2010-06-21 11:48:50

by Daniel Golle

[permalink] [raw]
Subject: [RFC] refactor ath9k_platform to sound sane for use in both ath9k and ath5k

I'm developing on OpenWrt to work with EEPROM-less ath5k mini-pci
devices. This is needed to support the ar71xx-based Senao EAP7660D board
which got two AR5413 modules soldered into its mini-pci slots. In the
original Firmware, MAC addresses as well as calibration-data seems to be
part of the firmware blob stored on the board flash.
See
https://lists.openwrt.org/pipermail/openwrt-devel/2010-June/007366.html
for the corresponding discussion on OpenWrt-devel.
Similarly to the way this happens for ath9k, I managed to get ath5k to
use the MAC address and eeprom-data supplied by ath9k_platform_data.
Including ath9k_platform.h in ath5k-sources looks confusing as the name
ath9k_platform suggests that the supplied ath9k_platform_info struct is
specific to ath9k devices.
Consequently, I believe ath9k_platform.h should be renamed into
ath_platform.h, the macro ATH9K_PLAT_EEP_MAX_WORDS into
ATH_PLAT_EEP_MAX_WORDS and struct ath9k_platform_data should be
refactored to be struct ath_platform_data.
Please let me know if you agree with this in theory; if yes, I'll start
posting the patches to OpenWrt.

Cheers

Daniel


2010-06-21 18:02:53

by Luis R. Rodriguez

[permalink] [raw]
Subject: Re: [RFC] refactor ath9k_platform to sound sane for use in both ath9k and ath5k

On Mon, Jun 21, 2010 at 4:48 AM, Daniel Golle <[email protected]> wrote:
> I'm developing on OpenWrt to work with EEPROM-less ath5k mini-pci devices.
> This is needed to support the ar71xx-based Senao EAP7660D board which got
> two AR5413 modules soldered into its mini-pci slots. In the original
> Firmware, MAC addresses as well as calibration-data seems to be part of the
> firmware blob stored on the board flash.
> See https://lists.openwrt.org/pipermail/openwrt-devel/2010-June/007366.html
> for the corresponding discussion on OpenWrt-devel.

I don't see much "discussion" there.

> Similarly to the way this happens for ath9k, I managed to get ath5k to use
> the MAC address and eeprom-data supplied by ath9k_platform_data.
> Including ath9k_platform.h in ath5k-sources looks confusing as the name
> ath9k_platform suggests that the supplied ath9k_platform_info struct is
> specific to ath9k devices.
> Consequently, I believe ath9k_platform.h should be renamed into
> ath_platform.h, the macro ATH9K_PLAT_EEP_MAX_WORDS into
> ATH_PLAT_EEP_MAX_WORDS and struct ath9k_platform_data should be refactored
> to be struct ath_platform_data.
> Please let me know if you agree with this in theory; if yes, I'll start
> posting the patches to OpenWrt.

You can stuff any general stuff into the ath module which is shared by both.

Luis

2010-07-23 02:02:52

by Daniel Golle

[permalink] [raw]
Subject: Re: [RFC] refactor ath9k_platform to sound sane for use in both ath9k and ath5k

Hi!
In order to support the Senao EAP7660D board, I now added
ath5k_platform.h and patched ath5k (and madwifi).
In this way, the infrastructure for ath9k and ath5k remains separate, as
discussed in
https://lists.openwrt.org/pipermail/openwrt-devel/2010-July/007506.html

Is there any chance to have the patches added in
https://dev.openwrt.org/changeset/22188
included to linux-wireless?

Thank you!

Daniel

On 06/21/2010 08:02 PM, Luis R. Rodriguez wrote:
> On Mon, Jun 21, 2010 at 4:48 AM, Daniel Golle<[email protected]> wrote:
>
>> I'm developing on OpenWrt to work with EEPROM-less ath5k mini-pci devices.
>> This is needed to support the ar71xx-based Senao EAP7660D board which got
>> two AR5413 modules soldered into its mini-pci slots. In the original
>> Firmware, MAC addresses as well as calibration-data seems to be part of the
>> firmware blob stored on the board flash.
>> See https://lists.openwrt.org/pipermail/openwrt-devel/2010-June/007366.html
>> for the corresponding discussion on OpenWrt-devel.
>>
> I don't see much "discussion" there.
>
>
>> Similarly to the way this happens for ath9k, I managed to get ath5k to use
>> the MAC address and eeprom-data supplied by ath9k_platform_data.
>> Including ath9k_platform.h in ath5k-sources looks confusing as the name
>> ath9k_platform suggests that the supplied ath9k_platform_info struct is
>> specific to ath9k devices.
>> Consequently, I believe ath9k_platform.h should be renamed into
>> ath_platform.h, the macro ATH9K_PLAT_EEP_MAX_WORDS into
>> ATH_PLAT_EEP_MAX_WORDS and struct ath9k_platform_data should be refactored
>> to be struct ath_platform_data.
>> Please let me know if you agree with this in theory; if yes, I'll start
>> posting the patches to OpenWrt.
>>
> You can stuff any general stuff into the ath module which is shared by both.
>
> Luis
>