lkml.org 
[lkml]   [2021]   [Feb]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: linux-next: build warning after merge of the v4l-dvb tree
Em Mon, 15 Feb 2021 11:20:24 +0100
Mauro Carvalho Chehab <mchehab@kernel.org> escreveu:

> Em Mon, 08 Feb 2021 15:53:00 -0300
> Ezequiel Garcia <ezequiel@collabora.com> escreveu:
>
> > On Mon, 2021-02-08 at 18:40 +0100, Mauro Carvalho Chehab wrote:
> > > Em Mon, 08 Feb 2021 13:57:56 -0300
> > > Ezequiel Garcia <ezequiel@collabora.com> escreveu:
> > >
> > > > On Mon, 2021-02-08 at 18:46 +0200, Sakari Ailus wrote:
> > > > > Hi Ezequiel,
> > > > >
> > > > > Thanks for addressing this.
> > > > >
> > > > > On Mon, Feb 08, 2021 at 01:42:21PM -0300, Ezequiel Garcia wrote: 
> > > > > > Hi Stephen,
> > > > > >
> > > > > > On Mon, 2021-02-08 at 23:37 +1100, Stephen Rothwell wrote: 
> > > > > > > Hi all,
> > > > > > >
> > > > > > > After merging the v4l-dvb tree, today's linux-next build (htmldocs)
> > > > > > > produced this warning:
> > > > > > >
> > > > > > > include/media/v4l2-async.h:178: warning: expecting prototype for v4l2_async_notifier_add_fwnode_subdev(). Prototype was for
> > > > > > > __v4l2_async_notifier_add_fwnode_subdev() instead
> > > > > > > include/media/v4l2-async.h:207: warning: expecting prototype for v4l2_async_notifier_add_fwnode_remote_subdev(). Prototype was for
> > > > > > > __v4l2_async_notifier_add_fwnode_remote_subdev() instead
> > > > > > > include/media/v4l2-async.h:230: warning: expecting prototype for v4l2_async_notifier_add_i2c_subdev(). Prototype was for
> > > > > > > __v4l2_async_notifier_add_i2c_subdev() instead
> > > > > > >
> > > > > > > Maybe introduced by commit
> > > > > > >
> > > > > > >   c1cc23625062 ("media: v4l2-async: Discourage use of v4l2_async_notifier_add_subdev")
> > > > > > >  
> > > > > >
> > > > > > Thanks for spotting this. Should be fixed by:
> > > > > >
> > > > > > diff --git a/include/media/v4l2-async.h b/include/media/v4l2-async.h
> > > > > > index 6f22daa6f067..3785445282fc 100644
> > > > > > --- a/include/media/v4l2-async.h
> > > > > > +++ b/include/media/v4l2-async.h
> > > > > > @@ -157,7 +157,7 @@ int __v4l2_async_notifier_add_subdev(struct v4l2_async_notifier *notifier,
> > > > > >                                    struct v4l2_async_subdev *asd);
> > > > > >  
> > > > > >  /**
> > > > > > - * v4l2_async_notifier_add_fwnode_subdev - Allocate and add a fwnode async
> > > > > > + * __v4l2_async_notifier_add_fwnode_subdev - Allocate and add a fwnode async 
> > > > >
> > > > > The problem with the approach is that this no longer documents the API that
> > > > > drivers are intended to use, but the intermediate one.
> > >
> > > Yep. the better would be to keep documenting what will be used.
> > >
> >
> > Is there a way to silence/ignore the warning for a specific function(s)?
>
> No. The warning is there for a very good reason: if you do something like:
>
>
> /**
> * delete - delete some file
> *
> * This function deletes something.
> */
> void insert() {}
> /**
> * insert - creates a new file
> *
> * This function creates a file and inserts something.
> */
> void delete() {}
>
> the output of it will be:
>
> $ ./scripts/kernel-doc -sphinx-version 3.0.0 silly.c
> .. c:function:: void insert ()
>
> delete some file
>
> **Parameters**
>
> **Description**
>
>
> This function deletes something.
>
>
> .. c:function:: void delete ()
>
> creates a new file
>
> **Parameters**
>
> **Description**
>
>
> This function creates a file and inserts something.
>
> Which is completely wrong. If someone tries to write a code using the
> above, the result will be just the opposite than what it was documented.
>
> Btw, I noticed several places where things like that happened, because
> some code were added between a Kernel-doc markup and some function.

Btw, in the specific case of this change, something like the above
quick hack would fix it.

PS: I didn't think much when I wrote the __type description.

Also, keeping the symbols that should be documented as __foo
doesn't seem the right thing to me :-)


---

[PATCH] v4l2-async.h: Fix kerneldoc markups

Document the functions that should be used by the kAPI clients,
instead of the internal functions.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

diff --git a/include/media/v4l2-async.h b/include/media/v4l2-async.h
index 6f22daa6f067..91dbeaa4e6ee 100644
--- a/include/media/v4l2-async.h
+++ b/include/media/v4l2-async.h
@@ -156,42 +156,38 @@ void v4l2_async_notifier_init(struct v4l2_async_notifier *notifier);
int __v4l2_async_notifier_add_subdev(struct v4l2_async_notifier *notifier,
struct v4l2_async_subdev *asd);

+struct v4l2_async_subdev *
+__v4l2_async_notifier_add_fwnode_subdev(struct v4l2_async_notifier *notifier,
+ struct fwnode_handle *fwnode,
+ unsigned int asd_struct_size);
/**
* v4l2_async_notifier_add_fwnode_subdev - Allocate and add a fwnode async
* subdev to the notifier's master asd_list.
*
- * @notifier: pointer to &struct v4l2_async_notifier
- * @fwnode: fwnode handle of the sub-device to be matched
- * @asd_struct_size: size of the driver's async sub-device struct, including
- * sizeof(struct v4l2_async_subdev). The &struct
- * v4l2_async_subdev shall be the first member of
- * the driver's async sub-device struct, i.e. both
- * begin at the same memory address.
+ * @__notifier: pointer to &struct v4l2_async_notifier
+ * @__fwnode: fwnode handle of the sub-device to be matched
+ * @__type: type of the struct that contains a struct v4l2_async_subdev.
*
* Allocate a fwnode-matched asd of size asd_struct_size, and add it to the
* notifiers @asd_list. The function also gets a reference of the fwnode which
* is released later at notifier cleanup time.
*/
-struct v4l2_async_subdev *
-__v4l2_async_notifier_add_fwnode_subdev(struct v4l2_async_notifier *notifier,
- struct fwnode_handle *fwnode,
- unsigned int asd_struct_size);
#define v4l2_async_notifier_add_fwnode_subdev(__notifier, __fwnode, __type) \
((__type *)__v4l2_async_notifier_add_fwnode_subdev(__notifier, __fwnode, \
sizeof(__type)))

+struct v4l2_async_subdev *
+__v4l2_async_notifier_add_fwnode_remote_subdev(struct v4l2_async_notifier *notif,
+ struct fwnode_handle *endpoint,
+ unsigned int asd_struct_size);
/**
* v4l2_async_notifier_add_fwnode_remote_subdev - Allocate and add a fwnode
* remote async subdev to the
* notifier's master asd_list.
*
- * @notif: pointer to &struct v4l2_async_notifier
- * @endpoint: local endpoint pointing to the remote sub-device to be matched
- * @asd_struct_size: size of the driver's async sub-device struct, including
- * sizeof(struct v4l2_async_subdev). The &struct
- * v4l2_async_subdev shall be the first member of
- * the driver's async sub-device struct, i.e. both
- * begin at the same memory address.
+ * @__notifier: pointer to &struct v4l2_async_notifier
+ * @__ep: local endpoint pointing to the remote sub-device to be matched
+ * @__type: type of the struct that contains a struct v4l2_async_subdev.
*
* Gets the remote endpoint of a given local endpoint, set it up for fwnode
* matching and adds the async sub-device to the notifier's @asd_list. The
@@ -201,33 +197,25 @@ __v4l2_async_notifier_add_fwnode_subdev(struct v4l2_async_notifier *notifier,
* This is just like @v4l2_async_notifier_add_fwnode_subdev, but with the
* exception that the fwnode refers to a local endpoint, not the remote one.
*/
-struct v4l2_async_subdev *
-__v4l2_async_notifier_add_fwnode_remote_subdev(struct v4l2_async_notifier *notif,
- struct fwnode_handle *endpoint,
- unsigned int asd_struct_size);
#define v4l2_async_notifier_add_fwnode_remote_subdev(__notifier, __ep, __type) \
((__type *)__v4l2_async_notifier_add_fwnode_remote_subdev(__notifier, __ep, \
sizeof(__type)))

+struct v4l2_async_subdev *
+__v4l2_async_notifier_add_i2c_subdev(struct v4l2_async_notifier *notifier,
+ int adapter_id, unsigned short address,
+ unsigned int asd_struct_size);
/**
* v4l2_async_notifier_add_i2c_subdev - Allocate and add an i2c async
* subdev to the notifier's master asd_list.
*
- * @notifier: pointer to &struct v4l2_async_notifier
- * @adapter_id: I2C adapter ID to be matched
- * @address: I2C address of sub-device to be matched
- * @asd_struct_size: size of the driver's async sub-device struct, including
- * sizeof(struct v4l2_async_subdev). The &struct
- * v4l2_async_subdev shall be the first member of
- * the driver's async sub-device struct, i.e. both
- * begin at the same memory address.
+ * @__notifier: pointer to &struct v4l2_async_notifier
+ * @__adap: I2C adapter ID to be matched
+ * @__addr: I2C address of sub-device to be matched
+ * @__type: type of the struct that contains a struct v4l2_async_subdev.
*
* Same as above but for I2C matched sub-devices.
*/
-struct v4l2_async_subdev *
-__v4l2_async_notifier_add_i2c_subdev(struct v4l2_async_notifier *notifier,
- int adapter_id, unsigned short address,
- unsigned int asd_struct_size);
#define v4l2_async_notifier_add_i2c_subdev(__notifier, __adap, __addr, __type) \
((__type *)__v4l2_async_notifier_add_i2c_subdev(__notifier, __adap, __addr, \
sizeof(__type)))

Thanks,
Mauro

\
 
 \ /
  Last update: 2021-02-15 11:52    [W:2.020 / U:0.048 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site