The return type of ACCESS_ONCE is configuration dependent and may be either
'int' or 'long int' for the writecache_has_error() macro, so we get a warning
like this for either format string:
In file included from drivers/md/dm-writecache.c:8:
drivers/md/dm-writecache.c: In function 'writecache_status':
drivers/md/dm-writecache.c:2227:10: error: format '%ld' expects argument of type 'long int', but argument 4 has type 'int' [-Werror=format=]
DMEMIT("%ld %llu %llu %llu", writecache_has_error(wc),
^~~~~~~~~~~~~~~~~~~~
include/linux/device-mapper.h:549:46: note: in definition of macro 'DMEMIT'
0 : scnprintf(result + sz, maxlen - sz, x))
^
The code is otherwise correct, so we just need to shut up the warning,
which can be done using an extra type cast.
Fixes: bb15b431d650 ("dm: add writecache target")
Signed-off-by: Arnd Bergmann <[email protected]>
---
drivers/md/dm-writecache.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/md/dm-writecache.c b/drivers/md/dm-writecache.c
index 1ef06e738eb6..772ac3a57287 100644
--- a/drivers/md/dm-writecache.c
+++ b/drivers/md/dm-writecache.c
@@ -2224,7 +2224,7 @@ static void writecache_status(struct dm_target *ti, status_type_t type,
switch (type) {
case STATUSTYPE_INFO:
- DMEMIT("%ld %llu %llu %llu", writecache_has_error(wc),
+ DMEMIT("%ld %llu %llu %llu", (long)writecache_has_error(wc),
(unsigned long long)wc->n_blocks, (unsigned long long)wc->freelist_size,
(unsigned long long)wc->writeback_size);
break;
--
2.9.0
On Mon, May 28 2018 at 11:54am -0400,
Arnd Bergmann <[email protected]> wrote:
> The return type of ACCESS_ONCE is configuration dependent and may be either
> 'int' or 'long int' for the writecache_has_error() macro, so we get a warning
> like this for either format string:
>
> In file included from drivers/md/dm-writecache.c:8:
> drivers/md/dm-writecache.c: In function 'writecache_status':
> drivers/md/dm-writecache.c:2227:10: error: format '%ld' expects argument of type 'long int', but argument 4 has type 'int' [-Werror=format=]
> DMEMIT("%ld %llu %llu %llu", writecache_has_error(wc),
> ^~~~~~~~~~~~~~~~~~~~
> include/linux/device-mapper.h:549:46: note: in definition of macro 'DMEMIT'
> 0 : scnprintf(result + sz, maxlen - sz, x))
> ^
>
> The code is otherwise correct, so we just need to shut up the warning,
> which can be done using an extra type cast.
>
> Fixes: bb15b431d650 ("dm: add writecache target")
> Signed-off-by: Arnd Bergmann <[email protected]>
Thanks, I've picked this up.
On Mon, 28 May 2018, Arnd Bergmann wrote:
> The return type of ACCESS_ONCE is configuration dependent and may be either
> 'int' or 'long int' for the writecache_has_error() macro, so we get a warning
> like this for either format string:
__builtin_expect returns always long, see the GCC documentation (it used
to return int in very old gcc versions such as 2.96).
I think this is a bug in the macro __branch_check__. The variable ______r
should be long, but it is int. This bug may cause misbehavior of other
kernel parts (i.e. truncation of long value to int), so it should be fixed
in __branch_check__ - not in dm-writecache.
Mikulas
> In file included from drivers/md/dm-writecache.c:8:
> drivers/md/dm-writecache.c: In function 'writecache_status':
> drivers/md/dm-writecache.c:2227:10: error: format '%ld' expects argument of type 'long int', but argument 4 has type 'int' [-Werror=format=]
> DMEMIT("%ld %llu %llu %llu", writecache_has_error(wc),
> ^~~~~~~~~~~~~~~~~~~~
> include/linux/device-mapper.h:549:46: note: in definition of macro 'DMEMIT'
> 0 : scnprintf(result + sz, maxlen - sz, x))
> ^
>
> The code is otherwise correct, so we just need to shut up the warning,
> which can be done using an extra type cast.
>
> Fixes: bb15b431d650 ("dm: add writecache target")
> Signed-off-by: Arnd Bergmann <[email protected]>
> ---
> drivers/md/dm-writecache.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/md/dm-writecache.c b/drivers/md/dm-writecache.c
> index 1ef06e738eb6..772ac3a57287 100644
> --- a/drivers/md/dm-writecache.c
> +++ b/drivers/md/dm-writecache.c
> @@ -2224,7 +2224,7 @@ static void writecache_status(struct dm_target *ti, status_type_t type,
>
> switch (type) {
> case STATUSTYPE_INFO:
> - DMEMIT("%ld %llu %llu %llu", writecache_has_error(wc),
> + DMEMIT("%ld %llu %llu %llu", (long)writecache_has_error(wc),
> (unsigned long long)wc->n_blocks, (unsigned long long)wc->freelist_size,
> (unsigned long long)wc->writeback_size);
> break;
> --
> 2.9.0
>
The function __builtin_expect returns long type (see the gcc
documentation), and so do macros likely and unlikely. Unfortunatelly, when
CONFIG_PROFILE_ANNOTATED_BRANCHES is selected, the macros likely and
unlikely expand to __branch_check__ and __branch_check__ truncates the
long type to int. This unintended truncation may cause bugs in various
kernel code (we found a bug in dm-writecache because of it), so it's
better to fix __branch_check__ to return long.
Signed-off-by: Mikulas Patocka <[email protected]>
Cc: [email protected]
---
include/linux/compiler.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: linux-2.6/include/linux/compiler.h
===================================================================
--- linux-2.6.orig/include/linux/compiler.h 2018-02-26 20:34:17.000000000 +0100
+++ linux-2.6/include/linux/compiler.h 2018-05-30 14:11:53.000000000 +0200
@@ -21,7 +21,7 @@ void ftrace_likely_update(struct ftrace_
#define unlikely_notrace(x) __builtin_expect(!!(x), 0)
#define __branch_check__(x, expect, is_constant) ({ \
- int ______r; \
+ long ______r; \
static struct ftrace_likely_data \
__attribute__((__aligned__(4))) \
__attribute__((section("_ftrace_annotated_branch"))) \
On Wed, 30 May 2018 08:19:22 -0400 (EDT)
Mikulas Patocka <[email protected]> wrote:
> The function __builtin_expect returns long type (see the gcc
> documentation), and so do macros likely and unlikely. Unfortunatelly, when
> CONFIG_PROFILE_ANNOTATED_BRANCHES is selected, the macros likely and
> unlikely expand to __branch_check__ and __branch_check__ truncates the
Nice catch.
> long type to int. This unintended truncation may cause bugs in various
> kernel code (we found a bug in dm-writecache because of it), so it's
I'm curious to what that bug was.
> better to fix __branch_check__ to return long.
>
> Signed-off-by: Mikulas Patocka <[email protected]>
> Cc: [email protected]
Anyway, I can pull this in my tree and test it.
-- Steve
>
> ---
> include/linux/compiler.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: linux-2.6/include/linux/compiler.h
> ===================================================================
> --- linux-2.6.orig/include/linux/compiler.h 2018-02-26 20:34:17.000000000 +0100
> +++ linux-2.6/include/linux/compiler.h 2018-05-30 14:11:53.000000000 +0200
> @@ -21,7 +21,7 @@ void ftrace_likely_update(struct ftrace_
> #define unlikely_notrace(x) __builtin_expect(!!(x), 0)
>
> #define __branch_check__(x, expect, is_constant) ({ \
> - int ______r; \
> + long ______r; \
> static struct ftrace_likely_data \
> __attribute__((__aligned__(4))) \
> __attribute__((section("_ftrace_annotated_branch"))) \
On Mon, 4 Jun 2018, Steven Rostedt wrote:
> On Wed, 30 May 2018 08:19:22 -0400 (EDT)
> Mikulas Patocka <[email protected]> wrote:
>
> > The function __builtin_expect returns long type (see the gcc
> > documentation), and so do macros likely and unlikely. Unfortunatelly, when
> > CONFIG_PROFILE_ANNOTATED_BRANCHES is selected, the macros likely and
> > unlikely expand to __branch_check__ and __branch_check__ truncates the
>
> Nice catch.
>
> > long type to int. This unintended truncation may cause bugs in various
> > kernel code (we found a bug in dm-writecache because of it), so it's
>
> I'm curious to what that bug was.
printk("%ld", writecache_has_error(wc))
... and writecache_has_error was defined as
#define writecache_has_error(wc) (unlikely(READ_ONCE((wc)->error)))
Mikulas