New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 782924 link

Starred by 1 user
Status: Assigned
Owner:
briannorris@chromium.org
Cc:
derat@chromium.org
snanda@chromium.org
█ kirtika@chromium.org
diand...@chromium.org
ecgh@chromium.org
grundler@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

improve powerd/set_wifi_transmit_power logic (e.g., for ath10k)

Project Member Reported by briannorris@chromium.org, Nov 8 2017 
In investigating  bug 781838 , we noticed a few problems about assumptions:

(a) we assume we're using either mwifiex or iwlwifi
(b) we assume devices are named either "mlan0" or "wlan0"

Both of those are liable to break, but (a) is definitely wrong for ath10k.

As a side note...do we have a dynamic TX power interface for ath10k yet? Or did we ever try to push upstream to make a generic API? Custom netlink commands, rebased on each Chrome OS release, is not exactly a scalable, maintainable solution...

Comment 1 by diand...@chromium.org, Nov 8 2017 

> Both of those are liable to break, but (a) is definitely wrong for ath10k.

This doesn't cause a problem in practice because the only devices that have:

  overlays/overlay-*/chromeos-base/chromeos-bsp-*/files/powerd_prefs/set_wifi_transmit_power_for_tablet_mode

...as 1 are devices that have supported WiFi modules.  

$ grep "" $(find -name set_wifi_transmit_power_for_tablet_mode)
./overlay-bob/chromeos-base/chromeos-bsp-bob/files/powerd_prefs/set_wifi_transmit_power_for_tablet_mode:1
./overlay-kevin/chromeos-base/chromeos-bsp-kevin/files/powerd_prefs/set_wifi_transmit_power_for_tablet_mode:1
./overlay-eve/chromeos-base/chromeos-bsp-eve/files/powerd_prefs/set_wifi_transmit_power_for_tablet_mode:1
./baseboard-reef/chromeos-base/chromeos-bsp-baseboard-reef/files/powerd_prefs/model_specific/reef/set_wifi_transmit_power_for_tablet_mode:1

Comment 2 by briannorris@chromium.org, Nov 8 2017 

That may change though. ath10k is getting used on more devices, and I'm currently not familiar which ones (if any) might have dynamic TX power considerations.

Comment 3 by derat@chromium.org, Nov 9 2017

Owner: wnhuang@chromium.org
Status: Assigned (was: Untriaged)
Wei-Ning, can you fix these issues in set_wifi_transmit_power?

Comment 4 by derat@chromium.org, Nov 9 2017

Components: OS>Kernel>Power

Comment 5 by tbroch@chromium.org, Mar 24 2018

Owner: ----
Status: Available (was: Assigned)

Comment 6 by kirtika@chromium.org, Aug 7 

Making a note regarding Brian's initial comment about QCA not being covered - seems like a clean way would be to do a "vendor" lookup based on querying the vid:pid in pci.

Comment 7 by briannorris@chromium.org, Aug 8

Cc: -wnhuang@chromium.org -cernekee@chromium.org
Not everything is PCI(e). We have a lot of SDIO (not any convertibles yet, but we may if we continue to have PCIe-less ARM devices), and there are varieties of custom/integrated Wifi.

Comment 8 by briannorris@chromium.org, Nov 10

Labels: -Pri-3 M-72 Pri-2
Owner: briannorris@chromium.org
Status: Assigned (was: Available)
I'm pretty sure we need to get this in shape for ath10k soon, so I'll claim this.
►

Sign in to add a comment