[Meego-kernel] [PATCH] ar6003 : Support reading BT address from OTP file

[Chang, Samuel]

Hi Samuel,

>How do we handle the case for chips without an OTP area ?
>ar6000_update_bdaddr(ar) would fail, and we should fall back to reading our
>MAC from ar3kbdaddr.pst. Also, I would expect the driver to know if the target
>comes with an OTP or not. This patch doesn't even check for that.

Good question. 
So, I have a patch to fix this.
But this patch related to some files for patch, ar6003_skip_unnecessary_scan.patch

I will upstream another patch to support OPTand non-OTP devices when previous patch is finished.
 

Samuel Chang
Engineer
Qualcomm Atheros, Inc.
Taipei Office, Taiwan
12F., No.407, Ruiguang Rd., Neihu Dist.
Taipei City 11492, Taiwan (R.O.C.)
t + 886 2.8751.6385
f + 886 2.8751.6397

________________________________________
寄件者: Samuel Ortiz [sameo at linux.intel.com]
寄件日期: 2011年7月28日 上午 01:57
收件者: Chang, Samuel
副本: meego-kernel at lists.meego.com; kristen.c.accardi at intel.com; Wu, Ricky; Huang, Maggie; Yang, Jack
主旨: Re: [Meego-kernel] [PATCH] ar6003 : Support reading BT address from OTP file

Hi Samuel,

My comments below:

On Fri, Jul 22, 2011 at 05:13:45AM +0000, Chang, Samuel wrote:
> diff -ruN kernel-2.6.37.6-151.5_skip_unnecessasy_scan_patch_P1/drivers/staging/ar6003/os/linux/ar6000_drv.c kernel-2.6.37.6-151.5_bdaddress_otp_patch_P2/drivers/staging/ar6003/os/linux/ar6000_drv.c
> --- kernel-2.6.37.6-151.5_skip_unnecessasy_scan_patch_P1/drivers/staging/ar6003/os/linux/ar6000_drv.c 2011-07-18 16:16:07.816871251 +0800
> +++ kernel-2.6.37.6-151.5_bdaddress_otp_patch_P2/drivers/staging/ar6003/os/linux/ar6000_drv.c 2011-07-22 12:55:41.505195148 +0800
> @@ -1384,6 +1384,33 @@
>  #endif /* INIT_MODE_DRV_ENABLED */
>
>  A_STATUS
> +ar6000_update_bdaddr(AR_SOFTC_T *ar)
> +{
> +
> +        if (setupbtdev != 0) {
I usually find the following construct:

        if (setupbtdev == 0)
                return A_ERROR;

        if (BMIReadMemory(ar->arHifDevice,....

a lot easier to read.


> +A_STATUS
>  ar6000_sysfs_bmi_get_config(AR_SOFTC_T *ar, A_UINT32 mode)
>  {
>      AR_DEBUG_PRINTF(ATH_DEBUG_INFO,("BMI: Requesting device specific configuration\n"));
> @@ -2896,6 +2923,10 @@
>
>      ar = arPriv->arSoftc;
>
> +    if (wlaninitmode == WLAN_INIT_MODE_USR || wlaninitmode == WLAN_INIT_MODE_DRV) {
> +
> +        ar6000_update_bdaddr(ar);
> +    }
How do we handle the case for chips without an OTP area ?
ar6000_update_bdaddr(ar) would fail, and we should fall back to reading our
MAC from ar3kbdaddr.pst. Also, I would expect the driver to know if the target
comes with an OTP or not. This patch doesn't even check for that.

Cheers,
Samuel.

--
Intel Open Source Technology Centre
http://oss.intel.com/