Re: [2.6.24-rc6-mm1]Build failure in drivers/net/ehea/ehea_main.c

[Jan-Bernd Themann <ossthema@de.ibm.com>]

On Monday 28 January 2008 20:22, you wrote:
> Greg KH wrote:
> > On Fri, Jan 25, 2008 at 01:10:48PM -0600, Nathan Lynch wrote:
> > > Jan-Bernd Themann wrote:
> > > 
> > > This is now broken in mainline...
> > > 
> > > drivers/net/ehea/ehea_main.c: In function 'ehea_driver_sysfs_add':
> > > drivers/net/ehea/ehea_main.c:2812: error: 'struct device_driver' has
> > > no member named 'kobj'
> > > drivers/net/ehea/ehea_main.c:2815: error: 'struct device_driver' has
> > > no member named 'kobj'
> > > drivers/net/ehea/ehea_main.c:2818: error: 'struct device_driver' has
> > > no member named 'kobj'
> > > drivers/net/ehea/ehea_main.c: In function 'ehea_driver_sysfs_remove':
> > > drivers/net/ehea/ehea_main.c:2830: error: 'struct device_driver' has
> > > no member named 'kobj'
> > 
> > Does the patch below fix this?  That driver should not have been trying
> > to create symlinks that the driver core has already created for it.
> 
> Yes, it fixes the build error, by just removing the code that got
> broken.  Jan-Bernd gave a rationale for creating the symlink that
> didn't really seem to be answered.
> 

> > > > The eHEA driver tries to orginize its sys-entries as close as possible to
> > > > other ethernet drivers. Each eHEA NIC has multiple ports which is not that
> > > > common in PCI. This means that each port is represented by a subdirectory
> > > > which has not the "driver" sys-link, only the root directory has.
> > > > Some tools expect to have this driver link in each port directory.
> > > > That is the reason why this link is created manually.

In addition to the explaination above:

Just to be sure that we talk about the same thing:
The eHEA driver controlls multiple eHEAs which have multiple
ethernet ports each. So the logical structure looks like this:
eHEA1 ---> port1 (ethX)
        -> port2 (ethX+1)
        -> port3 (ethX+2)
   ...
eHEA2 ---> port1 (eth?)
        -> port2
        -> port3 
   ...
This structure is represented in /sys/bus/ibmebus/devices in the same way
described above.

If you want to determine the driver belonging to an ethX, you would go
the following path:
/sys/class/net/ethX/device/driver


ls /sys/class/net/ethX/device:
drwxr-xr-x 2 root root    0 2008-01-29 10:00 .
drwxr-xr-x 4 root root    0 2008-01-29 10:00 ..
lrwxrwxrwx 1 root root    0 2008-01-29 10:00 bus -> ../../../../bus/ibmebus
-r--r--r-- 1 root root 4096 2008-01-29 10:00 devspec
lrwxrwxrwx 1 root root    0 2008-01-29 10:00 driver -> ../../../../bus/ibmebus/drivers/ehea
                                             ^^^^^
	In our case this one is created by the code that does not work any longer.

The sym-link is not gereated automatically as the device for portX is added
to the eHEA device (as subnode) where the eHEA device is not a bus.

If this sym-link is of interest (which I guess is the case as most devices
have it) we have to create it somehow.

Would the following proposal work?

There is an exported kernel function called 
"device_bind_driver" which creates the same links. However, it does some
additional stuff as well and there is currently no 
"device_unbind_driver" yet. Either "device_unbind_driver" needs to
be added or the following two already existing functions could be exported:

driver_sysfs_add
driver_sysfs_remove


Regards,
Jan-Bernd & Christoph Raisch