Some wl1251 hardware configurations (like in WG7210 module) have
EEPROM attached where NVS data is kept, which includes MAC address.
In such configurations, let's read default MAC address from EEPROM,
instead of using random one.
Signed-off-by: Grazvydas Ignotas <[email protected]>
---
drivers/net/wireless/wl12xx/wl1251_main.c | 63 +++++++++++++++++++++++++++++
drivers/net/wireless/wl12xx/wl1251_reg.h | 7 +++
2 files changed, 70 insertions(+), 0 deletions(-)
diff --git a/drivers/net/wireless/wl12xx/wl1251_main.c b/drivers/net/wireless/wl12xx/wl1251_main.c
index 56b78e4..7bc109d 100644
--- a/drivers/net/wireless/wl12xx/wl1251_main.c
+++ b/drivers/net/wireless/wl12xx/wl1251_main.c
@@ -1196,6 +1196,66 @@ static const struct ieee80211_ops wl1251_ops = {
.conf_tx = wl1251_op_conf_tx,
};
+static int wl1251_read_eeprom_byte(struct wl1251 *wl, off_t offset, u8 *data)
+{
+ unsigned long timeout;
+
+ wl1251_reg_write32(wl, EE_ADDR, offset);
+ wl1251_reg_write32(wl, EE_CTL, EE_CTL_READ);
+
+ /* EE_CTL_READ clears when data is ready */
+ timeout = jiffies + msecs_to_jiffies(100);
+ while (1) {
+ if (!(wl1251_reg_read32(wl, EE_CTL) & EE_CTL_READ))
+ break;
+
+ if (time_after(jiffies, timeout))
+ return -ETIMEDOUT;
+
+ udelay(200);
+ }
+
+ *data = wl1251_reg_read32(wl, EE_DATA);
+ return 0;
+}
+
+static int wl1251_read_eeprom(struct wl1251 *wl, off_t offset,
+ u8 *data, size_t len)
+{
+ size_t i;
+ int ret;
+
+ wl1251_reg_write32(wl, EE_START, 0);
+
+ for (i = 0; i < len; i++) {
+ ret = wl1251_read_eeprom_byte(wl, offset + i, &data[i]);
+ if (ret < 0)
+ return ret;
+ }
+
+ return 0;
+}
+
+static int wl1251_read_eeprom_mac(struct wl1251 *wl)
+{
+ u8 mac[ETH_ALEN];
+ int i, ret;
+
+ wl1251_set_partition(wl, 0, 0, REGISTERS_BASE, REGISTERS_DOWN_SIZE);
+
+ ret = wl1251_read_eeprom(wl, 0x1c, mac, sizeof(mac));
+ if (ret < 0) {
+ wl1251_warning("failed to read MAC address from EEPROM");
+ return ret;
+ }
+
+ /* MAC is stored in reverse order */
+ for (i = 0; i < ETH_ALEN; i++)
+ wl->mac_addr[i] = mac[ETH_ALEN - i - 1];
+
+ return 0;
+}
+
static int wl1251_register_hw(struct wl1251 *wl)
{
int ret;
@@ -1242,6 +1302,9 @@ int wl1251_init_ieee80211(struct wl1251 *wl)
wl->hw->queues = 4;
+ if (wl->use_eeprom)
+ wl1251_read_eeprom_mac(wl);
+
ret = wl1251_register_hw(wl);
if (ret)
goto out;
diff --git a/drivers/net/wireless/wl12xx/wl1251_reg.h b/drivers/net/wireless/wl12xx/wl1251_reg.h
index 0ca3b43..d16edd9 100644
--- a/drivers/net/wireless/wl12xx/wl1251_reg.h
+++ b/drivers/net/wireless/wl12xx/wl1251_reg.h
@@ -46,7 +46,14 @@
#define SOR_CFG (REGISTERS_BASE + 0x0800)
#define ECPU_CTRL (REGISTERS_BASE + 0x0804)
#define HI_CFG (REGISTERS_BASE + 0x0808)
+
+/* EEPROM registers */
#define EE_START (REGISTERS_BASE + 0x080C)
+#define EE_CTL (REGISTERS_BASE + 0x2000)
+#define EE_DATA (REGISTERS_BASE + 0x2004)
+#define EE_ADDR (REGISTERS_BASE + 0x2008)
+
+#define EE_CTL_READ 2
#define CHIP_ID_B (REGISTERS_BASE + 0x5674)
--
1.6.3.3
Grazvydas Ignotas <[email protected]> writes:
> Some wl1251 hardware configurations (like in WG7210 module) have
> EEPROM attached where NVS data is kept, which includes MAC address.
>
> In such configurations, let's read default MAC address from EEPROM,
> instead of using random one.
This also looks good, just one comment:
> +static int wl1251_read_eeprom_byte(struct wl1251 *wl, off_t offset, u8 *data)
> +{
> + unsigned long timeout;
> +
> + wl1251_reg_write32(wl, EE_ADDR, offset);
> + wl1251_reg_write32(wl, EE_CTL, EE_CTL_READ);
> +
> + /* EE_CTL_READ clears when data is ready */
> + timeout = jiffies + msecs_to_jiffies(100);
> + while (1) {
> + if (!(wl1251_reg_read32(wl, EE_CTL) & EE_CTL_READ))
> + break;
> +
> + if (time_after(jiffies, timeout))
> + return -ETIMEDOUT;
> +
> + udelay(200);
> + }
Is the udelay() really needed here? In my opinion msleep(), for
example with a value of 5 ms, would be nicer from system point of
view. It doesn't matter if this function takes few milliseconds
longer, because this is called only in op_start(). What do you think?
--
Kalle Valo
Grazvydas Ignotas <[email protected]> writes:
>>> + udelay(200);
>>> + }
>>
>> Is the udelay() really needed here? In my opinion msleep(), for
>> example with a value of 5 ms, would be nicer from system point of
>> view. It doesn't matter if this function takes few milliseconds
>> longer, because this is called only in op_start(). What do you think?
>
> It normally takes something like 700-1000us, so I guess I can replace
> with msleep(1) if you think it's better.
Yes, that's better. It's always good to avoid busylooping.
--
Kalle Valo
On Tue, Apr 13, 2010 at 8:05 PM, Kalle Valo <[email protected]> wrote:
> Grazvydas Ignotas <[email protected]> writes:
>
>> Some wl1251 hardware configurations (like in WG7210 module) have
>> EEPROM attached where NVS data is kept, which includes MAC address.
>>
>> In such configurations, let's read default MAC address from EEPROM,
>> instead of using random one.
>
> This also looks good, just one comment:
>
>> +static int wl1251_read_eeprom_byte(struct wl1251 *wl, off_t offset, u8 *data)
>> +{
>> + ? ? unsigned long timeout;
>> +
>> + ? ? wl1251_reg_write32(wl, EE_ADDR, offset);
>> + ? ? wl1251_reg_write32(wl, EE_CTL, EE_CTL_READ);
>> +
>> + ? ? /* EE_CTL_READ clears when data is ready */
>> + ? ? timeout = jiffies + msecs_to_jiffies(100);
>> + ? ? while (1) {
>> + ? ? ? ? ? ? if (!(wl1251_reg_read32(wl, EE_CTL) & EE_CTL_READ))
>> + ? ? ? ? ? ? ? ? ? ? break;
>> +
>> + ? ? ? ? ? ? if (time_after(jiffies, timeout))
>> + ? ? ? ? ? ? ? ? ? ? return -ETIMEDOUT;
>> +
>> + ? ? ? ? ? ? udelay(200);
>> + ? ? }
>
> Is the udelay() really needed here? In my opinion msleep(), for
> example with a value of 5 ms, would be nicer from system point of
> view. It doesn't matter if this function takes few milliseconds
> longer, because this is called only in op_start(). What do you think?
It normally takes something like 700-1000us, so I guess I can replace
with msleep(1) if you think it's better.