Received-SPF: pass (google.com: domain of linux-bluetooth-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
To:     linux-bluetooth@vger.kernel.org, luiz.von.dentz@intel.com
From:   Chris Clayton <chris2553@googlemail.com>
Subject: BT Intel: Regression loading firmware in linux-5.13-rc3
Message-ID: <81eddb07-f346-c12a-345b-d929fd2edb10@googlemail.com>
Date:   Tue, 25 May 2021 10:58:17 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.10.2
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Language: en-GB
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi,

I've built and installed 5.13-rc3+ (freshly pulled from Linus' tree). On booting it I find that none of my bluetooth
devices can connect. In the output from dmesg I find:

[    8.170332] Bluetooth: hci0: Firmware revision 0.0 build 121 week 7 2021
[    8.171339] bluetooth hci0: Direct firmware load for intel/ibt-19-16-0.sfi failed with error -2
[    8.171871] Bluetooth: hci0: Failed to load Intel firmware file intel/ibt-19-16-0.sfi (-2)

As you can see it is looking for a non-existent firmware file intel/ibt-19-16-0.sfi. intel/ibt-19-16-4.{ddc,sfi} do
exist and both 5.12.6 and 5.10.39 successfully load those.

If I create symbolic links from the ibt-19-16-4 files to ibt-19-16-0 equivalents, rc3 loads them and my bluetooth
devices work fine. I can't find any ibt-19-16-0 files in the lkinux-firmware tree on kernel.org (or anywhere else for
that matter).

I looked at the changes introduced in 5.13 and found a couple of suspects - d68903da4e220d1e6b7c6ecdb853c36144c6acc9
(Bluetooth: btintel: Consolidate intel_version parsing) and its parent commit 0a460d8fe2db6887169a19b048ea0c90f8bdc3b7
(Bluetooth: btintel: Consolidate intel_version_tlv parsing). Both were authored by Luiz Augusto von Dentz and make
changes to code that processes the firmware version data. I built a kernel with 0a460d8fe2db6887169a19b048ea0c90f8bdc3b7
checked out and that loads the installed firmware fine. d68903da4e220d1e6b7c6ecdb853c36144c6acc9, however, seems to be
the the change that breaks things because it tries (and obviously fails) to load ibt-19-16-0.sfi.

I can provide any additional diagnostics or test a fix, but please copy me in to any email because I'm not subscribed to
the list.

In the meantime, I'll use my symlinks workaround.

Thanks,

Chris