RFC: Wireless Extensions for rndis_host

PATCH In review Bjorge Dijkstra 6 participants 23 messages Dec 22, 2007 – Jan 17, 2008 8 messages not in the archive

Subsystems (5)

Networking Network drivers USB networking drivers USB cdc Ethernet driver +1 more subsystems

lore

Share & tools

lore

Bjorge Dijkstra

author

2007-12-22 21:51 UTC #0 lore raw

cover RFC: Wireless Extensions for rndis_host

Hello all,

I have here a patchset that needs some review. This patchset adds wireless extensions for rndis_host to enable support for RNDIS based USB wireless LAN adapters (e.g. Linksys WUSB54GS, Belkin F05D7051).

In this patchset:

 1.  Fix sparse warning: returning void valued expression
 2.  Hardwire CDC descriptors when missing
 3.  Use 1KB buffer in rndis_unbind
 4.  Halt device if rndis_bind fails
 5.  Fix rndis packet filter flags
 6.  Split up rndis_host.c
 7.  Add Wireless Extensions for rndis_host
 8.  Use wlan device name for RNDIS wireless devices

The first five patches are more generic fixes that I think can be applied as they are. Of these, patch 2 is required to get these wireless devices working.
Patches 6-8 introduce some larger changes that may need a closer look.

All these patches should be applied in order. The entire series should apply cleanly to current linux kernel and net-2.6.25 trees.

regards,
Bjorge Dijkstra

A message that is not in the archive (sent off-list or not yet received): [email protected]

Bjorge Dijkstra

2007-12-22 21:51 UTC #2 lore raw

PATCH 1/8 Fix sparse warning: returning void-valued expression

rndis_unbind and usbnet_cdc_unbind don't return anything.

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/rndis_host.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_host.c +1 -1

index 1ebe325..96ef6a9 100644
@@ -585,7 +585,7 @@ static void rndis_unbind(struct usbnet *dev, struct usb_interface *intf)
 		kfree(halt);
 	}
 
-	return usbnet_cdc_unbind(dev, intf);
+	usbnet_cdc_unbind(dev, intf);
 }
 
 /*

Signature · 1 line

1.5.2.5

A message that is not in the archive (sent off-list or not yet received): [email protected]

David Brownell

Acked-by

2007-12-23 00:42 UTC #4 lore raw

Re: [PATCH 1/8] Fix sparse warning: returning void-valued expression

Quoted 10 lines

From [email protected] Sat Dec 22 13:52:53 2007
From: Bjorge Dijkstra <[email protected]>
To: [email protected]
Cc: [email protected], [email protected]
Subject: [PATCH 1/8] Fix sparse warning: returning void-valued expression
Date: Sat, 22 Dec 2007 22:51:27 +0100
rndis_unbind and usbnet_cdc_unbind don't return anything.
Signed-off-by Bjorge Dijkstra <[email protected]>

Acked-by David Brownell <[email protected]>

Quoted 20 lines

---
 drivers/net/usb/rndis_host.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_host.c +1 -1

index 1ebe325..96ef6a9 100644
@@ -585,7 +585,7 @@ static void rndis_unbind(struct usbnet *dev, struct usb_interface *intf)
 		kfree(halt);
 	}
 
-	return usbnet_cdc_unbind(dev, intf);
+	usbnet_cdc_unbind(dev, intf);
 }
 
 /*

Signature · 1 line

1.5.2.5

Bjorge Dijkstra

2007-12-22 21:51 UTC #5 lore raw

PATCH 4/8 Halt device if rndis_bind fails.

In reply to a message not in the archive

From: Jussi Kivilinna <[email protected]>

When bind fails after device was initialized, shutdown device properly by sending RNDIS_MSG_HALT.

Signed-off-by Jussi Kivilinna <[email protected]>

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/rndis_host.c |   12 +++++++++---
 1 files changed, 9 insertions(+), 3 deletions(-)

drivers/net/usb/rndis_host.c +9 -3

index 42b161c..c686025 100644
@@ -467,6 +467,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 		struct rndis_query_c	*get_c;
 		struct rndis_set	*set;
 		struct rndis_set_c	*set_c;
+		struct rndis_halt	*halt;
 	} u;
 	u32			tmp;
 	int			reply_len;

@@ -517,7 +518,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 				"dev can't take %u byte packets (max %u)\n",
 				dev->hard_mtu, tmp);
 			retval = -EINVAL;
-			goto fail_and_release;
+			goto halt_fail_and_release;
 		}
 		dev->hard_mtu = tmp;
 		net->mtu = dev->hard_mtu - net->hard_header_len;

@@ -539,7 +540,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 			48, (void **) &bp, &reply_len);
 	if (unlikely(retval< 0)) {
 		dev_err(&intf->dev, "rndis get ethaddr, %d\n", retval);
-		goto fail_and_release;
+		goto halt_fail_and_release;
 	}
 	memcpy(net->dev_addr, bp, ETH_ALEN);

@@ -555,7 +556,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 	retval = rndis_command(dev, u.header);
 	if (unlikely(retval < 0)) {
 		dev_err(&intf->dev, "rndis set packet filter, %d\n", retval);
-		goto fail_and_release;
+		goto halt_fail_and_release;
 	}
 
 	retval = 0;

@@ -563,6 +564,11 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 	kfree(u.buf);
 	return retval;
 
+halt_fail_and_release:
+	memset(u.halt, 0, sizeof *u.halt);
+	u.halt->msg_type = RNDIS_MSG_HALT;
+	u.halt->msg_len = ccpu2(sizeof *u.halt);
+	(void) rndis_command(dev, (void *)u.halt);
 fail_and_release:
 	usb_set_intfdata(info->data, NULL);
 	usb_driver_release_interface(driver_of(intf), info->data);

Signature · 1 line

1.5.2.5

David Brownell

Acked-by on v1 4/8

2007-12-23 01:00 UTC #6 lore raw

Re: [PATCH 4/8] [PATCH] Halt device if rndis_bind fails.

Quoted 8 lines

From [email protected] Sat Dec 22 13:53:07 2007
From: Bjorge Dijkstra <[email protected]>
To: [email protected]
Cc: [email protected], [email protected],
Jussi Kivilinna <[email protected]>
Subject: [PATCH 4/8] [PATCH] Halt device if rndis_bind fails.
Date: Sat, 22 Dec 2007 22:51:30 +0100

Quoted 7 lines

From: Jussi Kivilinna <[email protected]>
When bind fails after device was initialized, shutdown device properly by sending RNDIS_MSG_HALT.
Signed-off-by Jussi Kivilinna <[email protected]>
Signed-off-by Bjorge Dijkstra <[email protected]>

Acked-by David Brownell <[email protected]>

Quoted 58 lines

---
 drivers/net/usb/rndis_host.c |   12 +++++++++---
 1 files changed, 9 insertions(+), 3 deletions(-)

drivers/net/usb/rndis_host.c +9 -3

index 42b161c..c686025 100644
@@ -467,6 +467,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 		struct rndis_query_c	*get_c;
 		struct rndis_set	*set;
 		struct rndis_set_c	*set_c;
+		struct rndis_halt	*halt;
 	} u;
 	u32			tmp;
 	int			reply_len;

@@ -517,7 +518,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 				"dev can't take %u byte packets (max %u)\n",
 				dev->hard_mtu, tmp);
 			retval = -EINVAL;
-			goto fail_and_release;
+			goto halt_fail_and_release;
 		}
 		dev->hard_mtu = tmp;
 		net->mtu = dev->hard_mtu - net->hard_header_len;

@@ -539,7 +540,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 			48, (void **) &bp, &reply_len);
 	if (unlikely(retval< 0)) {
 		dev_err(&intf->dev, "rndis get ethaddr, %d\n", retval);
-		goto fail_and_release;
+		goto halt_fail_and_release;
 	}
 	memcpy(net->dev_addr, bp, ETH_ALEN);

@@ -555,7 +556,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 	retval = rndis_command(dev, u.header);
 	if (unlikely(retval < 0)) {
 		dev_err(&intf->dev, "rndis set packet filter, %d\n", retval);
-		goto fail_and_release;
+		goto halt_fail_and_release;
 	}
 
 	retval = 0;

@@ -563,6 +564,11 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 	kfree(u.buf);
 	return retval;
 
+halt_fail_and_release:
+	memset(u.halt, 0, sizeof *u.halt);
+	u.halt->msg_type = RNDIS_MSG_HALT;
+	u.halt->msg_len = ccpu2(sizeof *u.halt);
+	(void) rndis_command(dev, (void *)u.halt);
 fail_and_release:
 	usb_set_intfdata(info->data, NULL);
 	usb_driver_release_interface(driver_of(intf), info->data);

Signature · 1 line

1.5.2.5

Bjorge Dijkstra

2007-12-22 21:51 UTC #7 lore raw

PATCH 5/8 Fix rndis packet filter flags.

In reply to a message not in the archive

From: Jussi Kivilinna <[email protected]>

RNDIS packet filter flags are not exactly the same as CDC flags so we cannot reuse them.

Signed-off-by Jussi Kivilinna <[email protected]>

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/rndis_host.c |   23 ++++++++++++++++++++++-
 1 files changed, 22 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_host.c +22 -1

index c686025..3c116f9 100644
@@ -256,6 +256,27 @@ struct rndis_keepalive_c {	/* IN (optionally OUT) */
 #define OID_GEN_MAXIMUM_FRAME_SIZE	ccpu2(0x00010106)
 #define OID_GEN_CURRENT_PACKET_FILTER	ccpu2(0x0001010e)
 
+/* packet filter bits used by OID_GEN_CURRENT_PACKET_FILTER */
+#define RNDIS_PACKET_TYPE_DIRECTED		ccpu2(0x00000001)
+#define RNDIS_PACKET_TYPE_MULTICAST		ccpu2(0x00000002)
+#define RNDIS_PACKET_TYPE_ALL_MULTICAST		ccpu2(0x00000004)
+#define RNDIS_PACKET_TYPE_BROADCAST		ccpu2(0x00000008)
+#define RNDIS_PACKET_TYPE_SOURCE_ROUTING	ccpu2(0x00000010)
+#define RNDIS_PACKET_TYPE_PROMISCUOUS		ccpu2(0x00000020)
+#define RNDIS_PACKET_TYPE_SMT			ccpu2(0x00000040)
+#define RNDIS_PACKET_TYPE_ALL_LOCAL		ccpu2(0x00000080)
+#define RNDIS_PACKET_TYPE_GROUP			ccpu2(0x00001000)
+#define RNDIS_PACKET_TYPE_ALL_FUNCTIONAL	ccpu2(0x00002000)
+#define RNDIS_PACKET_TYPE_FUNCTIONAL		ccpu2(0x00004000)
+#define RNDIS_PACKET_TYPE_MAC_FRAME		ccpu2(0x00008000)
+
+/* default filter used with RNDIS devices */
+#define RNDIS_DEFAULT_FILTER ( \
+	RNDIS_PACKET_TYPE_DIRECTED | \
+	RNDIS_PACKET_TYPE_BROADCAST | \
+	RNDIS_PACKET_TYPE_ALL_MULTICAST | \
+	RNDIS_PACKET_TYPE_PROMISCUOUS )
+
 /*
  * RNDIS notifications from device: command completion; "reverse"
  * keepalives; etc

@@ -551,7 +572,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 	u.set->oid = OID_GEN_CURRENT_PACKET_FILTER;
 	u.set->len = ccpu2(4);
 	u.set->offset = ccpu2((sizeof *u.set) - 8);
-	*(__le32 *)(u.buf + sizeof *u.set) = ccpu2(DEFAULT_FILTER);
+	*(__le32 *)(u.buf + sizeof *u.set) = RNDIS_DEFAULT_FILTER;
 
 	retval = rndis_command(dev, u.header);
 	if (unlikely(retval < 0)) {

Signature · 1 line

1.5.2.5

David Brownell

Acked-by on v1 5/8

2007-12-23 01:06 UTC #8 lore raw

Re: [PATCH 5/8] Fix rndis packet filter flags.

Quoted 8 lines

From [email protected] Sat Dec 22 13:53:12 2007
From: Bjorge Dijkstra <[email protected]>
To: [email protected]
Cc: [email protected], [email protected],
Jussi Kivilinna <[email protected]>
Subject: [PATCH 5/8] Fix rndis packet filter flags.
Date: Sat, 22 Dec 2007 22:51:31 +0100

Quoted 7 lines

From: Jussi Kivilinna <[email protected]>
RNDIS packet filter flags are not exactly the same as CDC flags so we cannot reuse them.
Signed-off-by Jussi Kivilinna <[email protected]>
Signed-off-by Bjorge Dijkstra <[email protected]>

Acked-by David Brownell <[email protected]>

Hmm, the list seems to have grown since I last looked at it. Or maybe it's just that the MSFT documentation was so hard to make sense of ... I never could figure out what the RNDIS power management messages were supposed to be doing, or what the extra seemingly-undocumented messages were there for.

Quoted 48 lines

---
 drivers/net/usb/rndis_host.c |   23 ++++++++++++++++++++++-
 1 files changed, 22 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_host.c +22 -1

index c686025..3c116f9 100644
@@ -256,6 +256,27 @@ struct rndis_keepalive_c {	/* IN (optionally OUT) */
 #define OID_GEN_MAXIMUM_FRAME_SIZE	ccpu2(0x00010106)
 #define OID_GEN_CURRENT_PACKET_FILTER	ccpu2(0x0001010e)
 
+/* packet filter bits used by OID_GEN_CURRENT_PACKET_FILTER */
+#define RNDIS_PACKET_TYPE_DIRECTED		ccpu2(0x00000001)
+#define RNDIS_PACKET_TYPE_MULTICAST		ccpu2(0x00000002)
+#define RNDIS_PACKET_TYPE_ALL_MULTICAST		ccpu2(0x00000004)
+#define RNDIS_PACKET_TYPE_BROADCAST		ccpu2(0x00000008)
+#define RNDIS_PACKET_TYPE_SOURCE_ROUTING	ccpu2(0x00000010)
+#define RNDIS_PACKET_TYPE_PROMISCUOUS		ccpu2(0x00000020)
+#define RNDIS_PACKET_TYPE_SMT			ccpu2(0x00000040)
+#define RNDIS_PACKET_TYPE_ALL_LOCAL		ccpu2(0x00000080)
+#define RNDIS_PACKET_TYPE_GROUP			ccpu2(0x00001000)
+#define RNDIS_PACKET_TYPE_ALL_FUNCTIONAL	ccpu2(0x00002000)
+#define RNDIS_PACKET_TYPE_FUNCTIONAL		ccpu2(0x00004000)
+#define RNDIS_PACKET_TYPE_MAC_FRAME		ccpu2(0x00008000)
+
+/* default filter used with RNDIS devices */
+#define RNDIS_DEFAULT_FILTER ( \
+	RNDIS_PACKET_TYPE_DIRECTED | \
+	RNDIS_PACKET_TYPE_BROADCAST | \
+	RNDIS_PACKET_TYPE_ALL_MULTICAST | \
+	RNDIS_PACKET_TYPE_PROMISCUOUS )
+
 /*
  * RNDIS notifications from device: command completion; "reverse"
  * keepalives; etc

@@ -551,7 +572,7 @@ static int rndis_bind(struct usbnet *dev, struct usb_interface *intf)
 	u.set->oid = OID_GEN_CURRENT_PACKET_FILTER;
 	u.set->len = ccpu2(4);
 	u.set->offset = ccpu2((sizeof *u.set) - 8);
-	*(__le32 *)(u.buf + sizeof *u.set) = ccpu2(DEFAULT_FILTER);
+	*(__le32 *)(u.buf + sizeof *u.set) = RNDIS_DEFAULT_FILTER;
 
 	retval = rndis_command(dev, u.header);
 	if (unlikely(retval < 0)) {

Signature · 1 line

1.5.2.5

Bjorge Dijkstra

2007-12-22 21:51 UTC #9 lore raw

PATCH 8/8 Use wlan device name for RNDIS wireless devices

In reply to a message not in the archive

From: Jussi Kivilinna <[email protected]>

Use wlan device name for RNDIS wireless devices.

Signed-off-by Jussi Kivilinna <[email protected]>

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/rndis_wext.c |    2 +-
 drivers/net/usb/usbnet.c     |    3 +++
 drivers/net/usb/usbnet.h     |    2 ++
 3 files changed, 6 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_wext.c +1 -1

index a9ce944..1c28b2a 100644
@@ -2166,7 +2166,7 @@ static int rndis_wext_reset(struct usbnet *dev)
 
 struct driver_info rndis_wext_info = {
 	.description =	"Wireless RNDIS device",
-	.flags =	FLAG_ETHER | FLAG_FRAMING_RN | FLAG_NO_SETINT,
+	.flags =	FLAG_WLAN | FLAG_FRAMING_RN | FLAG_NO_SETINT,
 	.bind =		rndis_wext_bind,
 	.unbind =	rndis_wext_unbind,
 	.status =	rndis_status,

drivers/net/usb/usbnet.c +3 -0

index 8ed1fc5..a2a2d5e 100644
@@ -1204,6 +1204,9 @@ usbnet_probe (struct usb_interface *udev, const struct usb_device_id *prod)
 		if ((dev->driver_info->flags & FLAG_ETHER) != 0
 				&& (net->dev_addr [0] & 0x02) == 0)
 			strcpy (net->name, "eth%d");
+		/* WLAN devices should always be named "wlan%d" */
+		if ((dev->driver_info->flags & FLAG_WLAN) != 0)
+			strcpy(net->name, "wlan%d");
 
 		/* maybe the remote can't receive an Ethernet MTU */
 		if (net->mtu > (dev->hard_mtu - net->hard_header_len))

drivers/net/usb/usbnet.h +2 -0

index 83860a0..5c98ddc 100644
@@ -88,6 +88,8 @@ struct driver_info {
 #define FLAG_ETHER	0x0020		/* maybe use "eth%d" names */
 
 #define FLAG_FRAMING_AX 0x0040		/* AX88772/178 packets */
+#define FLAG_WLAN	0x0080		/* use "wlan%d" names */
+
 
 	/* init device ... can sleep, or cause probe() failure */
 	int	(*bind)(struct usbnet *, struct usb_interface *);

Signature · 1 line

1.5.2.5

A message that is not in the archive (sent off-list or not yet received): [email protected]

David Brownell

Acked-by

2007-12-23 01:30 UTC #11 lore raw

Re: [PATCH 8/8] [PATCH] Use wlan device name for RNDIS wireless devices

Quoted 10 lines

From: Bjorge Dijkstra <[email protected]>
Subject: [PATCH 8/8] [PATCH] Use wlan device name for RNDIS wireless devices
Date: Sat, 22 Dec 2007 22:51:34 +0100
From: Jussi Kivilinna <[email protected]>
Use wlan device name for RNDIS wireless devices.
Signed-off-by Jussi Kivilinna <[email protected]>
Signed-off-by Bjorge Dijkstra <[email protected]>

Acked-by David Brownell <[email protected]>

... though it'd be a bit nicer IMO to see this before patch #7, and just have one (or two) patches that only add infrastructure to the usbnet core code, before the rndis_wext patch which uses that new infrastructure.

Quoted 49 lines

---
 drivers/net/usb/rndis_wext.c |    2 +-
 drivers/net/usb/usbnet.c     |    3 +++
 drivers/net/usb/usbnet.h     |    2 ++
 3 files changed, 6 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_wext.c +1 -1

index a9ce944..1c28b2a 100644
@@ -2166,7 +2166,7 @@ static int rndis_wext_reset(struct usbnet *dev)
 
 struct driver_info rndis_wext_info = {
 	.description =	"Wireless RNDIS device",
-	.flags =	FLAG_ETHER | FLAG_FRAMING_RN | FLAG_NO_SETINT,
+	.flags =	FLAG_WLAN | FLAG_FRAMING_RN | FLAG_NO_SETINT,
 	.bind =		rndis_wext_bind,
 	.unbind =	rndis_wext_unbind,
 	.status =	rndis_status,

drivers/net/usb/usbnet.c +3 -0

index 8ed1fc5..a2a2d5e 100644
@@ -1204,6 +1204,9 @@ usbnet_probe (struct usb_interface *udev, const struct usb_device_id *prod)
 		if ((dev->driver_info->flags & FLAG_ETHER) != 0
 				&& (net->dev_addr [0] & 0x02) == 0)
 			strcpy (net->name, "eth%d");
+		/* WLAN devices should always be named "wlan%d" */
+		if ((dev->driver_info->flags & FLAG_WLAN) != 0)
+			strcpy(net->name, "wlan%d");
 
 		/* maybe the remote can't receive an Ethernet MTU */
 		if (net->mtu > (dev->hard_mtu - net->hard_header_len))

drivers/net/usb/usbnet.h +2 -0

index 83860a0..5c98ddc 100644
@@ -88,6 +88,8 @@ struct driver_info {
 #define FLAG_ETHER	0x0020		/* maybe use "eth%d" names */
 
 #define FLAG_FRAMING_AX 0x0040		/* AX88772/178 packets */
+#define FLAG_WLAN	0x0080		/* use "wlan%d" names */
+
 
 	/* init device ... can sleep, or cause probe() failure */
 	int	(*bind)(struct usbnet *, struct usb_interface *);

Signature · 1 line

1.5.2.5

David Brownell

2007-12-23 01:49 UTC #12 lore raw

In reply to a message not in the archive

Quoted 10 lines

From: Bjorge Dijkstra <[email protected]>
Subject: [PATCH 0/8] RFC: Wireless Extensions for rndis_host
Date: Sat, 22 Dec 2007 22:51:26 +0100
Hello all,
I have here a patchset that needs some review. This patchset adds wireless extensions for rndis_host to enable support for RNDIS based USB wireless LAN adapters (e.g. Linksys WUSB54GS, Belkin F05D7051).

Cool! We always like to see support for more USB peripherals. Even (especially?) if they'e previously been Windows-only, penguin-deprived hardware. ;)

Is all that stuff adequately documented? What I recall from working with RNDIS before is that docs were lacking in some critical details, and were sometimes wrong.

Quoted 10 lines

In this patchset:

 1.  Fix sparse warning: returning void valued expression
 2.  Hardwire CDC descriptors when missing
 3.  Use 1KB buffer in rndis_unbind
 4.  Halt device if rndis_bind fails
 5.  Fix rndis packet filter flags
 6.  Split up rndis_host.c
 7.  Add Wireless Extensions for rndis_host
 8.  Use wlan device name for RNDIS wireless devices

Quoted 3 lines

The first five patches are more generic fixes that I think can be applied as they are. Of these, patch 2 is required to get these wireless devices working.

I had no problem with those first five. I take it none of those should be prioritized for 2.6.24 integration?

Quoted 2 lines

Patches 6-8 introduce some larger changes that may need a closer look.

And you saw my feedback on those. Minimally, please split out the usbnet core extensions, instead of combining them with rndis_wext patches ... and on first princples I'd think rndis_wext should live in its own module.

I just skimmed patch #7. There are people far more aware of the issues with Linux wireless drivers than me. :)

- Dave

Quoted 7 lines

All these patches should be applied in order. The entire series should apply cleanly to current linux kernel and net-2.6.25 trees.
regards,
Bjorge Dijkstra

Bjorge Dijkstra

author

2007-12-22 21:51 UTC #13 lore raw

PATCH 2/8 Hardwire CDC descriptors when missing

In reply to Bjorge Dijkstra

Just as ActiveSync devices, some regular RNDIS devices also lack the CDC descriptors (e.g. devices based on BCM4320 WLAN chip). This patch hardwires the CDC descriptors for all RNDIS style devices when they are missing.

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/cdc_ether.c |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

drivers/net/usb/cdc_ether.c +5 -5

index a42acc3..97c17bb 100644
@@ -228,15 +228,16 @@ next_desc:
 		buf += buf [0];
 	}
 
-	/* Microsoft ActiveSync based RNDIS devices lack the CDC descriptors,
-	 * so we'll hard-wire the interfaces and not check for descriptors.
+	/* Microsoft ActiveSync based and some regular RNDIS devices lack the
+	 * CDC descriptors, so we'll hard-wire the interfaces and not check
+	 * for descriptors.
 	 */
-	if (is_activesync(&intf->cur_altsetting->desc) && !info->u) {
+	if (rndis && !info->u) {
 		info->control = usb_ifnum_to_if(dev->udev, 0);
 		info->data = usb_ifnum_to_if(dev->udev, 1);
 		if (!info->control || !info->data) {
 			dev_dbg(&intf->dev,
-				"activesync: master #0/%p slave #1/%p\n",
+				"rndis: master #0/%p slave #1/%p\n",
 				info->control,
 				info->data);
 			goto bad_desc;

@@ -316,7 +317,6 @@ void usbnet_cdc_unbind(struct usbnet *dev, struct usb_interface *intf)
 }
 EXPORT_SYMBOL_GPL(usbnet_cdc_unbind);
 
-
 /*-------------------------------------------------------------------------
  *
  * Communications Device Class, Ethernet Control model

Signature · 1 line

1.5.2.5

David Brownell

Acked-by on v1 2/8

2007-12-23 00:49 UTC #14 lore raw

Re: [PATCH 2/8] [PATCH] Hardwire CDC descriptors when missing

Quoted 8 lines

From [email protected] Sat Dec 22 13:53:00 2007
X-SourceIP: 213.93.131.145
From: Bjorge Dijkstra <[email protected]>
To: [email protected]
Cc: [email protected], [email protected]
Subject: [PATCH 2/8] [PATCH] Hardwire CDC descriptors when missing
Date: Sat, 22 Dec 2007 22:51:28 +0100

Quoted 6 lines

Just as ActiveSync devices, some regular RNDIS devices also lack the CDC descriptors (e.g. devices based on BCM4320 WLAN chip). This patch hardwires the CDC descriptors for all RNDIS style devices when they are missing.
Signed-off-by Bjorge Dijkstra <[email protected]>

Acked-by David Brownell <[email protected]>

Note what this tells us about how little Microsoft cares about the interoperability specs they claim to conform to. Having those descriptors is *NOT* optional in their (incomplete and in adequate) RNDIS specification. (I've not seen ActiveSync specs, so that support is pure guesswork.)

Quoted 40 lines

---
 drivers/net/usb/cdc_ether.c |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

drivers/net/usb/cdc_ether.c +5 -5

index a42acc3..97c17bb 100644
@@ -228,15 +228,16 @@ next_desc:
 		buf += buf [0];
 	}
 
-	/* Microsoft ActiveSync based RNDIS devices lack the CDC descriptors,
-	 * so we'll hard-wire the interfaces and not check for descriptors.
+	/* Microsoft ActiveSync based and some regular RNDIS devices lack the
+	 * CDC descriptors, so we'll hard-wire the interfaces and not check
+	 * for descriptors.
 	 */
-	if (is_activesync(&intf->cur_altsetting->desc) && !info->u) {
+	if (rndis && !info->u) {
 		info->control = usb_ifnum_to_if(dev->udev, 0);
 		info->data = usb_ifnum_to_if(dev->udev, 1);
 		if (!info->control || !info->data) {
 			dev_dbg(&intf->dev,
-				"activesync: master #0/%p slave #1/%p\n",
+				"rndis: master #0/%p slave #1/%p\n",
 				info->control,
 				info->data);
 			goto bad_desc;

@@ -316,7 +317,6 @@ void usbnet_cdc_unbind(struct usbnet *dev, struct usb_interface *intf)
 }
 EXPORT_SYMBOL_GPL(usbnet_cdc_unbind);
 
-?
 /*-------------------------------------------------------------------------
  *
  * Communications Device Class, Ethernet Control model

Signature · 1 line

1.5.2.5

Bjorge Dijkstra

author

2007-12-22 21:51 UTC #15 lore raw

PATCH 3/8 Use 1KB buffer in rndis_unbind

In reply to Bjorge Dijkstra

From: Jussi Kivilinna <[email protected]>

rndis_command requires the caller to pass in a buffer of at least 1KB.

Signed-off-by Jussi Kivilinna <[email protected]>

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/rndis_host.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_host.c +1 -1

index 96ef6a9..42b161c 100644
@@ -577,7 +577,7 @@ static void rndis_unbind(struct usbnet *dev, struct usb_interface *intf)
 	struct rndis_halt	*halt;
 
 	/* try to clear any rndis state/activity (no i/o from stack!) */
-	halt = kzalloc(sizeof *halt, GFP_KERNEL);
+	halt = kzalloc(CONTROL_BUFFER_SIZE, GFP_KERNEL);
 	if (halt) {
 		halt->msg_type = RNDIS_MSG_HALT;
 		halt->msg_len = ccpu2(sizeof *halt);

Signature · 1 line

1.5.2.5

A message that is not in the archive (sent off-list or not yet received): [email protected]

David Brownell

Acked-by

2007-12-23 00:51 UTC #17 lore raw

Re: [PATCH 3/8] [PATCH] Use 1KB buffer in rndis_unbind

Quoted 8 lines

From [email protected] Sat Dec 22 13:53:03 2007
From: Bjorge Dijkstra <[email protected]>
To: [email protected]
Cc: [email protected], [email protected],
Jussi Kivilinna <[email protected]>
Subject: [PATCH 3/8] [PATCH] Use 1KB buffer in rndis_unbind
Date: Sat, 22 Dec 2007 22:51:29 +0100

Quoted 6 lines

From: Jussi Kivilinna <[email protected]>
rndis_command requires the caller to pass in a buffer of at least 1KB.
Signed-off-by Jussi Kivilinna <[email protected]>
Signed-off-by Bjorge Dijkstra <[email protected]>

Acked-by David Brownell <[email protected]>

Quoted 20 lines

---
 drivers/net/usb/rndis_host.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

drivers/net/usb/rndis_host.c +1 -1

index 96ef6a9..42b161c 100644
@@ -577,7 +577,7 @@ static void rndis_unbind(struct usbnet *dev, struct usb_interface *intf)
 	struct rndis_halt	*halt;
 
 	/* try to clear any rndis state/activity (no i/o from stack!) */
-	halt = kzalloc(sizeof *halt, GFP_KERNEL);
+	halt = kzalloc(CONTROL_BUFFER_SIZE, GFP_KERNEL);
 	if (halt) {
 		halt->msg_type = RNDIS_MSG_HALT;
 		halt->msg_len = ccpu2(sizeof *halt);

Signature · 1 line

1.5.2.5

Bjorge Dijkstra

author

2007-12-22 21:51 UTC #18 lore raw

PATCH 6/8 Split up rndis_host.c

In reply to Bjorge Dijkstra

Split up rndis_host.c into rndis_host.h and rndis_base.c and change Makefile accordingly. This is done so we can add extra source files to the rndis_host module later on.

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/Makefile     |    1 +
 drivers/net/usb/rndis_base.c |  548 ++++++++++++++++++++++++++++++
 drivers/net/usb/rndis_host.c |  763 ------------------------------------------
 drivers/net/usb/rndis_host.h |  256 ++++++++++++++
 4 files changed, 805 insertions(+), 763 deletions(-)
 create mode 100644 drivers/net/usb/rndis_base.c
 delete mode 100644 drivers/net/usb/rndis_host.c
 create mode 100644 drivers/net/usb/rndis_host.h

drivers/net/usb/Makefile +1 -0

index 595a539..611ab62 100644
@@ -13,6 +13,7 @@ obj-$(CONFIG_USB_NET_GL620A)	+= gl620a.o
 obj-$(CONFIG_USB_NET_NET1080)	+= net1080.o
 obj-$(CONFIG_USB_NET_PLUSB)	+= plusb.o
 obj-$(CONFIG_USB_NET_RNDIS_HOST)	+= rndis_host.o
+rndis_host-objs			+= rndis_base.o
 obj-$(CONFIG_USB_NET_CDC_SUBSET)	+= cdc_subset.o
 obj-$(CONFIG_USB_NET_ZAURUS)	+= zaurus.o
 obj-$(CONFIG_USB_NET_MCS7830)	+= mcs7830.o

drivers/net/usb/rndis_base.c new file +548 -0

548 lines not on this page · raw message

drivers/net/usb/rndis_host.c deleted +0 ~~-763~~

763 lines not on this page · raw message

drivers/net/usb/rndis_host.h new file +256 -0

256 lines not on this page · raw message

Signature · 1 line

1.5.2.5

David Brownell

2007-12-23 01:17 UTC #19 lore raw

Re: [PATCH 6/8] [PATCH] Split up rndis_host.c

Quoted 7 lines

From: Bjorge Dijkstra <[email protected]>
Subject: [PATCH 6/8] [PATCH] Split up rndis_host.c
Date: Sat, 22 Dec 2007 22:51:32 +0100
Split up rndis_host.c into rndis_host.h and rndis_base.c and change Makefile accordingly. This is done so we can add extra source files to the rndis_host module later on.

I'm fine with splitting out a header file and the EXPORT_SYMBOL_GPL. But why not just have a separate "rndis_wext" module?

Quoted 9 lines

---
 drivers/net/usb/Makefile     |    1 +
 drivers/net/usb/rndis_base.c |  548 ++++++++++++++++++++++++++++++
 drivers/net/usb/rndis_host.c |  763 ------------------------------------------
 drivers/net/usb/rndis_host.h |  256 ++++++++++++++
 4 files changed, 805 insertions(+), 763 deletions(-)
 create mode 100644 drivers/net/usb/rndis_base.c
 delete mode 100644 drivers/net/usb/rndis_host.c
 create mode 100644 drivers/net/usb/rndis_host.h

A message that is not in the archive (sent off-list or not yet received): 20071223011707.1798C2360C6-ZcXrCSuhvln6VZ3dlLfH/[email protected]

Jussi Kivilinna

2008-01-02 20:13 UTC #21 lore raw

Re: [PATCH 6/8] [PATCH] Split up rndis_host.c

Hello,

Bjorge was not comfortable with double probing rndis_wext requires, wireless RNDIS devices have same device id as the rest of RNDIS. Our module version checks OID_GEN_PHYSICAL_MEDIUM in generic_rndis_bind, with rndis_host bind fails if OID is supported and wireless media type is returned, with rndis_wext if OID isn't supported or type isn't wireless. Should this be ok?

Should separate rndis_wext be located in drivers/net/wireless instead of drivers/net/usb?

- Jussi Kivilinna

Quoted 26 lines · David Brownell

Quoted 7 lines
From: Bjorge Dijkstra <[email protected]>
Subject: [PATCH 6/8] [PATCH] Split up rndis_host.c
Date: Sat, 22 Dec 2007 22:51:32 +0100
Split up rndis_host.c into rndis_host.h and rndis_base.c and change Makefile accordingly. This is done so we can add extra source files to the rndis_host module later on.
I'm fine with splitting out a header file and the EXPORT_SYMBOL_GPL. But why not just have a separate "rndis_wext" module?
Quoted 9 lines
---
 drivers/net/usb/Makefile     |    1 +
 drivers/net/usb/rndis_base.c |  548 ++++++++++++++++++++++++++++++
 drivers/net/usb/rndis_host.c |  763 ------------------------------------------
 drivers/net/usb/rndis_host.h |  256 ++++++++++++++
 4 files changed, 805 insertions(+), 763 deletions(-)
 create mode 100644 drivers/net/usb/rndis_base.c
 delete mode 100644 drivers/net/usb/rndis_host.c
 create mode 100644 drivers/net/usb/rndis_host.h
-
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

David Brownell

2008-01-08 11:19 UTC #22 lore raw

Re: [PATCH 6/8] [PATCH] Split up rndis_host.c

Quoted 4 lines · Jussi Kivilinna

Hello,
Bjorge was not comfortable with double probing rndis_wext requires, wireless RNDIS devices have same device id as the rest of RNDIS.

I should have known Microsoft wouldn't start by assuming systems software module boundaries should be natural and obvious ones. :(

I see that the rndis_wext.c Kconfig won't kick in unless the 802.11 stuff is available ... what additional dependencies does that imply for a fatter rndis_host module? If it doesn't cause extra modules to be loaded (even for non-WLAN devices), then I suppose I could just live with the RNDIS module being bigger than it needs to be. It was pretty thin to start with!

What would bother me would be seeing that systems with this WLAN support option configured could no longer load the basic RNDIS code without also loading a bunch of extra modules that weren't needed for all RNDIS devices.

Quoted 5 lines

Our
module version checks OID_GEN_PHYSICAL_MEDIUM in generic_rndis_bind, with rndis_host bind fails if OID is supported and wireless media type is returned, with rndis_wext if OID isn't supported or type isn't wireless. Should this be ok?

Sounds like if you can go the different-modules route, then rndis_bind() should maybe accept a parameter giving it that option. Then

	rndis_wlan_bind(...) == rndis_bind(..., check_wlan_media)
	rndis_generic_bind(...) == rndis_bind(..., non_wlan_media)

That would be getting pretty complicated, since as you say the same USB-level binding (hence hotplugging) would need to kick in ... but having different modules for different RNDIS flavors would need a new hotplug mechanism keying on media type. Ugh.

Quoted 2 lines

Should separate rndis_wext be located in drivers/net/wireless instead of drivers/net/usb?

If that doesn't complicate maintaining that software, I can't see why it shouldn't go there. It might imply moving the usbnet interface declarations to someplace linux/include though. (And that presumes there's actually a sane way to split that stuff into its own module, too...)

- Dave

Quoted 26 lines

- Jussi Kivilinna
Quoted 22 lines · David Brownell
Quoted 7 lines
From: Bjorge Dijkstra <[email protected]>
Subject: [PATCH 6/8] [PATCH] Split up rndis_host.c
Date: Sat, 22 Dec 2007 22:51:32 +0100
Split up rndis_host.c into rndis_host.h and rndis_base.c and change Makefile accordingly. This is done so we can add extra source files to the rndis_host module later on.
I'm fine with splitting out a header file and the EXPORT_SYMBOL_GPL. But why not just have a separate "rndis_wext" module?
Quoted 9 lines
---
 drivers/net/usb/Makefile     |    1 +
 drivers/net/usb/rndis_base.c |  548 ++++++++++++++++++++++++++++++
 drivers/net/usb/rndis_host.c |  763 ------------------------------------------
 drivers/net/usb/rndis_host.h |  256 ++++++++++++++
 4 files changed, 805 insertions(+), 763 deletions(-)
 create mode 100644 drivers/net/usb/rndis_base.c
 delete mode 100644 drivers/net/usb/rndis_host.c
 create mode 100644 drivers/net/usb/rndis_host.h

A message that is not in the archive (sent off-list or not yet received): 200801080319.23299.david-b-yBeKhBN/[email protected]

Johannes Berg

2008-01-08 17:32 UTC #24 lore raw

Re: [PATCH 6/8] [PATCH] Split up rndis_host.c

Quoted 4 lines

I see that the rndis_wext.c Kconfig won't kick in unless the 802.11 stuff is available ... what additional dependencies does that imply for a fatter rndis_host module?

No extra modules are currently required for just plain wext [1]. In the future, we hope to migrate stuff to cfg80211 though, which is a separate module.

johannes

[1] mostly because all the code is simply compiled into the kernel...

A message that is not in the archive (sent off-list or not yet received): [email protected]

David Brownell

2008-01-08 18:49 UTC #26 lore raw

Re: [PATCH 6/8] [PATCH] Split up rndis_host.c

Quoted 8 lines · Johannes Berg

Quoted 4 lines
I see that the rndis_wext.c Kconfig won't kick in unless the 802.11 stuff is available ... what additional dependencies does that imply for a fatter rndis_host module?
No extra modules are currently required for just plain wext [1]. In the future, we hope to migrate stuff to cfg80211 though, which is a separate module.

Hmmm. Well, go the current route then ... but come up with a good plan to split it up later, so that connecting a typical non-wireless RNDIS device won't incur such needless extra modules.

- Dave

Quoted 4 lines

johannes
[1] mostly because all the code is simply compiled into the kernel...

Jussi Kivilinna

2008-01-17 20:40 UTC #27 lore raw

Re: [PATCH 6/8] [PATCH] Split up rndis_host.c

In reply to a message not in the archive

Quoted 3 lines · David Brownell

Quoted 3 lines
Our
module version checks OID_GEN_PHYSICAL_MEDIUM in generic_rndis_bind, with rndis_host bind fails if OID is supported and wireless media

type

Quoted 4 lines

Quoted 2 lines
is returned, with rndis_wext if OID isn't supported or type isn't wireless. Should this be ok?
Sounds like if you can go the different-modules route, then

rndis_bind()

Quoted 9 lines

should maybe accept a parameter giving it that option. Then
      rndis_wlan_bind(...) == rndis_bind(..., check_wlan_media)
      rndis_generic_bind(...) == rndis_bind(..., non_wlan_media)
That would be getting pretty complicated, since as you say the same USB-level binding (hence hotplugging) would need to kick in ... but having different modules for different RNDIS flavors would need a new hotplug mechanism keying on media type. Ugh.

Different modules works with rndis_generic_bind(..., media type option). dmesg shows following:

[11214.496773] usb 3-3: new high speed USB device using ehci_hcd and
address 7
[11214.622861] usb 3-3: configuration #1 chosen from 1 choice
[11214.784874] usbcore: registered new interface driver cdc_ether
[11214.805700] rndis_host 3-3:1.0: driver requires wired physical
medium, but device is wireless.
[11214.806823] usbcore: registered new interface driver rndis_host
[11215.190824] wlan0: register 'rndis_wext' at usb-0000:00:10.3-3,
Wireless RNDIS device, 00:xx:xx:xx:xx:xx
[11215.190853] usbcore: registered new interface driver rndis_wext
[11215.398850] rndis_wext 3-3:1.0: rndis media disconnect

I assume that rndis_host gets always bind first as it's first in alphabetic, so rndis_wext is only loaded if RNDIS device has wireless media type.

I have new patchset that prepares rndis_host/usbnet for separate rndis_wext but not sure should I submit (I haven't managed to get into contact with Bjorge Dijkstra for 3,5 weeks now to have him decide how to proceed).

- Jussi Kivilinna

Bjorge Dijkstra

author

2007-12-22 21:51 UTC #28 lore raw

PATCH 7/8 Add Wireless Extensions to rndis_host

In reply to Bjorge Dijkstra

The bulk of this patch is the addition of a new file that implements the wireless extensions for RNDIS devices. The rest are some smaller changes to usbnet and rndis_host to hook the wireless extensions into them:
* a private data pointer is added to usbnet.
* a callback is added to driver_info to signal link state changes.
* a physical medium type check is added to rndis_bind to check for wireless lan devices and turn on the wireless extensions.
* and finally a Kconfig option to enable/disable this all.

Signed-off-by Jussi Kivilinna <[email protected]>

Signed-off-by Bjorge Dijkstra <[email protected]>

---
 drivers/net/usb/Kconfig      |    7 +
 drivers/net/usb/Makefile     |    4 +
 drivers/net/usb/rndis_base.c |   50 +-
 drivers/net/usb/rndis_host.h |   18 +
 drivers/net/usb/rndis_wext.c | 2177 ++++++++++++++++++++++++++++++++++++++++++
 drivers/net/usb/usbnet.h     |    4 +
 6 files changed, 2255 insertions(+), 5 deletions(-)
 create mode 100644 drivers/net/usb/rndis_wext.c

drivers/net/usb/Kconfig +7 -0

index a12c9c4..cc38294 100644
@@ -239,6 +239,13 @@ config USB_NET_RNDIS_HOST
 	  The protocol specification is incomplete, and is controlled by
 	  (and for) Microsoft; it isn't an "Open" ecosystem or market.
 
+config USB_NET_RNDIS_WEXT
+	boolean "Wireless Extensions for RNDIS host driver"
+	depends on USB_NET_RNDIS_HOST && WLAN_80211 && EXPERIMENTAL
+	select WIRELESS_EXT
+	help
+	  This option enables support for RNDIS based wireless LAN devices.
+
 config USB_NET_CDC_SUBSET
 	tristate "Simple USB Network Links (CDC Ethernet subset)"
 	depends on USB_USBNET

drivers/net/usb/Makefile +4 -0

index 611ab62..83c14f1 100644
@@ -14,6 +14,10 @@ obj-$(CONFIG_USB_NET_NET1080)	+= net1080.o
 obj-$(CONFIG_USB_NET_PLUSB)	+= plusb.o
 obj-$(CONFIG_USB_NET_RNDIS_HOST)	+= rndis_host.o
 rndis_host-objs			+= rndis_base.o
+ifeq ($(CONFIG_USB_NET_RNDIS_WEXT),y)
+rndis_host-objs += rndis_wext.o
+endif
+
 obj-$(CONFIG_USB_NET_CDC_SUBSET)	+= cdc_subset.o
 obj-$(CONFIG_USB_NET_ZAURUS)	+= zaurus.o
 obj-$(CONFIG_USB_NET_MCS7830)	+= mcs7830.o

drivers/net/usb/rndis_base.c +45 -5

68 lines not on this page · raw message

drivers/net/usb/rndis_host.h +18 -0

index 56de532..6427315 100644
@@ -82,6 +82,18 @@ struct rndis_msg_hdr {
 #define	RNDIS_STATUS_MEDIA_CONNECT	ccpu2(0x4001000b)
 #define	RNDIS_STATUS_MEDIA_DISCONNECT	ccpu2(0x4001000c)
 
+/* codes for OID_GEN_PHYSICAL_MEDIUM */
+#define RNDIS_PHYSICAL_MEDIUM_UNSPECIFIED       ccpu2(0x00000000)
+#define RNDIS_PHYSICAL_MEDIUM_WIRELESS_LAN      ccpu2(0x00000001)
+#define RNDIS_PHYSICAL_MEDIUM_CABLE_MODEM       ccpu2(0x00000002)
+#define RNDIS_PHYSICAL_MEDIUM_PHONE_LINE        ccpu2(0x00000003)
+#define RNDIS_PHYSICAL_MEDIUM_POWER_LINE        ccpu2(0x00000004)
+#define RNDIS_PHYSICAL_MEDIUM_DSL               ccpu2(0x00000005)
+#define RNDIS_PHYSICAL_MEDIUM_FIBRE_CHANNEL     ccpu2(0x00000006)
+#define RNDIS_PHYSICAL_MEDIUM_1394              ccpu2(0x00000007)
+#define RNDIS_PHYSICAL_MEDIUM_WIRELESS_WAN      ccpu2(0x00000008)
+#define RNDIS_PHYSICAL_MEDIUM_MAX               ccpu2(0x00000009)
+
 
 struct rndis_data_hdr {
 	__le32	msg_type;		/* RNDIS_MSG_PACKET */

@@ -222,6 +234,7 @@ struct rndis_keepalive_c {	/* IN (optionally OUT) */
 #define OID_802_3_PERMANENT_ADDRESS	ccpu2(0x01010101)
 #define OID_GEN_MAXIMUM_FRAME_SIZE	ccpu2(0x00010106)
 #define OID_GEN_CURRENT_PACKET_FILTER	ccpu2(0x0001010e)
+#define OID_GEN_PHYSICAL_MEDIUM		ccpu2(0x00010202)
 
 /* packet filter bits used by NDIS_OID_PACKET_FILTER */
 #define RNDIS_PACKET_TYPE_DIRECTED		ccpu2(0x00000001)

@@ -252,5 +265,10 @@ extern int rndis_rx_fixup(struct usbnet *dev, struct sk_buff *skb);
 extern struct sk_buff *
 rndis_tx_fixup(struct usbnet *dev, struct sk_buff *skb, gfp_t flags);
 
+#ifdef CONFIG_USB_NET_RNDIS_WEXT
+/* from rndis_wext.c */
+extern struct driver_info rndis_wext_info;
+#endif
+
 #endif	/* __RNDIS_HOST_H */

drivers/net/usb/rndis_wext.c new file +2177 -0

2,177 lines not on this page · raw message

drivers/net/usb/usbnet.h +4 -0

index 1fae434..83860a0 100644
@@ -31,6 +31,7 @@ struct usbnet {
 	struct usb_interface	*intf;
 	struct driver_info	*driver_info;
 	const char		*driver_name;
+	void			*driver_priv;
 	wait_queue_head_t	*wait;
 	struct mutex		phy_mutex;
 	unsigned char		suspend_count;

@@ -109,6 +110,9 @@ struct driver_info {
 	/* fixup rx packet (strip framing) */
 	int	(*rx_fixup)(struct usbnet *dev, struct sk_buff *skb);
 
+	/* called when link state changes */
+	void    (*link_change)(struct usbnet *dev, int state);
+
 	/* fixup tx packet (add framing) */
 	struct sk_buff	*(*tx_fixup)(struct usbnet *dev,
 				struct sk_buff *skb, gfp_t flags);

Signature · 1 line

1.5.2.5

A message that is not in the archive (sent off-list or not yet received): [email protected]

David Brownell

2007-12-23 01:27 UTC #30 lore raw

Re: [PATCH 7/8] Add Wireless Extensions to rndis_host

Quoted 14 lines

From: Bjorge Dijkstra <[email protected]>
Subject: [PATCH 7/8] Add Wireless Extensions to rndis_host
Date: Sat, 22 Dec 2007 22:51:33 +0100
The bulk of this patch is the addition of a new file that implements the wireless extensions for RNDIS devices. The rest are some smaller changes to usbnet and rndis_host to hook the wireless extensions into them:
* a private data pointer is added to usbnet.
* a callback is added to driver_info to signal link state changes.
* a physical medium type check is added to rndis_bind to check for wireless lan devices and turn on the wireless extensions.
* and finally a Kconfig option to enable/disable this all.

Quoted 2 lines

Signed-off-by Jussi Kivilinna <[email protected]>
Signed-off-by Bjorge Dijkstra <[email protected]>

I'm basically fine with this, but I'll hold off any ack until the issue with the preceding patch (6/8) is resolved. That would boil down to making this a standalone module, or knowing why that's not such a good idea.

I'd prefer to see the usbnet core extension packaged separately instead of buried in this patch. :)

And I think a MAINTAINERS update for rndis_wext would be appropriate. No way can I take care of that code!

(In fact, would you feel comfortable taking over rndis_host too?)

- Dave

Quoted 9 lines

---
 drivers/net/usb/Kconfig      |    7 +
 drivers/net/usb/Makefile     |    4 +
 drivers/net/usb/rndis_base.c |   50 +-
 drivers/net/usb/rndis_host.h |   18 +
 drivers/net/usb/rndis_wext.c | 2177 ++++++++++++++++++++++++++++++++++++++++++
 drivers/net/usb/usbnet.h     |    4 +
 6 files changed, 2255 insertions(+), 5 deletions(-)
 create mode 100644 drivers/net/usb/rndis_wext.c