2009-09-16 19:37:57

by Thomas Ilnseher

[permalink] [raw]
Subject: [PATCH 2/2] b43: Add lpphy_clear_tx_power_offsets to improve TX Power handling

This patch adds the lpphy_clear_tx_power_offsets to b43.

Signed-off-by: Thomas Ilnseher <[email protected]>

---
diff -uNr a/drivers/net/wireless/b43/phy_lp.c b/drivers/net/wireless/b43/phy_lp.c
--- a/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:52:17.501318374 +0200
+++ b/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:53:36.593319452 +0200
@@ -1125,6 +1125,18 @@
dev->phy.lp->tssi_idx = (b43_phy_read(dev, B43_LPPHY_TX_PWR_CTL_STAT) & 0x7F00) >> 8;
}

+static void lpphy_clear_tx_power_offsets(struct b43_wldev *dev)
+{
+ int i;
+ int id = 7;
+ if (dev->phy.rev < 2)
+ id = 10;
+ for (i = 0; i < 12; i++)
+ b43_lptab_write(dev, B43_LPTAB32(id, 0x40 + i), 0);
+ for (i = 0; i < 64; i++)
+ b43_lptab_write(dev, B43_LPTAB32(id, 0x80 + i), 0);
+}
+
static void lpphy_set_tx_power_control(struct b43_wldev *dev,
enum b43_lpphy_txpctl_mode mode)
{
@@ -1139,7 +1151,7 @@

if (oldmode == B43_LPPHY_TXPCTL_HW) {
lpphy_update_tx_power_npt(dev);
- //TODO Clear all TX Power offsets
+ lpphy_clear_tx_power_offsets(dev);
} else {
if (mode == B43_LPPHY_TXPCTL_HW) {
//TODO Recalculate target TX power



2009-09-16 19:41:10

by Gábor Stefanik

[permalink] [raw]
Subject: Re: [PATCH 2/2] b43: Add lpphy_clear_tx_power_offsets to improve TX Power handling

You are essentially implementing dead code at this point - this will
only ever be called if hardware-accelerated TX power control is
enabled - and HW TX power control is unsupported, even for G-PHYs.

On Wed, Sep 16, 2009 at 9:37 PM, Thomas Ilnseher <[email protected]> wrote:
> This patch adds the lpphy_clear_tx_power_offsets to b43.
>
> Signed-off-by: Thomas Ilnseher <[email protected]>
>
> ---
> diff -uNr a/drivers/net/wireless/b43/phy_lp.c b/drivers/net/wireless/b43/phy_lp.c
> --- a/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:52:17.501318374 +0200
> +++ b/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:53:36.593319452 +0200
> @@ -1125,6 +1125,18 @@
> ? ? ? ?dev->phy.lp->tssi_idx = (b43_phy_read(dev, B43_LPPHY_TX_PWR_CTL_STAT) & 0x7F00) >> 8;
> ?}
>
> +static void lpphy_clear_tx_power_offsets(struct b43_wldev *dev)
> +{
> + ? ? ? int i;
> + ? ? ? int id = 7;
> + ? ? ? if (dev->phy.rev < 2)
> + ? ? ? ? ? ? ? id = 10;
> + ? ? ? for (i = 0; i < 12; i++)
> + ? ? ? ? ? ? ? b43_lptab_write(dev, B43_LPTAB32(id, 0x40 + i), 0);
> + ? ? ? for (i = 0; i < 64; i++)
> + ? ? ? ? ? ? ? b43_lptab_write(dev, B43_LPTAB32(id, 0x80 + i), 0);
> +}
> +
> ?static void lpphy_set_tx_power_control(struct b43_wldev *dev,
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? enum b43_lpphy_txpctl_mode mode)
> ?{
> @@ -1139,7 +1151,7 @@
>
> ? ? ? ?if (oldmode == B43_LPPHY_TXPCTL_HW) {
> ? ? ? ? ? ? ? ?lpphy_update_tx_power_npt(dev);
> - ? ? ? ? ? ? ? //TODO Clear all TX Power offsets
> + ? ? ? ? ? ? ? lpphy_clear_tx_power_offsets(dev);
> ? ? ? ?} else {
> ? ? ? ? ? ? ? ?if (mode == B43_LPPHY_TXPCTL_HW) {
> ? ? ? ? ? ? ? ? ? ? ? ?//TODO Recalculate target TX power
>
>



--
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)

2009-09-16 20:44:48

by Gábor Stefanik

[permalink] [raw]
Subject: Re: [PATCH 2/2] b43: Add lpphy_clear_tx_power_offsets to improve TX Power handling

2009/9/16 Thomas Ilnseher <[email protected]>:
> Am Mittwoch, den 16.09.2009, 21:40 +0200 schrieb G?bor Stefanik:
>> You are essentially implementing dead code at this point - this will
>> only ever be called if hardware-accelerated TX power control is
>> enabled - and HW TX power control is unsupported, even for G-PHYs.
> Then the question remains, why this brings my device to 54 MBit/s ?
>
> I did double check again with the old driver:
>
> wlan0 ? ? IEEE 802.11bg ?ESSID:"tommy"
> ? ? ? ? ?Mode:Managed ?Frequency:2.412 GHz ?Access Point:
> ? ? ? ? ?Bit Rate=9 Mb/s ? Tx-Power=20 dBm
> ? ? ? ? ?Retry ?long limit:7 ? RTS thr:off ? Fragment thr:off
> ? ? ? ? ?Encryption key:off
> ? ? ? ? ?Power Management:off
> ? ? ? ? ?Link Quality=70/70 ?Signal level=5 dBm
> ? ? ? ? ?Rx invalid nwid:0 ?Rx invalid crypt:0 ?Rx invalid frag:0
> ? ? ? ? ?Tx excessive retries:0 ?Invalid misc:0 ? Missed beacon:0
>
> Patched driver:
>
> wlan0 ? ? IEEE 802.11bg ?ESSID:"tommy"
> ? ? ? ? ?Mode:Managed ?Frequency:2.412 GHz ?Access Point: XXX
> ? ? ? ? ?Bit Rate=54 Mb/s ? Tx-Power=20 dBm
> ? ? ? ? ?Retry ?long limit:7 ? RTS thr:off ? Fragment thr:off
> ? ? ? ? ?Encryption key:off
> ? ? ? ? ?Power Management:off
> ? ? ? ? ?Link Quality=70/70 ?Signal level=10 dBm
> ? ? ? ? ?Rx invalid nwid:0 ?Rx invalid crypt:0 ?Rx invalid frag:0
> ? ? ? ? ?Tx excessive retries:0 ?Invalid misc:0 ? Missed beacon:0
>
>
>
>> > Signed-off-by: Thomas Ilnseher <[email protected]>
>> >
>> > ---
>> > diff -uNr a/drivers/net/wireless/b43/phy_lp.c b/drivers/net/wireless/b43/phy_lp.c
>> > --- a/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:52:17.501318374 +0200
>> > +++ b/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:53:36.593319452 +0200
>> > @@ -1125,6 +1125,18 @@
>> > ? ? ? ?dev->phy.lp->tssi_idx = (b43_phy_read(dev, B43_LPPHY_TX_PWR_CTL_STAT) & 0x7F00) >> 8;
>> > ?}
>> >
>> > +static void lpphy_clear_tx_power_offsets(struct b43_wldev *dev)
>> > +{
>> > + ? ? ? int i;
>> > + ? ? ? int id = 7;
>> > + ? ? ? if (dev->phy.rev < 2)
>> > + ? ? ? ? ? ? ? id = 10;
>> > + ? ? ? for (i = 0; i < 12; i++)
>> > + ? ? ? ? ? ? ? b43_lptab_write(dev, B43_LPTAB32(id, 0x40 + i), 0);
>> > + ? ? ? for (i = 0; i < 64; i++)
>> > + ? ? ? ? ? ? ? b43_lptab_write(dev, B43_LPTAB32(id, 0x80 + i), 0);
>> > +}
>> > +
>> > ?static void lpphy_set_tx_power_control(struct b43_wldev *dev,
>> > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? enum b43_lpphy_txpctl_mode mode)
>> > ?{
>> > @@ -1139,7 +1151,7 @@
>> >
>> > ? ? ? ?if (oldmode == B43_LPPHY_TXPCTL_HW) {
>> > ? ? ? ? ? ? ? ?lpphy_update_tx_power_npt(dev);
>> > - ? ? ? ? ? ? ? //TODO Clear all TX Power offsets
>> > + ? ? ? ? ? ? ? lpphy_clear_tx_power_offsets(dev);

Put a printk here to see if this branch is getting hit.

(BTW, are you loading b43 with the "hwpctl" modparam? That enables
experimental HW TX power control support, which might explain what you
were seeing.)

>> > ? ? ? ?} else {
>> > ? ? ? ? ? ? ? ?if (mode == B43_LPPHY_TXPCTL_HW) {
>> > ? ? ? ? ? ? ? ? ? ? ? ?//TODO Recalculate target TX power
>> >
>> >
>>
>>
>>
>
>



--
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)

2009-09-16 20:17:51

by Thomas Ilnseher

[permalink] [raw]
Subject: Re: [PATCH 2/2] b43: Add lpphy_clear_tx_power_offsets to improve TX Power handling

Am Mittwoch, den 16.09.2009, 21:40 +0200 schrieb Gábor Stefanik:
> You are essentially implementing dead code at this point - this will
> only ever be called if hardware-accelerated TX power control is
> enabled - and HW TX power control is unsupported, even for G-PHYs.
Then the question remains, why this brings my device to 54 MBit/s ?

I did double check again with the old driver:

wlan0 IEEE 802.11bg ESSID:"tommy"
Mode:Managed Frequency:2.412 GHz Access Point:
Bit Rate=9 Mb/s Tx-Power=20 dBm
Retry long limit:7 RTS thr:off Fragment thr:off
Encryption key:off
Power Management:off
Link Quality=70/70 Signal level=5 dBm
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0 Missed beacon:0

Patched driver:

wlan0 IEEE 802.11bg ESSID:"tommy"
Mode:Managed Frequency:2.412 GHz Access Point: XXX
Bit Rate=54 Mb/s Tx-Power=20 dBm
Retry long limit:7 RTS thr:off Fragment thr:off
Encryption key:off
Power Management:off
Link Quality=70/70 Signal level=10 dBm
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0 Missed beacon:0



> > Signed-off-by: Thomas Ilnseher <[email protected]>
> >
> > ---
> > diff -uNr a/drivers/net/wireless/b43/phy_lp.c b/drivers/net/wireless/b43/phy_lp.c
> > --- a/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:52:17.501318374 +0200
> > +++ b/drivers/net/wireless/b43/phy_lp.c 2009-09-16 20:53:36.593319452 +0200
> > @@ -1125,6 +1125,18 @@
> > dev->phy.lp->tssi_idx = (b43_phy_read(dev, B43_LPPHY_TX_PWR_CTL_STAT) & 0x7F00) >> 8;
> > }
> >
> > +static void lpphy_clear_tx_power_offsets(struct b43_wldev *dev)
> > +{
> > + int i;
> > + int id = 7;
> > + if (dev->phy.rev < 2)
> > + id = 10;
> > + for (i = 0; i < 12; i++)
> > + b43_lptab_write(dev, B43_LPTAB32(id, 0x40 + i), 0);
> > + for (i = 0; i < 64; i++)
> > + b43_lptab_write(dev, B43_LPTAB32(id, 0x80 + i), 0);
> > +}
> > +
> > static void lpphy_set_tx_power_control(struct b43_wldev *dev,
> > enum b43_lpphy_txpctl_mode mode)
> > {
> > @@ -1139,7 +1151,7 @@
> >
> > if (oldmode == B43_LPPHY_TXPCTL_HW) {
> > lpphy_update_tx_power_npt(dev);
> > - //TODO Clear all TX Power offsets
> > + lpphy_clear_tx_power_offsets(dev);
> > } else {
> > if (mode == B43_LPPHY_TXPCTL_HW) {
> > //TODO Recalculate target TX power
> >
> >
>
>
>


2009-09-16 21:00:14

by Thomas Ilnseher

[permalink] [raw]
Subject: Re: [PATCH 2/2] b43: Add lpphy_clear_tx_power_offsets to improve TX Power handling


>
> Put a printk here to see if this branch is getting hit.
>
> (BTW, are you loading b43 with the "hwpctl" modparam? That enables
> experimental HW TX power control support, which might explain what you
> were seeing.)
Hi Gabor,

I'm stupid.

I applied 3 Patches, the analog switch patch, and the other two patches.
The analog switch patch did it.



> >>
> >>
> >
> >
>
>
>


2009-09-16 21:00:10

by Michael Büsch

[permalink] [raw]
Subject: Re: [PATCH 2/2] b43: Add lpphy_clear_tx_power_offsets to improve TX Power handling

On Wednesday 16 September 2009 22:17:55 Thomas Ilnseher wrote:
> Am Mittwoch, den 16.09.2009, 21:40 +0200 schrieb Gábor Stefanik:
> > You are essentially implementing dead code at this point - this will
> > only ever be called if hardware-accelerated TX power control is
> > enabled - and HW TX power control is unsupported, even for G-PHYs.
> Then the question remains, why this brings my device to 54 MBit/s ?
>
> I did double check again with the old driver:
>
> wlan0 IEEE 802.11bg ESSID:"tommy"
> Mode:Managed Frequency:2.412 GHz Access Point:
> Bit Rate=9 Mb/s Tx-Power=20 dBm
> Retry long limit:7 RTS thr:off Fragment thr:off
> Encryption key:off
> Power Management:off
> Link Quality=70/70 Signal level=5 dBm
> Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
> Tx excessive retries:0 Invalid misc:0 Missed beacon:0
>
> Patched driver:
>
> wlan0 IEEE 802.11bg ESSID:"tommy"
> Mode:Managed Frequency:2.412 GHz Access Point: XXX
> Bit Rate=54 Mb/s Tx-Power=20 dBm
> Retry long limit:7 RTS thr:off Fragment thr:off
> Encryption key:off
> Power Management:off
> Link Quality=70/70 Signal level=10 dBm
> Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
> Tx excessive retries:0 Invalid misc:0 Missed beacon:0

The iwconfig output doesn't tell you anything about the actual net TX rate.
Please benchmark it with iperf.

--
Greetings, Michael.