Al 26/02/11 22:15, En/na Joerg Albert ha escrit:
> On 02/26/2011 06:21 PM, Luca Olivetti wrote:
>> Al 26/02/11 17:22, En/na Luca Olivetti ha escrit:
>
>>> Yes, I also have two areas beginning with 0xa5 0x5a, maybe you're right,
>>> the first one is just
>>> for the pci fixup and the second one is the real calibration data.
>>> I'll try with the second one.
>>
>> Well, that was not it: if I use what appears to be the second area as
>> the calibration data I get this:
>>
>> ath: Bad EEPROM checksum 0x0 or revision 0x0000
>
> I wasn't aware that you already passed the WLAN calibration data from the
> flash
> into the ath9k driver - how do you do this?
static struct ath9k_platform_data arv7518_ath9k_wmac_data = {
.led_pin = -1,
.check_endianness = 1, /* I patched this in the ath9k driver, see patch
to mac80211 posted here some days ago */
};
static int arv7518_pci_plat_dev_init(struct pci_dev *dev)
{
dev->dev.platform_data = &arv7518_ath9k_wmac_data;
return 0;
}
void __init
arv7518_register_ath9k(unsigned char *mac)
{
#define ARV7518_BRN_ATH 0xa07f0400
u16 *eepdata;
memcpy_fromio(arv7518_ath9k_wmac_data.eeprom_data,
(void *)KSEG1ADDR(LQ_FLASH_START + ARV7518_BRN_ATH
),sizeof(arv7518_ath9k_wmac_data.eeprom_data));
eepdata = (u16 *) (&arv7518_ath9k_wmac_data.eeprom_data);
arv7518_ath9k_wmac_data.macaddr = mac;
lqpci_plat_dev_init = arv7518_pci_plat_dev_init;
pci_enable_ath9k_fixup(14, arv7518_ath9k_wmac_data.eeprom_data);
}
pci_enable_ath9k_fixup is the same as for ar71xx, with the endianness problem
discussed earlier. Of course I'm not sure I doing things right, due to the bad
performance.
> I'm still struggling with getting the PCI fixup into the code, bcm63xx and
> ar71xx
> are quite different here.
>
>> Once I go back to my previous, apparently working, version, I noticed that
>> the regdomain is US, and that's surely means there's something wrong in the
>> eeprom,
>> or my treatment of it: I'm in Spain and this device is custom made for a
>> local
>> ISP.
>
> Maybe the vendor WLAN driver allowed override of the regdomain on commandline.
> Have you got access to any source code from the vendor of your device
> (similar to the Linksys GPL tarballs etc.)?
Nope, the device uses a closed source crappy as hell firmware.
Bye
--
Luca
_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
