linux-kernel - [PATCH 0/6] wifi: wilc1000: update power management

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20240613-wilc_suspend-v1-0-c2f766d0988c@bootlin.com>
Date: Thu, 13 Jun 2024 16:06:39 +0200
From: Alexis Lothoré <alexis.lothore@...tlin.com>
To: linux-wireless@...r.kernel.org
Cc: Ajay Singh <ajay.kathat@...rochip.com>, 
 Claudiu Beznea <claudiu.beznea@...on.dev>, Kalle Valo <kvalo@...nel.org>, 
 Thomas Petazzoni <thomas.petazzoni@...tlin.com>, 
 linux-kernel@...r.kernel.org, 
 Alexis Lothoré <alexis.lothore@...tlin.com>
Subject: [PATCH 0/6] wifi: wilc1000: update power management

Despite being implemented, power management for wilc1000 does not work as
expected and suffer some issues. Those issues can be observed by
executing "echo devices > /sys/power/pm_test && echo mem >
/sys/power/state":

- system does not resume correctly (at least on my testing platform,
  sama5d27-wlsom1-evk, either with devices test mode or user button
  configured as wake up source)
- warnings are triggered during suspend

This series brings fixes for those issues, and more generally try to
improve things a bit:
- implement only one pair of suspend/resume callbacks (those are currently
  implemented for both 802.11 class and driver, with one depending on the
  other)
- fix some warnings (for example, IRQ still being enabled on SDIO func
  while going to suspend)
- fix some scenarios leading to chip becoming unresponsive to SDIO
  re-enumeration (error -110)

With this series the following sequence works:
  - echo devices > /sys/power/pm_test
  - echo mem > /sys/power/state
- chip is properly disconnected from any AP when entering suspend
- chip properly reconnects on resume

The patches comes both from Microchip internal patches (slightly reworked)
and new patches based on encountered issues.
Tested on SAMA5D27 EVK with a WILC 1000 SD module, tested both in SDIO
mode and SPI mode

Signed-off-by: Alexis Lothoré <alexis.lothore@...tlin.com>
---
Ajay Singh (1):
      wifi: wilc1000: disable power sequencer

Alexis Lothoré (5):
      wifi: wilc1000: let host->chip suspend/resume notifications manage chip wake/sleep
      wifi: wilc1000: do not keep sdio bus claimed during suspend/resume
      wifi: wilc1000: move sdio suspend method next to resume and pm ops definition
      wifi: wilc1000: remove suspend/resume from cfg80211 part
      wifi: wilc1000: disable SDIO func IRQ before suspend

 drivers/net/wireless/microchip/wilc1000/cfg80211.c | 19 ------
 drivers/net/wireless/microchip/wilc1000/netdev.h   |  1 -
 drivers/net/wireless/microchip/wilc1000/sdio.c     | 74 ++++++++--------------
 drivers/net/wireless/microchip/wilc1000/wlan.c     |  8 +--
 drivers/net/wireless/microchip/wilc1000/wlan.h     |  1 -
 5 files changed, 30 insertions(+), 73 deletions(-)
---
base-commit: 3971a5a8f911615f9a816f76c83c37d51ac522a9
change-id: 20240123-wilc_suspend-441a5944d941

Best regards,
-- 
Alexis Lothoré, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com