* [PATCH 2/4] wl12xx: set the irq polarity before loading the fw
2012-05-08 11:02 [PATCH 1/4] wlcore: use psd_type indexing according to spec Yoni Divinsky
@ 2012-05-08 11:02 ` Yoni Divinsky
2012-05-08 11:02 ` [PATCH 3/4] wlcore: do not send stop fwlog cmd if fw is hanged Yoni Divinsky
` (2 subsequent siblings)
3 siblings, 0 replies; 6+ messages in thread
From: Yoni Divinsky @ 2012-05-08 11:02 UTC (permalink / raw)
To: linux-wireless; +Cc: arik, coelho
The polarity should be set before the firmware is loaded
since the firmware touches the same register. Access
of the firmware and driver to the same register will
cause a collision since there is no exclusion scheme.
Signed-off-by: Yoni Divinsky <yoni.divinsky@ti.com>
---
drivers/net/wireless/ti/wl12xx/main.c | 12 ++++++------
1 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/net/wireless/ti/wl12xx/main.c b/drivers/net/wireless/ti/wl12xx/main.c
index d7dd3de..d2ea424 100644
--- a/drivers/net/wireless/ti/wl12xx/main.c
+++ b/drivers/net/wireless/ti/wl12xx/main.c
@@ -989,7 +989,7 @@ out:
static void wl12xx_pre_upload(struct wl1271 *wl)
{
- u32 tmp;
+ u32 tmp, polarity;
/* write firmware's last address (ie. it's length) to
* ACX_EEPROMLESS_IND_REG */
@@ -1009,18 +1009,18 @@ static void wl12xx_pre_upload(struct wl1271 *wl)
if (wl->chip.id == CHIP_ID_1283_PG20)
wl12xx_top_reg_write(wl, SDIO_IO_DS, HCI_IO_DS_6MA);
-}
-
-static void wl12xx_enable_interrupts(struct wl1271 *wl)
-{
- u32 polarity;
+ /* polarity must be set before the firmware is loaded */
polarity = wl12xx_top_reg_read(wl, OCP_REG_POLARITY);
/* We use HIGH polarity, so unset the LOW bit */
polarity &= ~POLARITY_LOW;
wl12xx_top_reg_write(wl, OCP_REG_POLARITY, polarity);
+}
+
+static void wl12xx_enable_interrupts(struct wl1271 *wl)
+{
wlcore_write_reg(wl, REG_INTERRUPT_MASK, WL1271_ACX_ALL_EVENTS_VECTOR);
wlcore_enable_interrupts(wl);
--
1.7.0.4
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH 3/4] wlcore: do not send stop fwlog cmd if fw is hanged
2012-05-08 11:02 [PATCH 1/4] wlcore: use psd_type indexing according to spec Yoni Divinsky
2012-05-08 11:02 ` [PATCH 2/4] wl12xx: set the irq polarity before loading the fw Yoni Divinsky
@ 2012-05-08 11:02 ` Yoni Divinsky
2012-05-11 7:46 ` Arik Nemtsov
2012-05-08 11:02 ` [PATCH 4/4] wlcore: fix the CONF_TX_AC_ANY_TID to be 0xff Yoni Divinsky
2012-06-06 8:17 ` [PATCH 1/4] wlcore: use psd_type indexing according to spec Luciano Coelho
3 siblings, 1 reply; 6+ messages in thread
From: Yoni Divinsky @ 2012-05-08 11:02 UTC (permalink / raw)
To: linux-wireless; +Cc: arik, coelho
If the driver received a watchdog interrupt then the
assumption is that the fw is hanged. Avoid sending
the stop fwlog command in case of a watchdog recovey
to avoid waiting for the 2 seconds timeout of the command.
Signed-off-by: Yoni Divinsky <yoni.divinsky@ti.com>
---
drivers/net/wireless/ti/wlcore/main.c | 9 +++++++--
drivers/net/wireless/ti/wlcore/wlcore.h | 1 +
2 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers/net/wireless/ti/wlcore/main.c b/drivers/net/wireless/ti/wlcore/main.c
index 45fe911..b8dcef3 100644
--- a/drivers/net/wireless/ti/wlcore/main.c
+++ b/drivers/net/wireless/ti/wlcore/main.c
@@ -582,6 +582,7 @@ static irqreturn_t wl1271_irq(int irq, void *cookie)
if (unlikely(intr & WL1271_ACX_INTR_WATCHDOG)) {
wl1271_error("watchdog interrupt received! "
"starting recovery.");
+ wl->watchdog_recovery = true;
wl12xx_queue_recovery_work(wl);
/* restarting the chip. ignore any other interrupt. */
@@ -820,10 +821,12 @@ static void wl12xx_read_fwlog_panic(struct wl1271 *wl)
/*
* Make sure the chip is awake and the logger isn't active.
- * This might fail if the firmware hanged.
+ * Do not send a stop fwlog command if the fw is hanged.
*/
- if (!wl1271_ps_elp_wakeup(wl))
+ if (!wl1271_ps_elp_wakeup(wl) && !wl->watchdog_recovery)
wl12xx_cmd_stop_fwlog(wl);
+ else
+ goto out;
/* Read the first memory block address */
wl12xx_fw_status(wl, wl->fw_status);
@@ -926,6 +929,7 @@ static void wl1271_recovery_work(struct work_struct *work)
* to restart the HW.
*/
ieee80211_wake_queues(wl->hw);
+ wl->watchdog_recovery = false;
return;
out_unlock:
mutex_unlock(&wl->mutex);
@@ -1049,6 +1053,7 @@ int wl1271_plt_start(struct wl1271 *wl)
wl->plt = true;
wl->state = WL1271_STATE_ON;
+ wl->watchdog_recovery = false;
wl1271_notice("firmware booted in PLT mode (%s)",
wl->chip.fw_ver_str);
diff --git a/drivers/net/wireless/ti/wlcore/wlcore.h b/drivers/net/wireless/ti/wlcore/wlcore.h
index 0b3f0b5..d6a8ad4 100644
--- a/drivers/net/wireless/ti/wlcore/wlcore.h
+++ b/drivers/net/wireless/ti/wlcore/wlcore.h
@@ -228,6 +228,7 @@ struct wl1271 {
/* Hardware recovery work */
struct work_struct recovery_work;
+ bool watchdog_recovery;
/* Pointer that holds DMA-friendly block for the mailbox */
struct event_mailbox *mbox;
--
1.7.0.4
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH 3/4] wlcore: do not send stop fwlog cmd if fw is hanged
2012-05-08 11:02 ` [PATCH 3/4] wlcore: do not send stop fwlog cmd if fw is hanged Yoni Divinsky
@ 2012-05-11 7:46 ` Arik Nemtsov
0 siblings, 0 replies; 6+ messages in thread
From: Arik Nemtsov @ 2012-05-11 7:46 UTC (permalink / raw)
To: Yoni Divinsky; +Cc: linux-wireless, coelho
On Tue, May 8, 2012 at 2:02 PM, Yoni Divinsky <yoni.divinsky@ti.com> wrote:
> If the driver received a watchdog interrupt then the
> assumption is that the fw is hanged. Avoid sending
> the stop fwlog command in case of a watchdog recovey
> to avoid waiting for the 2 seconds timeout of the command.
>
> Signed-off-by: Yoni Divinsky <yoni.divinsky@ti.com>
> ---
> drivers/net/wireless/ti/wlcore/main.c | 9 +++++++--
> drivers/net/wireless/ti/wlcore/wlcore.h | 1 +
> 2 files changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/ti/wlcore/main.c b/drivers/net/wireless/ti/wlcore/main.c
> index 45fe911..b8dcef3 100644
> --- a/drivers/net/wireless/ti/wlcore/main.c
> +++ b/drivers/net/wireless/ti/wlcore/main.c
> @@ -582,6 +582,7 @@ static irqreturn_t wl1271_irq(int irq, void *cookie)
> if (unlikely(intr & WL1271_ACX_INTR_WATCHDOG)) {
> wl1271_error("watchdog interrupt received! "
> "starting recovery.");
> + wl->watchdog_recovery = true;
> wl12xx_queue_recovery_work(wl);
>
> /* restarting the chip. ignore any other interrupt. */
> @@ -820,10 +821,12 @@ static void wl12xx_read_fwlog_panic(struct wl1271 *wl)
>
> /*
> * Make sure the chip is awake and the logger isn't active.
> - * This might fail if the firmware hanged.
> + * Do not send a stop fwlog command if the fw is hanged.
> */
> - if (!wl1271_ps_elp_wakeup(wl))
> + if (!wl1271_ps_elp_wakeup(wl) && !wl->watchdog_recovery)
> wl12xx_cmd_stop_fwlog(wl);
> + else
> + goto out;
>
> /* Read the first memory block address */
> wl12xx_fw_status(wl, wl->fw_status);
> @@ -926,6 +929,7 @@ static void wl1271_recovery_work(struct work_struct *work)
> * to restart the HW.
> */
> ieee80211_wake_queues(wl->hw);
> + wl->watchdog_recovery = false;
> return;
Please change watchdog_recovery only inside the wl->mutex.
> out_unlock:
> mutex_unlock(&wl->mutex);
Please change watchdog_recovery to false here as well.
> @@ -1049,6 +1053,7 @@ int wl1271_plt_start(struct wl1271 *wl)
>
> wl->plt = true;
> wl->state = WL1271_STATE_ON;
> + wl->watchdog_recovery = false;
> wl1271_notice("firmware booted in PLT mode (%s)",
> wl->chip.fw_ver_str);
This is not needed here once watchdog_recovery is always set to false
indie recovery_work().
Arik
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 4/4] wlcore: fix the CONF_TX_AC_ANY_TID to be 0xff
2012-05-08 11:02 [PATCH 1/4] wlcore: use psd_type indexing according to spec Yoni Divinsky
2012-05-08 11:02 ` [PATCH 2/4] wl12xx: set the irq polarity before loading the fw Yoni Divinsky
2012-05-08 11:02 ` [PATCH 3/4] wlcore: do not send stop fwlog cmd if fw is hanged Yoni Divinsky
@ 2012-05-08 11:02 ` Yoni Divinsky
2012-06-06 8:17 ` [PATCH 1/4] wlcore: use psd_type indexing according to spec Luciano Coelho
3 siblings, 0 replies; 6+ messages in thread
From: Yoni Divinsky @ 2012-05-08 11:02 UTC (permalink / raw)
To: linux-wireless; +Cc: arik, coelho
In the enum conf_tx_ac CONF_TX_AC_ANY_TID should
be 0xff to match the firmware's implementation.
Signed-off-by: Yoni Divinsky <yoni.divinsky@ti.com>
---
drivers/net/wireless/ti/wlcore/conf.h | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/wireless/ti/wlcore/conf.h b/drivers/net/wireless/ti/wlcore/conf.h
index fef0db4..dbea4db 100644
--- a/drivers/net/wireless/ti/wlcore/conf.h
+++ b/drivers/net/wireless/ti/wlcore/conf.h
@@ -504,7 +504,7 @@ enum conf_tx_ac {
CONF_TX_AC_VI = 2, /* video */
CONF_TX_AC_VO = 3, /* voice */
CONF_TX_AC_CTS2SELF = 4, /* fictitious AC, follows AC_VO */
- CONF_TX_AC_ANY_TID = 0x1f
+ CONF_TX_AC_ANY_TID = 0xff
};
struct conf_tx_ac_category {
--
1.7.0.4
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH 1/4] wlcore: use psd_type indexing according to spec
2012-05-08 11:02 [PATCH 1/4] wlcore: use psd_type indexing according to spec Yoni Divinsky
` (2 preceding siblings ...)
2012-05-08 11:02 ` [PATCH 4/4] wlcore: fix the CONF_TX_AC_ANY_TID to be 0xff Yoni Divinsky
@ 2012-06-06 8:17 ` Luciano Coelho
3 siblings, 0 replies; 6+ messages in thread
From: Luciano Coelho @ 2012-06-06 8:17 UTC (permalink / raw)
To: Yoni Divinsky; +Cc: linux-wireless, arik
On Tue, 2012-05-08 at 14:02 +0300, Yoni Divinsky wrote:
> In ieee80211.h the uapsd bit mask is defined such that
> VO=BIT(0), VI=BIT(1), BK=BIT(2), BE=BIT(3).
> The firmware uses the indexing as defined in the ieee80211
> spec, meaning that VO=3, VI=2, BK=1, BE=0.
>
> In AP mode when adding peer wlcore needs to convert
> the indexing accordingly.
>
> Signed-off-by: Yoni Divinsky <yoni.divinsky@ti.com>
> ---
Applied and pushed this series, except for 3/4, which was sent as v2 in
a separate patchset.
--
Luca.
^ permalink raw reply [flat|nested] 6+ messages in thread