[PATCH] net/hyperv: Advertise hv_netvsc instead of netvsc as driver name

From: Olaf Hering <redacted>

Signed-off-by: Olaf Hering <redacted>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
---
 drivers/net/hyperv/netvsc_drv.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c
index 0ae7a1a..53d1ee5 100644
--- a/drivers/net/hyperv/netvsc_drv.c
+++ b/drivers/net/hyperv/netvsc_drv.c

@@ -40,6 +40,8 @@
 
 #include "hyperv_net.h"
 
+#define MODULE_NAME "hv_netvsc"
+
 struct net_device_context {
 	/* point back to our device context */
 	struct hv_device *device_ctx;

@@ -310,7 +312,7 @@ int netvsc_recv_callback(struct hv_device *device_obj,
 static void netvsc_get_drvinfo(struct net_device *net,
 			       struct ethtool_drvinfo *info)
 {
-	strcpy(info->driver, "hv_netvsc");
+	strcpy(info->driver, MODULE_NAME);
 	strcpy(info->version, HV_DRV_VERSION);
 	strcpy(info->fw_version, "N/A");
 }

@@ -482,7 +484,7 @@ MODULE_DEVICE_TABLE(vmbus, id_table);
 
 /* The one and only one */
 static struct  hv_driver netvsc_drv = {
-	.name = "netvsc",
+	.name = MODULE_NAME,
 	.id_table = id_table,
 	.probe = netvsc_probe,
 	.remove = netvsc_remove,

-- 
1.7.4.1

On Wed, Mar 07, 2012 at 11:05:31AM -0800, Haiyang Zhang wrote:

No, please use the built-in macro for this and don't define your own,
that's totally redundant.

greg k-h

-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Wednesday, March 07, 2012 2:22 PM
To: Haiyang Zhang
Cc: KY Srinivasan; davem@davemloft.net; netdev@vger.kernel.org; linux-
kernel@vger.kernel.org; devel@linuxdriverproject.org; Olaf Hering
Subject: Re: [PATCH] net/hyperv: Advertise hv_netvsc instead of netvsc as
driver name

On Wed, Mar 07, 2012 at 11:05:31AM -0800, Haiyang Zhang wrote:

quoted

From: Olaf Hering <redacted>

Signed-off-by: Olaf Hering <redacted>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
---
 drivers/net/hyperv/netvsc_drv.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/hyperv/netvsc_drv.c

b/drivers/net/hyperv/netvsc_drv.c

quoted

index 0ae7a1a..53d1ee5 100644

--- a/drivers/net/hyperv/netvsc_drv.c
+++ b/drivers/net/hyperv/netvsc_drv.c

@@ -40,6 +40,8 @@

 #include "hyperv_net.h"

+#define MODULE_NAME "hv_netvsc"

No, please use the built-in macro for this and don't define your own,
that's totally redundant.

I have submitted a patch using the built-in KBUILD_MODNAME.
Thanks,
- Haiyang

On Wed, Mar 07, Haiyang Zhang wrote:

It would be nice if this could be applied to 3.3.
hv_storvsc and hid-hyperv already use the same naming, this patch was
send last year already but was not applied for some reason.

Olaf

On Wed, Mar 07, 2012 at 09:17:46PM +0100, Olaf Hering wrote:

On Wed, Mar 07, Haiyang Zhang wrote:

quoted

@@ -482,7 +484,7 @@ MODULE_DEVICE_TABLE(vmbus, id_table);
 
 /* The one and only one */
 static struct  hv_driver netvsc_drv = {
-	.name = "netvsc",
+	.name = MODULE_NAME,

It would be nice if this could be applied to 3.3.
hv_storvsc and hid-hyperv already use the same naming, this patch was
send last year already but was not applied for some reason.

What userspace tools depend on this that requires it to be in 3.3?

greg k-h

On Wed, Mar 07, Greg KH wrote:

On Wed, Mar 07, 2012 at 09:17:46PM +0100, Olaf Hering wrote:

quoted

On Wed, Mar 07, Haiyang Zhang wrote:

quoted

@@ -482,7 +484,7 @@ MODULE_DEVICE_TABLE(vmbus, id_table);
 
 /* The one and only one */
 static struct  hv_driver netvsc_drv = {
-	.name = "netvsc",
+	.name = MODULE_NAME,

It would be nice if this could be applied to 3.3.
hv_storvsc and hid-hyperv already use the same naming, this patch was
send last year already but was not applied for some reason.

What userspace tools depend on this that requires it to be in 3.3?

It represents the module names, at least for pci and other bus types.
hwinfo uses it to recognize a network interface (storage in case of
hv_storvsc).

Olaf

On Thu, Mar 08, 2012 at 07:16:16AM +0100, Olaf Hering wrote:

On Wed, Mar 07, Greg KH wrote:

quoted

On Wed, Mar 07, 2012 at 09:17:46PM +0100, Olaf Hering wrote:

quoted

On Wed, Mar 07, Haiyang Zhang wrote:

quoted

@@ -482,7 +484,7 @@ MODULE_DEVICE_TABLE(vmbus, id_table);
 
 /* The one and only one */
 static struct  hv_driver netvsc_drv = {
-	.name = "netvsc",
+	.name = MODULE_NAME,

It would be nice if this could be applied to 3.3.
hv_storvsc and hid-hyperv already use the same naming, this patch was
send last year already but was not applied for some reason.

What userspace tools depend on this that requires it to be in 3.3?

It represents the module names, at least for pci and other bus types.

Yes, but no one should care what the name of the module is.

hwinfo uses it to recognize a network interface (storage in case of
hv_storvsc).

That tool should report the real module name, not rely on the name to
match based on the structure, if you really care about the real module
name.

greg k-h

On Thu, Mar 08, Greg KH wrote:

quoted

hwinfo uses it to recognize a network interface (storage in case of
hv_storvsc).

That tool should report the real module name, not rely on the name to
match based on the structure, if you really care about the real module
name.

Its about the driver symlink, not the module name. 

I think most, if not all, (relevant) drivers use some sort of
KBUILD_MODNAME. Why should this driver be any different?

Olaf

On Thu, Mar 08, 2012 at 05:12:21PM +0100, Olaf Hering wrote:

On Thu, Mar 08, Greg KH wrote:

quoted

quoted

hwinfo uses it to recognize a network interface (storage in case of
hv_storvsc).

That tool should report the real module name, not rely on the name to
match based on the structure, if you really care about the real module
name.

Its about the driver symlink, not the module name. 

I think most, if not all, (relevant) drivers use some sort of
KBUILD_MODNAME. Why should this driver be any different?

It shouldn't, I'm not saying that at all.  I'm only objecting to the
insistance that this be something worthwhile for 3.3-final which should
be out any day now.

greg k-h

On Thu, Mar 08, Greg KH wrote:

On Thu, Mar 08, 2012 at 05:12:21PM +0100, Olaf Hering wrote:

quoted

On Thu, Mar 08, Greg KH wrote:

quoted

quoted

hwinfo uses it to recognize a network interface (storage in case of
hv_storvsc).

That tool should report the real module name, not rely on the name to
match based on the structure, if you really care about the real module
name.

Its about the driver symlink, not the module name. 

I think most, if not all, (relevant) drivers use some sort of
KBUILD_MODNAME. Why should this driver be any different?

It shouldn't, I'm not saying that at all.  I'm only objecting to the
insistance that this be something worthwhile for 3.3-final which should
be out any day now.

Since its an API we could settle on the new string, which is already in
distro use.

Olaf

On Thu, Mar 08, 2012 at 07:38:20PM +0100, Olaf Hering wrote:

On Thu, Mar 08, Greg KH wrote:

quoted

On Thu, Mar 08, 2012 at 05:12:21PM +0100, Olaf Hering wrote:

quoted

On Thu, Mar 08, Greg KH wrote:

quoted

quoted

hwinfo uses it to recognize a network interface (storage in case of
hv_storvsc).

That tool should report the real module name, not rely on the name to
match based on the structure, if you really care about the real module
name.

Its about the driver symlink, not the module name. 

I think most, if not all, (relevant) drivers use some sort of
KBUILD_MODNAME. Why should this driver be any different?

It shouldn't, I'm not saying that at all.  I'm only objecting to the
insistance that this be something worthwhile for 3.3-final which should
be out any day now.

Since its an API we could settle on the new string, which is already in
distro use.

I don't really understand, but in the end, it's not up to me, it's up to
the network maintainer when he wants to merge this to Linus.  If it were
up to me, I would wait for 3.4.

good luck,

greg k-h