kasan: mark kasan_check_(read|write) as 'notrace'
Subsystems (2)
PATCH 1/1 kasan: mark kasan_check_(read|write) as 'notrace'
When option CONFIG_KASAN is enabled toghether with ftrace, function ftrace_graph_caller() gets in to a recursion, via functions kasan_check_read() and kasan_check_write().
Breakpoint 2, ftrace_graph_caller () at ../arch/arm64/kernel/entry-ftrace.S:179
179 mcount_get_pc x0 // function's pc
(gdb) bt
#0 ftrace_graph_caller () at ../arch/arm64/kernel/entry-ftrace.S:179 #1 0xffffff90101406c8 in ftrace_caller () at ../arch/arm64/kernel/entry-ftrace.S:151 #2 0xffffff90106fd084 in kasan_check_write (p=0xffffffc06c170878, size=4) at ../mm/kasan/common.c:105 #3 0xffffff90104a2464 in atomic_add_return (v=<optimized out>, i=<optimized out>) at ./include/generated/atomic-instrumented.h:71 #4 atomic_inc_return (v=<optimized out>) at ./include/generated/atomic-fallback.h:284 #5 trace_graph_entry (trace=0xffffffc03f5ff380) at ../kernel/trace/trace_functions_graph.c:441 #6 0xffffff9010481774 in trace_graph_entry_watchdog (trace=<optimized out>) at ../kernel/trace/trace_selftest.c:741 #7 0xffffff90104a185c in function_graph_enter (ret=<optimized out>, func=<optimized out>, frame_pointer=18446743799894897728, retp=<optimized out>) at ../kernel/trace/trace_functions_graph.c:196 #8 0xffffff9010140628 in prepare_ftrace_return (self_addr=18446743592948977792, parent=0xffffffc03f5ff418, frame_pointer=18446743799894897728) at ../arch/arm64/kernel/ftrace.c:231 #9 0xffffff90101406f4 in ftrace_graph_caller () at ../arch/arm64/kernel/entry-ftrace.S:182
Backtrace stopped: previous frame identical to this frame (corrupt stack?) (gdb)
Rework so that kasan_check_read() and kasan_check_write() is marked with 'notrace'.
--- mm/kasan/common.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
mm/kasan/common.c +2 -2
@@ -95,13 +95,13 @@ void kasan_disable_current(void) current->kasan_depth--; } -void kasan_check_read (const volatile void *p,unsigned int size) +void notrace kasan_check_read(const volatile void *p, unsigned int size) {check_memory_region ((unsigned long )p, size, false, _RET_IP_); }EXPORT_SYMBOL (kasan_check_read); -void kasan_check_write (const volatile void *p,unsigned int size) +void notrace kasan_check_write(const volatile void *p, unsigned int size) {check_memory_region ((unsigned long )p, size, true, _RET_IP_); }
Signature · 1 line
2.19.2
Dmitry Vyukov
reviewerQuoted 39 lines · Anders Roxell
When option
CONFIG_KASANis enabled toghether with ftrace, functionftrace_graph_caller()gets in to a recursion, via functionskasan_check_read()andkasan_check_write().Breakpoint 2,
ftrace_graph_caller() at../arch/arm64/kernel/entry-ftrace.S:179
179mcount_get_pcx0 // function's pc
(gdb) bt#0 ftrace_graph_caller () at ../arch/arm64/kernel/entry-ftrace.S:179 #1 0xffffff90101406c8 in ftrace_caller () at ../arch/arm64/kernel/entry-ftrace.S:151 #2 0xffffff90106fd084 in kasan_check_write (p=0xffffffc06c170878, size=4) at ../mm/kasan/common.c:105 #3 0xffffff90104a2464 in atomic_add_return (v=<optimized out>, i=<optimized out>) at ./include/generated/atomic-instrumented.h:71 #4 atomic_inc_return (v=<optimized out>) at ./include/generated/atomic-fallback.h:284 #5 trace_graph_entry (trace=0xffffffc03f5ff380) at ../kernel/trace/trace_functions_graph.c:441 #6 0xffffff9010481774 in trace_graph_entry_watchdog (trace=<optimized out>) at ../kernel/trace/trace_selftest.c:741 #7 0xffffff90104a185c in function_graph_enter (ret=<optimized out>, func=<optimized out>, frame_pointer=18446743799894897728, retp=<optimized out>) at ../kernel/trace/trace_functions_graph.c:196 #8 0xffffff9010140628 in prepare_ftrace_return (self_addr=18446743592948977792, parent=0xffffffc03f5ff418, frame_pointer=18446743799894897728) at ../arch/arm64/kernel/ftrace.c:231 #9 0xffffff90101406f4 in ftrace_graph_caller () at ../arch/arm64/kernel/entry-ftrace.S:182Backtrace stopped: previous frame identical to this frame (corrupt stack?) (gdb)
Rework so that
kasan_check_read()andkasan_check_write()is marked with 'notrace'.Signed-off-by Anders Roxell <[email protected]>--- mm/kasan/common.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)mm/kasan/common.c +1
-1@@ -95,13 +95,13 @@ void kasan_disable_current(void) current->kasan_depth--; } -void kasan_check_read (const volatile void *p,unsigned int size) +void notrace kasan_check_read(const volatile void *p, unsigned int size)
mm/kasan/common.c void kasan_disable_current(void) · line 107
Quoted 10 lines
mm/kasan/common.c +1
-1{check_memory_region ((unsigned long )p, size, false, _RET_IP_); }EXPORT_SYMBOL (kasan_check_read); -void kasan_check_write (const volatile void *p,unsigned int size) +void notrace kasan_check_write(const volatile void *p, unsigned int size) {check_memory_region ((unsigned long )p, size, true, _RET_IP_); }
Hi Anders,
Thanks for fixing this!
I wonder if there is some compiler/make flag to turn this off for the whole file?
We turn as much instrumentation as possible already for this file in Makefile:
KASAN_SANITIZE := n
UBSAN_SANITIZE_kasan.o := nKCOV_INSTRUMENT := n
CFLAGS_REMOVE_kasan.o = -pg
CFLAGS_kasan.o := $(call cc-option, -fno-conserve-stack -fno-stack-protector)
These functions call check_memory_region, which is presumably inlined. But if it's not inlined later in some configuration, or we just shuffle code a bit, we can get the same problem again.
Quoted 67 lines · Dmitry Vyukov
Quoted 39 lines · Anders Roxell
When option
CONFIG_KASANis enabled toghether with ftrace, functionftrace_graph_caller()gets in to a recursion, via functionskasan_check_read()andkasan_check_write().Breakpoint 2,
ftrace_graph_caller() at../arch/arm64/kernel/entry-ftrace.S:179
179mcount_get_pcx0 // function's pc
(gdb) bt#0 ftrace_graph_caller () at ../arch/arm64/kernel/entry-ftrace.S:179 #1 0xffffff90101406c8 in ftrace_caller () at ../arch/arm64/kernel/entry-ftrace.S:151 #2 0xffffff90106fd084 in kasan_check_write (p=0xffffffc06c170878, size=4) at ../mm/kasan/common.c:105 #3 0xffffff90104a2464 in atomic_add_return (v=<optimized out>, i=<optimized out>) at ./include/generated/atomic-instrumented.h:71 #4 atomic_inc_return (v=<optimized out>) at ./include/generated/atomic-fallback.h:284 #5 trace_graph_entry (trace=0xffffffc03f5ff380) at ../kernel/trace/trace_functions_graph.c:441 #6 0xffffff9010481774 in trace_graph_entry_watchdog (trace=<optimized out>) at ../kernel/trace/trace_selftest.c:741 #7 0xffffff90104a185c in function_graph_enter (ret=<optimized out>, func=<optimized out>, frame_pointer=18446743799894897728, retp=<optimized out>) at ../kernel/trace/trace_functions_graph.c:196 #8 0xffffff9010140628 in prepare_ftrace_return (self_addr=18446743592948977792, parent=0xffffffc03f5ff418, frame_pointer=18446743799894897728) at ../arch/arm64/kernel/ftrace.c:231 #9 0xffffff90101406f4 in ftrace_graph_caller () at ../arch/arm64/kernel/entry-ftrace.S:182Backtrace stopped: previous frame identical to this frame (corrupt stack?) (gdb)
Rework so that
kasan_check_read()andkasan_check_write()is marked with 'notrace'.Signed-off-by Anders Roxell <[email protected]>--- mm/kasan/common.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)mm/kasan/common.c +1
-1@@ -95,13 +95,13 @@ void kasan_disable_current(void) current->kasan_depth--; } -void kasan_check_read (const volatile void *p,unsigned int size) +void notrace kasan_check_read(const volatile void *p, unsigned int size)mm/kasan/common.c void kasan_disable_current(void) · line 107
Quoted 10 lines
mm/kasan/common.c +1
-1{check_memory_region ((unsigned long )p, size, false, _RET_IP_); }EXPORT_SYMBOL (kasan_check_read); -void kasan_check_write (const volatile void *p,unsigned int size) +void notrace kasan_check_write(const volatile void *p, unsigned int size) {check_memory_region ((unsigned long )p, size, true, _RET_IP_); }Hi Anders,
Thanks for fixing this!
I wonder if there is some compiler/make flag to turn this off for the whole file?
We turn as much instrumentation as possible already for this file in
Makefile:
KASAN_SANITIZE:= n
UBSAN_SANITIZE_kasan.o := nKCOV_INSTRUMENT:= n
CFLAGS_REMOVE_kasan.o = -pg
CFLAGS_kasan.o := $(call cc-option, -fno-conserve-stack -fno-stack-protector)
Quoted 3 lines
These functions call
check_memory_region, which is presumably inlined. But if it's not inlined later in some configuration, or we just shuffle code a bit, we can get the same problem again.
Thank you for the review.
I'll send out a v2 shortly where I turned off the hole file.
Cheers,
Anders