Skip to contentSkip to conversation

b43: rfkill: use status register based on core revision not (PHY's)

PATCH New Rafał Miłecki 2 participants 6 messages Oct 22, 2010 – Oct 25, 2010 Networking drivers (wireless)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 b43: rfkill: use status register based on core revision not (PHY's)

---
This makes my N-PHY radio at least running (no scanning results so far). The same register is used by wl on this device.

---
 drivers/net/wireless/b43/rfkill.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/net/wireless/b43/rfkill.c +1 -1
index 78016ae..d2734c9 100644
@@ -28,7 +28,7 @@
 /* Returns TRUE, if the radio is enabled in hardware. */
 bool b43_is_hw_radio_enabled(struct b43_wldev *dev)
 {
-	if (dev->phy.rev >= 3 || dev->phy.type == B43_PHYTYPE_LP) {
+	if (dev->dev->id.revision >= 3) {
 		if (!(b43_read32(dev, B43_MMIO_RADIO_HWENABLED_HI)
 		      & B43_MMIO_RADIO_HWENABLED_HI_MASK))
 			return 1;
Signature · 1 line
1.7.1
#1 lore raw

PATCH 1/1 b43: rfkill: use status register based on core revision not (PHY's)

W dniu 22 pa?dziernika 2010 22:24 u?ytkownik Rafa? Mi?ecki <[email protected]> napisa?:

Quoted 4 lines

---
This makes my N-PHY radio at least running (no scanning results so far). The same register is used by wl on this device.

John, if you care you can move left opening brace in subject to look as following:
(not PHY's)

Signature · 1 line
Rafa?
#2 lore raw

PATCH 1/1 b43: rfkill: use status register based on core revision not (PHY's)

In reply to Rafał Miłecki
Quoted 13 lines · Rafał Miłecki

---
This makes my N-PHY radio at least running (no scanning results so far). The same register is used by wl on this device.

---
 drivers/net/wireless/b43/rfkill.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/net/wireless/b43/rfkill.c +0 -0
index 78016ae..d2734c9 100644

drivers/net/wireless/b43/rfkill.c line 34

Quoted 8 lines
drivers/net/wireless/b43/rfkill.c +1 -1
@@ -28,7 +28,7 @@
 /* Returns TRUE, if the radio is enabled in hardware. */
 bool b43_is_hw_radio_enabled(struct b43_wldev *dev)
 {
-	if (dev->phy.rev >= 3 || dev->phy.type == B43_PHYTYPE_LP) {
+	if (dev->dev->id.revision >= 3) {
 		if (!(b43_read32(dev, B43_MMIO_RADIO_HWENABLED_HI)
 		      & B43_MMIO_RADIO_HWENABLED_HI_MASK))
 			return 

As that register is present only when the core revision >= 3, this change should not cause a regression. Perhaps this was an error in the RE from the start and testing the phy.rev has always been wrong.

John: Just in case my analysis is wrong, keep this change out of 2.6.37. I'll give it some testing with G PHYs before 2.6.38 merge time.

Larry

#3 lore raw

PATCH 1/1 b43: rfkill: use status register based on core revision not (PHY's)

W dniu 22 pa?dziernika 2010 23:38 u?ytkownik Larry Finger <[email protected]> napisa?:

Quoted 18 lines · Rafa? Mi?ecki
Quoted 17 lines

---
This makes my N-PHY radio at least running (no scanning results so far). The same register is used by wl on this device.
---
?drivers/net/wireless/b43/rfkill.c | ? ?2 +-
?1 files changed, 1 insertions(+), 1 deletions(-)

drivers/net/wireless/b43/rfkill.c +0 -0
index 78016ae..d2734c9 100644
@@ -28,7 +28,7 @@

?/* Returns TRUE, if the radio is enabled in hardware. */ ?bool b43_is_hw_radio_enabled(struct b43_wldev *dev)
?{
- ? ? if (dev->phy.rev >= 3 || dev->phy.type == B43_PHYTYPE_LP) {

Quoted 8 lines
Quoted 4 lines

+ ? ? if (dev->dev->id.revision >= 3) {
? ? ? ? ? ? ? if (!(b43_read32(dev, B43_MMIO_RADIO_HWENABLED_HI) ? ? ? ? ? ? ? ? ? ? & B43_MMIO_RADIO_HWENABLED_HI_MASK)) ? ? ? ? ? ? ? ? ? ? ? return

As that register is present only when the core revision >= 3, this change should not cause a regression. Perhaps this was an error in the RE from the start and testing the phy.rev has always been wrong.

Do not blame RE team ;) It was described correctly since "ever", you can check it in history of:
http://bcm-specs.sipsolutions.net/MMIO

I actually suspect this may be source of self-switching-on-off-radio issue described in:
"b43legacy-phy3: Radio hardware status changed to XXX" thread.

I already mailed reporter to ask if he still has access to problematic card.

Signature · 1 line
Rafa?
#4 lore raw

PATCH 1/1 b43: rfkill: use status register based on core revision not (PHY's)

Quoted 12 lines · Rafał Miłecki

W dniu 22 października 2010 23:38 użytkownik Larry Finger

Quoted 5 lines · <[email protected]> napisał:

On 10/22/2010 03:24 PM, Rafał Miłecki wrote:

As that register is present only when the core revision >= 3, this change should not cause a regression. Perhaps this was an error in the RE from the start and testing the phy.rev has always been wrong.

Do not blame RE team ;) It was described correctly since "ever", you can check it in history of:
http://bcm-specs.sipsolutions.net/MMIO

Quoted 5 lines

I actually suspect this may be source of self-switching-on-off-radio issue described in:
"b43legacy-phy3: Radio hardware status changed to XXX" thread.

I already mailed reporter to ask if he still has access to problematic card.

I have two cards that use b43legacy, but their physical format (PCI and Cardbus) do not include an RFKILL switch, thus I have never tested this code with b43legacy.

Larry

#5 lore raw

PATCH 1/1 b43: rfkill: use status register based on core revision not (PHY's)

In reply to Rafał Miłecki
Quoted 13 lines · Rafał Miłecki

---
This makes my N-PHY radio at least running (no scanning results so far). The same register is used by wl on this device.

---
 drivers/net/wireless/b43/rfkill.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/net/wireless/b43/rfkill.c +0 -0
index 78016ae..d2734c9 100644

drivers/net/wireless/b43/rfkill.c line 34

Quoted 8 lines
drivers/net/wireless/b43/rfkill.c +1 -1
@@ -28,7 +28,7 @@
 /* Returns TRUE, if the radio is enabled in hardware. */
 bool b43_is_hw_radio_enabled(struct b43_wldev *dev)
 {
-	if (dev->phy.rev >= 3 || dev->phy.type == B43_PHYTYPE_LP) {
+	if (dev->dev->id.revision >= 3) {
 		if (!(b43_read32(dev, B43_MMIO_RADIO_HWENABLED_HI)
 		      & B43_MMIO_RADIO_HWENABLED_HI_MASK))
 			return 1;

Over the weekend, I was able to test this patch using all my b43 hardware. There were no problems.

Despite having no problems, I still NACK the patch for the reason that b43 only applies to devices with id.revision >= 5, thus the revised "if" will always be true. The patch should be rewritten as only b43legacy needs to test the 802.11 core revision.

Larry