RE: [PATCH 22/59] Staging: hv: vmbus: Get rid of the unused name field in struct hv_driver

-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, August 25, 2011 5:28 PM
To: KY Srinivasan
Cc: gregkh@suse.de; linux-kernel@vger.kernel.org;
devel@linuxdriverproject.org; virtualization@lists.osdl.org; Haiyang Zhang
Subject: Re: [PATCH 22/59] Staging: hv: vmbus: Get rid of the unused name field
in struct hv_driver

On Thu, Aug 25, 2011 at 02:24:28PM -0700, Greg KH wrote:

quoted

On Thu, Aug 25, 2011 at 09:48:48AM -0700, K. Y. Srinivasan wrote:

quoted

Get rid of the unused "name" field in struct hv_driver.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
---
 drivers/staging/hv/hyperv.h |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/drivers/staging/hv/hyperv.h b/drivers/staging/hv/hyperv.h
index b8199f4..60ead66 100644
--- a/drivers/staging/hv/hyperv.h
+++ b/drivers/staging/hv/hyperv.h

@@ -802,8 +802,6 @@ struct hv_device_info {

 /* Base driver object */
 struct hv_driver {
-	const char *name;

Wait, why is this unused?  What are you going to use as your name for
the driver in sysfs then?  The module name?

As much as I love seeing things deleted, I really think you need this
field.

Ah, yeah, I see why you think it's unneeded, crud like this in the
drivers:

	drv->driver.name = driver_name;

No vmbus driver should ever have to touch the base struct driver on it's
own at all.  Your vmbus core should properly handle telling the driver
core what the name of the driver is.

As an example, see the __pci_register_driver() function, the first thing
that code does is set the name based on the name of the larger
pci_driver structure passed to it.

Man, if you want something done right, you have to do it yourself, let
me go make these changes so you don't have to do any new work at this
point in time, hopefully your other patches will apply...

What, vmbus_child_driver_register() takes a struct driver *?

No wonder things are so messed up here, and why you got confused.  Let
me pound on this for a bit to see if I can get it cleaned up to be more
"sane"...

Greg,

This was existing code and I realized it did not conform to the Linux
Driver Model. I fixed it in the patch set I had sent earlier (110/117).
Yesterday I only got up to 74 of the 117 patches that you commented on.
I was planning to send the rest today. If it is ok with you, please don't fix it here
As it would break everything else. This is existing code that I am fixing in the patch set
I will send you shortly.

Regards,

K. Y

greg k-h