On Tue, Nov 26, 2019 at 6:52 PM Jean Delvare <jdelvare@xxxxxxx> wrote:
>
> On Mon, 25 Nov 2019 15:48:57 +0100, Wolfram Sang wrote:
> > On Tue, Nov 19, 2019 at 07:23:52PM +0800, Lei YU wrote:
> > > In i2c/dev-interface.rst it said
> > >
> > > > All these transactions return -1 on failure
> > >
> > > But actually the i2c_smbus_xxx functions return negative error numbers
> > > on failure, instead of -1.
> > >
> > > Fix the document and remove the following sentence.
> > >
> > > Signed-off-by: Lei YU <mine260309@xxxxxxxxx>
> > > ---
> > > Documentation/i2c/dev-interface.rst | 9 ++++-----
> > > 1 file changed, 4 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/Documentation/i2c/dev-interface.rst b/Documentation/i2c/dev-interface.rst
> > > index 69c23a3..f2f2b28 100644
> > > --- a/Documentation/i2c/dev-interface.rst
> > > +++ b/Documentation/i2c/dev-interface.rst
> > > @@ -163,11 +163,10 @@ for details) through the following functions::
> > > __s32 i2c_smbus_write_block_data(int file, __u8 command, __u8 length,
> > > __u8 *values);
> > >
> > > -All these transactions return -1 on failure; you can read errno to see
> > > -what happened. The 'write' transactions return 0 on success; the
> > > -'read' transactions return the read value, except for read_block, which
> > > -returns the number of values read. The block buffers need not be longer
> > > -than 32 bytes.
> > > +All these transactions return a negative error number on failure.
> > > +The 'write' transactions return 0 on success; the 'read' transactions
> > > +return the read value, except for read_block, which returns the number
> > > +of values read. The block buffers need not be longer than 32 bytes.
> >
> > I think the correct solution is to remove this paragraph entirely.
> > Because the returned value does not depend on the kernel but on the
> > libi2c version. Check this commit from 2012 in the i2c-tools repo:
> >
> > 330bba2 ("libi2c: Properly propagate real error codes on read errors")
> >
> > So, I think we should document it there. Jean, what do you think?
>
> I would go further and move half of the document to i2c-tools. i2c-dev
> itself only provides the ioctls. Everything on top of that is in libi2c
> now, so the kernel documentation should point to libi2c and the
> detailed documentation should come with libi2c.
Yeah, I sent the patch to simply fix the "return -1" issue, which is misleading.
But if the whole paragraph or the whole document is not valid anymore,
it needs to be fixed.
>
> So I guess I should review the whole document now to see what needs to
> be updated, what should stay, and what should move.
Thanks, please help to fix the whole document so that others get the
correct information :)
>
> --
> Jean Delvare
> SUSE L3 Support
