APPLIED: [linux-firmware][PATCH 0/1][SRU][B/D] Update firmware file for Intel Bluetooth 9260
Seth Forshee
seth.forshee at canonical.com
Mon Nov 4 21:33:18 UTC 2019
On Fri, Oct 25, 2019 at 03:59:14PM +0800, You-Sheng Yang wrote:
> From: You-Sheng Yang <vicamo at gmail.com>
>
> BugLink: https://bugs.launchpad.net/bugs/1849779
>
> [Impact]
> On some platform, with Intel 9260 Bluetooth fw ver < REL0472, deep
> suspend (S3) will be interrupted and resume automatically once it has
> been paired with a BLE device. The device doesn't have to remain
> connected, as long as there are BLE devices near by broadcasting LE
> ADV.
>
> This doesn't seem to apply to all platforms with 9260, as there are
> other platforms with identical chip and fw version but are immune from
> this issue.
>
> [Fix]
> This can be fixed with newer (>= REL0472) release of fw blob
> intel/ibt-18-16-1.sfi:
>
> |commit |fw rev|4.15.0-66-generic|series|
> |------------|------|-----------------|------|
> |c2d8f1b7f820| 0329| waked up|bionic|
> |ae90c3bce108| 0386| waked up| |
> |046d1085b19e| 0420| waked up| disco|
> |1e8253bbaafb| 0450| waked up| |
> |7444ca40083a| 0472| ok| |
> |fe48882708de| 0482| ok| eoan|
>
> This patch will SRU 0482 to keep synced with eoan.
>
> [Test Case]
> 1. pair with BLE device and turn off the device. It should be listed
> as Disconnected in System Bluetooth settings. Close the settings
> app as well.
> 2. run `sudo btmon -t` in the background, make sure LE ADV packets
> pop up.
> 3. run `sudo btmon -t` on another host to make sure there are LE ADV
> packets during the suspend period.
> 3. trigger system deep suspend (S3) on DUT.
> 4. wait for around 1 minutes (depending on the surrounding
> environment) and see if the DUT is resumed unsolicitedly.
>
> [Regression Risk]
> The same process is also verified on Dell Precision 7530 that is also
> equipped with Intel 9260 Bluetooth. However, although also has
> linux-firmware 1.173.9 installed (fw REL0329), this platform doesn't
> have such unsolicited resume problem. With fw upgraded to proposed
> REL0482, there is no known regression found in the suspend/resume and
> other functions. Low.
Applied for bionic/disco, thanks!
More information about the kernel-team
mailing list