* [PATCH 1/2] ubsan: allow to disable the null sanitizer
2016-09-12 12:37 [PATCH 0/2] ubsan for s390 Christian Borntraeger
@ 2016-09-12 12:37 ` Christian Borntraeger
2016-09-13 6:46 ` Heiko Carstens
2016-09-12 12:37 ` [PATCH 2/2] s390/ubsan: enable UBSAN for s390 Christian Borntraeger
2016-09-12 16:38 ` Andrey Ryabinin
2 siblings, 1 reply; 6+ messages in thread
From: Christian Borntraeger @ 2016-09-12 12:37 UTC (permalink / raw)
To: linux-s390, linux-kernel, linux-kbuild
Cc: Heiko Carstens, Martin Schwidefsky, Michal Marek,
Andrey Ryabinin, Yang Shi, Andrew Morton, Christian Borntraeger
Some architectures use a hardware defined struct at address zero.
Checking for a null pointer will result in many ubsan reports.
Let users disable the null sanitizer.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
lib/Kconfig.ubsan | 11 +++++++++++
scripts/Makefile.ubsan | 5 ++++-
2 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/lib/Kconfig.ubsan b/lib/Kconfig.ubsan
index 39494af..851c3f2 100644
--- a/lib/Kconfig.ubsan
+++ b/lib/Kconfig.ubsan
@@ -1,6 +1,9 @@
config ARCH_HAS_UBSAN_SANITIZE_ALL
bool
+config ARCH_WANTS_UBSAN_NO_NULL
+ def_bool n
+
config UBSAN
bool "Undefined behaviour sanity checker"
help
@@ -34,3 +37,11 @@ config UBSAN_ALIGNMENT
This option enables detection of unaligned memory accesses.
Enabling this option on architectures that support unaligned
accesses may produce a lot of false positives.
+
+config UBSAN_NULL
+ bool "Enable checking of null pointers"
+ depends on UBSAN
+ default y if UBSAN && !ARCH_WANTS_UBSAN_NO_NULL
+ help
+ This option enables detection of memory accesses via a
+ null pointer.
diff --git a/scripts/Makefile.ubsan b/scripts/Makefile.ubsan
index 77ce538..3b1b138 100644
--- a/scripts/Makefile.ubsan
+++ b/scripts/Makefile.ubsan
@@ -3,7 +3,6 @@ ifdef CONFIG_UBSAN
CFLAGS_UBSAN += $(call cc-option, -fsanitize=integer-divide-by-zero)
CFLAGS_UBSAN += $(call cc-option, -fsanitize=unreachable)
CFLAGS_UBSAN += $(call cc-option, -fsanitize=vla-bound)
- CFLAGS_UBSAN += $(call cc-option, -fsanitize=null)
CFLAGS_UBSAN += $(call cc-option, -fsanitize=signed-integer-overflow)
CFLAGS_UBSAN += $(call cc-option, -fsanitize=bounds)
CFLAGS_UBSAN += $(call cc-option, -fsanitize=object-size)
@@ -15,6 +14,10 @@ ifdef CONFIG_UBSAN_ALIGNMENT
CFLAGS_UBSAN += $(call cc-option, -fsanitize=alignment)
endif
+ifdef CONFIG_UBSAN_NULL
+ CFLAGS_UBSAN += $(call cc-option, -fsanitize=null)
+endif
+
# -fsanitize=* options makes GCC less smart than usual and
# increase number of 'maybe-uninitialized false-positives
CFLAGS_UBSAN += $(call cc-option, -Wno-maybe-uninitialized)
--
2.5.5
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH 1/2] ubsan: allow to disable the null sanitizer
2016-09-12 12:37 ` [PATCH 1/2] ubsan: allow to disable the null sanitizer Christian Borntraeger
@ 2016-09-13 6:46 ` Heiko Carstens
0 siblings, 0 replies; 6+ messages in thread
From: Heiko Carstens @ 2016-09-13 6:46 UTC (permalink / raw)
To: Christian Borntraeger
Cc: linux-s390, linux-kernel, linux-kbuild, Martin Schwidefsky,
Michal Marek, Andrey Ryabinin, Yang Shi, Andrew Morton
On Mon, Sep 12, 2016 at 02:37:19PM +0200, Christian Borntraeger wrote:
> Some architectures use a hardware defined struct at address zero.
> Checking for a null pointer will result in many ubsan reports.
> Let users disable the null sanitizer.
>
> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
> ---
> lib/Kconfig.ubsan | 11 +++++++++++
> scripts/Makefile.ubsan | 5 ++++-
> 2 files changed, 15 insertions(+), 1 deletion(-)
>
> diff --git a/lib/Kconfig.ubsan b/lib/Kconfig.ubsan
> index 39494af..851c3f2 100644
> --- a/lib/Kconfig.ubsan
> +++ b/lib/Kconfig.ubsan
> @@ -1,6 +1,9 @@
> config ARCH_HAS_UBSAN_SANITIZE_ALL
> bool
>
> +config ARCH_WANTS_UBSAN_NO_NULL
> + def_bool n
> +
> config UBSAN
> bool "Undefined behaviour sanity checker"
> help
> @@ -34,3 +37,11 @@ config UBSAN_ALIGNMENT
> This option enables detection of unaligned memory accesses.
> Enabling this option on architectures that support unaligned
> accesses may produce a lot of false positives.
> +
> +config UBSAN_NULL
> + bool "Enable checking of null pointers"
> + depends on UBSAN
> + default y if UBSAN && !ARCH_WANTS_UBSAN_NO_NULL
I removed the not needed UBSAN from the "default y" and applied both
patches to the s390 tree.
Thanks!
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 2/2] s390/ubsan: enable UBSAN for s390
2016-09-12 12:37 [PATCH 0/2] ubsan for s390 Christian Borntraeger
2016-09-12 12:37 ` [PATCH 1/2] ubsan: allow to disable the null sanitizer Christian Borntraeger
@ 2016-09-12 12:37 ` Christian Borntraeger
2016-09-12 16:38 ` Andrey Ryabinin
2 siblings, 0 replies; 6+ messages in thread
From: Christian Borntraeger @ 2016-09-12 12:37 UTC (permalink / raw)
To: linux-s390, linux-kernel, linux-kbuild
Cc: Heiko Carstens, Martin Schwidefsky, Michal Marek,
Andrey Ryabinin, Yang Shi, Andrew Morton, Christian Borntraeger
This enables UBSAN for s390. We have to disable the null sanitizer
as s390 code does access memory via a null pointer (the prefix page).
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
arch/s390/Kconfig | 2 ++
arch/s390/boot/compressed/Makefile | 1 +
arch/s390/kernel/Makefile | 3 +++
arch/s390/kernel/vdso32/Makefile | 3 ++-
arch/s390/kernel/vdso64/Makefile | 3 ++-
5 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index e49ca114..5862a93 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -74,6 +74,7 @@ config S390
select ARCH_HAS_GCOV_PROFILE_ALL
select ARCH_HAS_KCOV
select ARCH_HAS_SG_CHAIN
+ select ARCH_HAS_UBSAN_SANITIZE_ALL
select ARCH_HAVE_NMI_SAFE_CMPXCHG
select ARCH_INLINE_READ_LOCK
select ARCH_INLINE_READ_LOCK_BH
@@ -110,6 +111,7 @@ config S390
select ARCH_USE_CMPXCHG_LOCKREF
select ARCH_WANTS_DYNAMIC_TASK_STRUCT
select ARCH_WANTS_PROT_NUMA_PROT_NONE
+ select ARCH_WANTS_UBSAN_NO_NULL
select ARCH_WANT_IPC_PARSE_VERSION
select BUILDTIME_EXTABLE_SORT
select CLONE_BACKWARDS2
diff --git a/arch/s390/boot/compressed/Makefile b/arch/s390/boot/compressed/Makefile
index 13723c3..66f42bb 100644
--- a/arch/s390/boot/compressed/Makefile
+++ b/arch/s390/boot/compressed/Makefile
@@ -17,6 +17,7 @@ KBUILD_CFLAGS += $(call cc-option,-mpacked-stack)
KBUILD_CFLAGS += $(call cc-option,-ffreestanding)
GCOV_PROFILE := n
+UBSAN_SANITIZE := n
OBJECTS := $(addprefix $(objtree)/arch/s390/kernel/, head.o sclp.o ebcdic.o als.o)
OBJECTS += $(obj)/head.o $(obj)/misc.o $(obj)/piggy.o
diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile
index 8bc18eb..30407e6 100644
--- a/arch/s390/kernel/Makefile
+++ b/arch/s390/kernel/Makefile
@@ -48,6 +48,9 @@ AFLAGS_head.o += -march=z900
endif
GCOV_PROFILE_sclp.o := n
GCOV_PROFILE_als.o := n
+UBSAN_SANITIZE_als.o := n
+UBSAN_SANITIZE_early.o := n
+UBSAN_SANITIZE_sclp.o := n
obj-y := traps.o time.o process.o base.o early.o setup.o idle.o vtime.o
obj-y += processor.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o nmi.o
diff --git a/arch/s390/kernel/vdso32/Makefile b/arch/s390/kernel/vdso32/Makefile
index 6814545..6cc9478 100644
--- a/arch/s390/kernel/vdso32/Makefile
+++ b/arch/s390/kernel/vdso32/Makefile
@@ -24,8 +24,9 @@ obj-y += vdso32_wrapper.o
extra-y += vdso32.lds
CPPFLAGS_vdso32.lds += -P -C -U$(ARCH)
-# Disable gcov profiling for VDSO code
+# Disable gcov profiling and ubsan for VDSO code
GCOV_PROFILE := n
+UBSAN_SANITIZE := n
# Force dependency (incbin is bad)
$(obj)/vdso32_wrapper.o : $(obj)/vdso32.so
diff --git a/arch/s390/kernel/vdso64/Makefile b/arch/s390/kernel/vdso64/Makefile
index 0b0fd22..2d54c18 100644
--- a/arch/s390/kernel/vdso64/Makefile
+++ b/arch/s390/kernel/vdso64/Makefile
@@ -24,8 +24,9 @@ obj-y += vdso64_wrapper.o
extra-y += vdso64.lds
CPPFLAGS_vdso64.lds += -P -C -U$(ARCH)
-# Disable gcov profiling for VDSO code
+# Disable gcov profiling and ubsan for VDSO code
GCOV_PROFILE := n
+UBSAN_SANITIZE := n
# Force dependency (incbin is bad)
$(obj)/vdso64_wrapper.o : $(obj)/vdso64.so
--
2.5.5
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 0/2] ubsan for s390
2016-09-12 12:37 [PATCH 0/2] ubsan for s390 Christian Borntraeger
@ 2016-09-12 16:38 ` Andrey Ryabinin
2016-09-12 12:37 ` [PATCH 2/2] s390/ubsan: enable UBSAN for s390 Christian Borntraeger
2016-09-12 16:38 ` Andrey Ryabinin
2 siblings, 0 replies; 6+ messages in thread
From: Andrey Ryabinin @ 2016-09-12 16:38 UTC (permalink / raw)
To: Christian Borntraeger, linux-s390, linux-kernel, linux-kbuild
Cc: Heiko Carstens, Martin Schwidefsky, Michal Marek, Yang Shi,
Andrew Morton
On 09/12/2016 03:37 PM, Christian Borntraeger wrote:
> This patch set enables ubsan for s390.
>
> We have to fence some files and can then enable
> ARCH_HAS_UBSAN_SANITIZE_ALL. (patch 2)
>
> We have to disable the null pointer sanitizer as
> we use the S390_lowcore macro to access several
> memory location after address 0 (patch 1)
>
> If patch 1 is ok, is the s390 tree the right place
> also for patch1?
>
Fine by me
Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
> Christian Borntraeger (2):
> ubsan: allow to disable the null sanitizer
> s390/ubsan: enable UBSAN for s390
>
> arch/s390/Kconfig | 2 ++
> arch/s390/boot/compressed/Makefile | 1 +
> arch/s390/kernel/Makefile | 3 +++
> arch/s390/kernel/vdso32/Makefile | 3 ++-
> arch/s390/kernel/vdso64/Makefile | 3 ++-
> lib/Kconfig.ubsan | 11 +++++++++++
> scripts/Makefile.ubsan | 5 ++++-
> 7 files changed, 25 insertions(+), 3 deletions(-)
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 0/2] ubsan for s390
@ 2016-09-12 16:38 ` Andrey Ryabinin
0 siblings, 0 replies; 6+ messages in thread
From: Andrey Ryabinin @ 2016-09-12 16:38 UTC (permalink / raw)
To: Christian Borntraeger, linux-s390, linux-kernel, linux-kbuild
Cc: Heiko Carstens, Martin Schwidefsky, Michal Marek, Yang Shi,
Andrew Morton
On 09/12/2016 03:37 PM, Christian Borntraeger wrote:
> This patch set enables ubsan for s390.
>
> We have to fence some files and can then enable
> ARCH_HAS_UBSAN_SANITIZE_ALL. (patch 2)
>
> We have to disable the null pointer sanitizer as
> we use the S390_lowcore macro to access several
> memory location after address 0 (patch 1)
>
> If patch 1 is ok, is the s390 tree the right place
> also for patch1?
>
Fine by me
Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
> Christian Borntraeger (2):
> ubsan: allow to disable the null sanitizer
> s390/ubsan: enable UBSAN for s390
>
> arch/s390/Kconfig | 2 ++
> arch/s390/boot/compressed/Makefile | 1 +
> arch/s390/kernel/Makefile | 3 +++
> arch/s390/kernel/vdso32/Makefile | 3 ++-
> arch/s390/kernel/vdso64/Makefile | 3 ++-
> lib/Kconfig.ubsan | 11 +++++++++++
> scripts/Makefile.ubsan | 5 ++++-
> 7 files changed, 25 insertions(+), 3 deletions(-)
>
^ permalink raw reply [flat|nested] 6+ messages in thread