Re: [PATCH 1/1] New driver: rtl8xxxu (mac80211)

[Jes Sorensen <Jes.Sorensen@redhat.com>]

Jakub Sitnicki <jsitnicki@gmail.com> writes:
> Hi Jes,
>
> On Sat, Aug 29, 2015 at 11:18 PM CEST, Jes.Sorensen@redhat.com wrote:
>> This is an alternate driver for a number of Realtek WiFi USB devices,
>> including RTL8723AU, RTL8188CU, RTL8188RU, RTL8191CU, and RTL8192CU.
>> It was written from scratch utilizing the Linux mac80211 stack.
>>
>> After spending months cleaning up the vendor provided rtl8723au
>> driver, which comes with it's own 802.11 stack included, I decided to
>> rewrite this driver from the bottom up.
>
> In the long term, do you plan for this driver to support the
> RTL8188EU-based devices as well?

I have an 8188EU dongle in my wifi dongle pouch, together with several
other yet to be supported ones :) I am currently trying to get the
8723BU and 8192EU working, but I haven't had time to work much on the
8188EU yet.

> I've been checking how far I can get through the rtl8xxxu initialization
> using a TP-Link TL-WN725N dongle without changing too much.
>
> Getting it up to read the EFUSE contents wasn't difficult but then you
> need to parse it and first differences come to light.  Namely, TX power
> levels are defined per each of 6 (not 3) channel groups, and power
> differences are defined per TX path (as opposed to per channel group).
> This is perhaps better described by the relevant structure from
> rtl8188eu driver:
>
> struct txpowerinfo24g {
> 	u8 IndexCCK_Base[MAX_RF_PATH][MAX_CHNL_GROUP_24G];
> 	u8 IndexBW40_Base[MAX_RF_PATH][MAX_CHNL_GROUP_24G];
> 	/* If only one tx, only BW20 and OFDM are used. */
> 	s8 CCK_Diff[MAX_RF_PATH][MAX_TX_COUNT];
> 	s8 OFDM_Diff[MAX_RF_PATH][MAX_TX_COUNT];
> 	s8 BW20_Diff[MAX_RF_PATH][MAX_TX_COUNT];
> 	s8 BW40_Diff[MAX_RF_PATH][MAX_TX_COUNT];
> };
>
> AFAICT, in this regard, rtl8188eu driver has more in common with the
> rtlwifi/rtl8188ee driver.
>
> I'm curious what is your opinion on this.  In particular, do you see any
> reason not to make an attempt to extend rtl8xxxu to support 8188EU
> chips?

Parsing the efuse is normally fairly easy, if you have a reference to
the layout. It can be deducted from the vendor driver code. The layout
pretty much differ for each chip, but I have seen enough of them by now
that I can parse most of it by looking at the hex dump.

I think the 8188EU uses IOL, which is an enhanced firmware assist API.
That will change the programming a bit, but I think it's well within
scope for this driver to support this as well.

In other words, it is my plan to look at the 8188EU. If you get it going
before me, I am very happy to work with you on integrating the patches.

Cheers,
Jes