Ok, can you send me a copy of the string values?
Cant really risk it due to NDA, this is all a gray area even now as those tools are not public, and its all under a ton of legal writing.
But, you can use the leaked SPF 10.0 which used 2.0 FW to maybe get a sense of it
One question: how sure we are that the BDF changes between firmware versions? Because if that is the case, our BDFs (used for both the AX3600 and AX6) are all incorrect. For example the board file for AX/ was extracted from a stock firmware with wifi version 2.0
My point is: if the BDF changes with lets say every major wifi fw release, then the BFD we use are all incorrect, and the regdb issues might be because of that.
It changes with the FW, that is how they add new stuff.
https://source.codeaurora.org/quic/qsdk/oss/ath11k-bdf/commit/?h=NHSS.QSDK.11.4.0.5&id=ee954d2fc34dc2ea93557c9fa02b16850d7dfc26
But it is supposed to be backwards compatible, meaning older BDF must work with newer FW.
Just out of fun, I took this BDF:
IPQ8074/hw2.0/WLAN.HK.2.5.0.1/WLAN.HK.2.5.0.1-01100-QCAHKSWPL_SILICONZ-1/board-2.bin
and used the qca swiss army knife to make it compatible with my device. So I am on FW 2.5.0.1 1100 and I used the paired BDF file. Everything works, but the regdb is still crap.
I really hate the crap they are pulling with regulatory db being embedded
Although the 30dBm the original BDF allowed on 2.4GHz is now gone. I wonder if the calfile format can also change between wifi and/or BDF versions...
Well, that makes sense as the BDF determines the output power limits.
Caldata format should never change as it is not upgradeable due to it being per device.
Hm... I converted my BDF to text with the tool on your link, and it gives meaningful output despite the fact that the BDF templte is version 503 and my BDF is 714.
Interesting flag:
REGULATORY_DB_SECTION.regDbFlag 0
It should be able to interpret all of the fields it has descriptions for just fine.
Yeah, that flag was the first thing that got me interested.
BTW, the friend that I contacted pointed out that the 2G and 5G channel definitions are completely broken.
In the 2G section they are defining start and end frequencies from the 5G range, 5G range same some channels without end frequencies, wrong max bandwidth etc.
Lets say this is true. How can I tell when I repack my extracted BDF file, if I am (the interpreter) not missing anything?
The only way I can think of would be to then copy the old _new.txt and convert the binary back to text again and diff the two .txt files.
It shouldn't miss anything as long as it was able to interpret it.
I am comparing the BDF came with the AX6 and the BDF from QCA (2.5.0.1 1100) and for example in the QCA version DPD is not enabled. Also, the regdb flag in the QCA released BDF is set to 0, and that is what I am using atm.
But that's the deal, QCA releases those for reference designs only so it's lacking all of the custom settings that ODM-s will tune.
DPD looks important to me for performance.
So they are really not meant to be used on custom boards as for sure the power settings will be incorrect as well.
OK, so the recommended BDF is the one extracted from the latest stock firmware of the target device. If I am not mistaken.
Yes, that is the only one you can know is tuned for that device.
2 Likes
I am looking at the extracted stock firmwares, but I can't seem to find BDFs in them. Can you point me to where they should be?
By the way the extraction of the original AX6 BDF was OK with the 503 template, but the latest BDF from QCA was total garbage. It seemingly extracts it, but if you look at the data, it is quite obvious that the extraction was failing.
I also found the mapping between countries and country codes in the BDF, so first I will try to find a country that matches mine. If that fails, I might be able to fix my country in the original AX6 BDF with this leaked tool.
Its in /lib/firmware/IPQ8074/
Thanks. I checked the BDF file we use and it is an exact match for the one found in the latest available stock firmware.
It should be as it was pulled from it