Hi Rusty,
After merging the modules tree, today's linux-next build (x86_64
allmodconfig) failed like this:
In file included from include/linux/linkage.h:4:0,
from include/linux/kernel.h:6,
from kernel/params.c:18:
kernel/params.c: In function 'param_array':
kernel/params.c:32:42: error: dereferencing pointer to incomplete type
#define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
^
include/linux/compiler.h:164:42: note: in definition of macro 'unlikely'
# define unlikely(x) __builtin_expect(!!(x), 0)
^
kernel/params.c:462:3: note: in expansion of macro 'BUG_ON'
BUG_ON(!KPARAM_IS_LOCKED(mod));
^
kernel/params.c:33:47: note: in expansion of macro 'KPARAM_MUTEX'
#define KPARAM_IS_LOCKED(mod) mutex_is_locked(KPARAM_MUTEX(mod))
^
kernel/params.c:462:11: note: in expansion of macro 'KPARAM_IS_LOCKED'
BUG_ON(!KPARAM_IS_LOCKED(mod));
^
kernel/params.c: In function 'param_array_get':
kernel/params.c:32:42: error: dereferencing pointer to incomplete type
#define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
^
include/linux/compiler.h:164:42: note: in definition of macro 'unlikely'
# define unlikely(x) __builtin_expect(!!(x), 0)
^
kernel/params.c:499:3: note: in expansion of macro 'BUG_ON'
BUG_ON(!KPARAM_IS_LOCKED(p.mod));
^
kernel/params.c:33:47: note: in expansion of macro 'KPARAM_MUTEX'
#define KPARAM_IS_LOCKED(mod) mutex_is_locked(KPARAM_MUTEX(mod))
^
kernel/params.c:499:11: note: in expansion of macro 'KPARAM_IS_LOCKED'
BUG_ON(!KPARAM_IS_LOCKED(p.mod));
^
In file included from include/linux/notifier.h:13:0,
from include/linux/memory_hotplug.h:6,
from include/linux/mmzone.h:789,
from include/linux/gfp.h:5,
from include/linux/kmod.h:22,
from include/linux/module.h:13,
from kernel/params.c:21:
kernel/params.c: In function 'kernel_param_lock':
kernel/params.c:32:42: error: dereferencing pointer to incomplete type
#define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
^
include/linux/mutex.h:146:44: note: in definition of macro 'mutex_lock'
#define mutex_lock(lock) mutex_lock_nested(lock, 0)
^
kernel/params.c:621:13: note: in expansion of macro 'KPARAM_MUTEX'
mutex_lock(KPARAM_MUTEX(mod));
^
kernel/params.c: In function 'kernel_param_unlock':
kernel/params.c:32:42: error: dereferencing pointer to incomplete type
#define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
^
kernel/params.c:626:15: note: in expansion of macro 'KPARAM_MUTEX'
mutex_unlock(KPARAM_MUTEX(mod));
^
Caused by commit b51d23e4e9fe ("module: add per-module param_lock").
I am not exactly sure why this fails to build. It may be an
interaction with some other change.
I have used the modules tree from next-20150623 for today.
--
Cheers,
Stephen Rothwell [email protected]
On Tue, Jun 23, 2015 at 9:37 PM, Stephen Rothwell <[email protected]> wrote:
> Hi Rusty,
>
> After merging the modules tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
that's weird. Are you sure it failed during allmodconfig? I can see
why it would fail like that if CONFIG_MODULES ins't defined, which
I'll send a patch for...
>
> In file included from include/linux/linkage.h:4:0,
> from include/linux/kernel.h:6,
> from kernel/params.c:18:
> kernel/params.c: In function 'param_array':
> kernel/params.c:32:42: error: dereferencing pointer to incomplete type
> #define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
> ^
> include/linux/compiler.h:164:42: note: in definition of macro 'unlikely'
> # define unlikely(x) __builtin_expect(!!(x), 0)
> ^
> kernel/params.c:462:3: note: in expansion of macro 'BUG_ON'
> BUG_ON(!KPARAM_IS_LOCKED(mod));
> ^
> kernel/params.c:33:47: note: in expansion of macro 'KPARAM_MUTEX'
> #define KPARAM_IS_LOCKED(mod) mutex_is_locked(KPARAM_MUTEX(mod))
> ^
> kernel/params.c:462:11: note: in expansion of macro 'KPARAM_IS_LOCKED'
> BUG_ON(!KPARAM_IS_LOCKED(mod));
> ^
> kernel/params.c: In function 'param_array_get':
> kernel/params.c:32:42: error: dereferencing pointer to incomplete type
> #define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
> ^
> include/linux/compiler.h:164:42: note: in definition of macro 'unlikely'
> # define unlikely(x) __builtin_expect(!!(x), 0)
> ^
> kernel/params.c:499:3: note: in expansion of macro 'BUG_ON'
> BUG_ON(!KPARAM_IS_LOCKED(p.mod));
> ^
> kernel/params.c:33:47: note: in expansion of macro 'KPARAM_MUTEX'
> #define KPARAM_IS_LOCKED(mod) mutex_is_locked(KPARAM_MUTEX(mod))
> ^
> kernel/params.c:499:11: note: in expansion of macro 'KPARAM_IS_LOCKED'
> BUG_ON(!KPARAM_IS_LOCKED(p.mod));
> ^
> In file included from include/linux/notifier.h:13:0,
> from include/linux/memory_hotplug.h:6,
> from include/linux/mmzone.h:789,
> from include/linux/gfp.h:5,
> from include/linux/kmod.h:22,
> from include/linux/module.h:13,
> from kernel/params.c:21:
> kernel/params.c: In function 'kernel_param_lock':
> kernel/params.c:32:42: error: dereferencing pointer to incomplete type
> #define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
> ^
> include/linux/mutex.h:146:44: note: in definition of macro 'mutex_lock'
> #define mutex_lock(lock) mutex_lock_nested(lock, 0)
> ^
> kernel/params.c:621:13: note: in expansion of macro 'KPARAM_MUTEX'
> mutex_lock(KPARAM_MUTEX(mod));
> ^
> kernel/params.c: In function 'kernel_param_unlock':
> kernel/params.c:32:42: error: dereferencing pointer to incomplete type
> #define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
> ^
> kernel/params.c:626:15: note: in expansion of macro 'KPARAM_MUTEX'
> mutex_unlock(KPARAM_MUTEX(mod));
> ^
>
> Caused by commit b51d23e4e9fe ("module: add per-module param_lock").
>
> I am not exactly sure why this fails to build. It may be an
> interaction with some other change.
>
> I have used the modules tree from next-20150623 for today.
>
> --
> Cheers,
> Stephen Rothwell [email protected]
Simplify params.c by replacing KPARAM_MUTEX() with an inline function
__param_lock(), and replacing KPARAM_IS_LOCKED() with direct calls to
mutex_is_locked(). Inside __param_lock(), only (conditionally) return
mod->param_lock #ifdef CONFIG_MODULES; if there are no modules just
return the built-in param_lock.
Since "struct module" isn't defined if !CONFIG_MODULES, we must leave
out the code that returns mod->param_lock if there are no modules.
Signed-off-by: Dan Streetman <[email protected]>
---
kernel/params.c | 18 ++++++++++++------
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/kernel/params.c b/kernel/params.c
index 8890d0b..9c955cd 100644
--- a/kernel/params.c
+++ b/kernel/params.c
@@ -29,8 +29,14 @@
static DEFINE_MUTEX(param_lock);
/* Use the module's mutex, or if built-in use the built-in mutex */
-#define KPARAM_MUTEX(mod) ((mod) ? &(mod)->param_lock : ¶m_lock)
-#define KPARAM_IS_LOCKED(mod) mutex_is_locked(KPARAM_MUTEX(mod))
+static inline struct mutex *__param_lock(struct module *mod)
+{
+#ifdef CONFIG_MODULES
+ return mod ? &mod->param_lock : ¶m_lock;
+#else
+ return ¶m_lock;
+#endif
+}
/* This just allows us to keep track of which parameters are kmalloced. */
struct kmalloced_param {
@@ -459,7 +465,7 @@ static int param_array(struct module *mod,
/* nul-terminate and parse */
save = val[len];
((char *)val)[len] = '\0';
- BUG_ON(!KPARAM_IS_LOCKED(mod));
+ BUG_ON(!mutex_is_locked(__param_lock(mod)));
ret = set(val, &kp);
if (ret != 0)
@@ -496,7 +502,7 @@ static int param_array_get(char *buffer, const struct kernel_param *kp)
if (i)
buffer[off++] = ',';
p.arg = arr->elem + arr->elemsize * i;
- BUG_ON(!KPARAM_IS_LOCKED(p.mod));
+ BUG_ON(!mutex_is_locked(__param_lock(p.mod)));
ret = arr->ops->get(buffer + off, &p);
if (ret < 0)
return ret;
@@ -618,12 +624,12 @@ static ssize_t param_attr_store(struct module_attribute *mattr,
void kernel_param_lock(struct module *mod)
{
- mutex_lock(KPARAM_MUTEX(mod));
+ mutex_lock(__param_lock(mod));
}
void kernel_param_unlock(struct module *mod)
{
- mutex_unlock(KPARAM_MUTEX(mod));
+ mutex_unlock(__param_lock(mod));
}
#ifdef CONFIG_SYSFS
--
2.1.0
Hi Dan,
On Wed, 24 Jun 2015 14:18:44 -0400 Dan Streetman <[email protected]> wrote:
>
> On Tue, Jun 23, 2015 at 9:37 PM, Stephen Rothwell <[email protected]> wrote:
> >
> > After merging the modules tree, today's linux-next build (x86_64
> > allmodconfig) failed like this:
>
> that's weird. Are you sure it failed during allmodconfig? I can see
> why it would fail like that if CONFIG_MODULES ins't defined, which
> I'll send a patch for...
Pretty sure - and, in any case, I don't do any CONFIG_MODULES=n builds
between tree merges (only later in the day). That is why I couldn't
figure out what went wrong.
I will apply your patch today and see if that helps.
--
Cheers,
Stephen Rothwell [email protected]
Hi Dan,
On Thu, 25 Jun 2015 08:57:06 +1000 Stephen Rothwell <[email protected]> wrote:
>
> On Wed, 24 Jun 2015 14:18:44 -0400 Dan Streetman <[email protected]> wrote:
> >
> > On Tue, Jun 23, 2015 at 9:37 PM, Stephen Rothwell <[email protected]> wrote:
> > >
> > > After merging the modules tree, today's linux-next build (x86_64
> > > allmodconfig) failed like this:
> >
> > that's weird. Are you sure it failed during allmodconfig? I can see
> > why it would fail like that if CONFIG_MODULES ins't defined, which
> > I'll send a patch for...
>
> Pretty sure - and, in any case, I don't do any CONFIG_MODULES=n builds
> between tree merges (only later in the day). That is why I couldn't
> figure out what went wrong.
>
> I will apply your patch today and see if that helps.
I built without your patch and it failed again, but applying your patch fixes it.
Rusty, you can consider this
Tested-by: Stephen Rothwell <[email protected]>
for "[PATCH] modules: only use mod->param_lock if CONFIG_MODULES"
--
Cheers,
Stephen Rothwell [email protected]