitimately happen).
+ */
+#define printk_index_subsys_emit(subsys_fmt_prefix, level, fmt, ...) \
+ __printk_index_emit(fmt, level, subsys_fmt_prefix)
+
+#define printk_index_wrap(_p_func, _fmt, ...) \
+ ({ \
+ __printk_index_emit(_fmt, NULL, NULL); \
+ _p_func(_fmt, ##__VA_ARGS__); \
+ })
+
+
+#define printk(fmt, ...) printk_index_wrap(_printk, fmt, ##__VA_ARGS__)
+#define printk_deferred(fmt, ...) \
+ printk_index_wrap(_printk_deferred, fmt, ##__VA_ARGS__)
+
/**
* pr_emerg - Print an emergency-level message
* @fmt: format string
diff --git a/init/Kconfig b/init/Kconfig
index bb63bf54f1e3..f7f6e8951458 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -775,6 +775,20 @@ config PRINTK_SAFE_LOG_BUF_SHIFT
13 => 8 KB for each CPU
12 => 4 KB for each CPU
+config PRINTK_INDEX
+ bool "Printk indexing debugfs interface"
+ depends on PRINTK && DEBUG_FS
+ help
+ Add support for indexing of all printk formats known at compile time
+ at <debugfs>/printk/index/<module>.
+
+ This can be used as part of maintaining daemons which monitor
+ /dev/kmsg, as it permits auditing the printk formats present in a
+ kernel, allowing detection of cases where monitored printks are
+ changed or no longer present.
+
+ There is no additional runtime cost to printk with this enabled.
+
#
# Architectures with an unreliable sched_clock() should select this:
#
diff --git a/kernel/module.c b/kernel/module.c
index c97f65a37756..0edabd47584e 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -3430,6 +3430,11 @@ static int find_module_sections(struct module *mod, struct load_info *info)
sizeof(unsigned long),
&mod->num_kprobe_blacklist);
#endif
+#ifdef CONFIG_PRINTK_INDEX
+ mod->printk_index_start = section_objs(info, ".printk_index",
+ sizeof(*mod->printk_index_start),
+ &mod->printk_index_size);
+#endif
#ifdef CONFIG_HAVE_STATIC_CALL_INLINE
mod->static_call_sites = section_objs(info, ".static_call_sites",
sizeof(*mod->static_call_sites),
diff --git a/kernel/printk/Makefile b/kernel/printk/Makefile
index eee3dc9b60a9..d118739874c0 100644
--- a/kernel/printk/Makefile
+++ b/kernel/printk/Makefile
@@ -3,3 +3,4 @@ obj-y = printk.o
obj-$(CONFIG_PRINTK) += printk_safe.o
obj-$(CONFIG_A11Y_BRAILLE_CONSOLE) += braille.o
obj-$(CONFIG_PRINTK) += printk_ringbuffer.o
+obj-$(CONFIG_PRINTK_INDEX) += index.o
diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index 8d72854f031a..919cb99e908e 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -2272,10 +2272,13 @@ int vprintk_default(const char *fmt, va_list args)
EXPORT_SYMBOL_GPL(vprintk_default);
/**
- * printk - print a kernel message
+ * _printk - print a kernel message
* @fmt: format string
*
- * This is printk(). It can be called from any context. We want it to work.
+ * This is _printk(). It can be called from any context. We want it to work.
+ *
+ * If printk indexing is enabled, _printk() is called from printk_index_wrap.
+ * Otherwise, printk is simply #defined to _printk.
*
* We try to grab the console_lock. If we succeed, it's easy - we log the
* output and call the console drivers. If we fail to get the semaphore, we
@@ -2292,7 +2295,7 @@ EXPORT_SYMBOL_GPL(vprintk_default);
*
* See the vsnprintf() documentation for format string extensions over C99.
*/
-asmlinkage __visible int printk(const char *fmt, ...)
+asmlinkage __visible int _printk(const char *fmt, ...)
{
va_list args;
int r;
@@ -2303,7 +2306,7 @@ asmlinkage __visible int printk(const char *fmt, ...)
return r;
}
-EXPORT_SYMBOL(printk);
+EXPORT_SYMBOL(_printk);
#else /* CONFIG_PRINTK */
@@ -3291,7 +3294,7 @@ int vprintk_deferred(const char *fmt, va_list args)
return r;
}
-int printk_deferred(const char *fmt, ...)
+int _printk_deferred(const char *fmt, ...)
{
va_list args;
int r;