Arch Linux

roma0104

Member

Registered: 2012-10-31

Posts: 3

rtl8192ce resume freeze -- upstream bug

For the last year(?, or longer) my laptop will not resume from a suspend.  Suspending the netbook with a command, time-out, or closing the lid results in a sucessful suspend, from what I can tell.  Waking it, the power light returns, there is brief disk activity but the backlight does not turn on and there does not appear to be any output to the screen.  The laptop cannot be SSH to after the resume, no response/route. The only solution when in this state is a hard reset.

I believe it is related to the wifi card and/or its driver.  If wireless is being managed by WICD or NetworkManager the netbook will not resume, however, if it is being managed by netctl it resumes without a problem.  My Arch install is up-to-date as is the kernel.  I believe the problem is upstream with the kernel: Any Ubuntu-based distro exhibits the same problem; an Lubuntu and Xubuntu with 3.16.0 will not wake.  Debian Wheezy 7.8 with the 3.4.105 does not have this problem.  Debian Jessie with the 3.16.0 only exhibits this issue on occasion, roughly less than 25% of time.  I have yet 'wasted' time to recompile kernels to find the exact version the break happens.

journalctl does not report an error.

I would like to report it upstream but I feel the issue is still fuzzy.  Can someone help me divine further the issue, namely what is the difference between netctl versus WICD and NetworkManager that might result in this behaviour?  I do not have enough knowledge.  Is the wake-up calls different?  Do the handle drivers differently? Etcetera...

# lspci
...
07:00.0 Network controller: Realtek Semiconductor Co., Ltd. RTL8188CE 802.11b/g/n WiFi Adapter (rev 01)
09:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8101E/RTL8102E PCI Express Fast Ethernet controller (rev 05)

# lsmod
...
rtl8192ce              56303  0
rtl_pci                25988  1 rtl8192ce
rtl8192c_common        44268  1 rtl8192ce
rtlwifi                63199  3 rtl_pci,rtl8192c_common,rtl8192ce
...

Journal after hard reset due to hang

# journalctl
...
Feb 15 05:23:35 localhost kernel: Initializing cgroup subsys cpuset
Feb 15 05:23:35 localhost systemd-journal[117]: Time spent on flushing to /var is 2.966ms for 2 entries.
Feb 15 05:23:35 localhost systemd-journal[117]: Permanent journal is using 136.0M (max allowed 2.3G, trying to 
Feb 15 05:23:34 localhost systemd-journal[117]: Runtime journal is using 6.1M (max allowed 49.3M, trying to lea
-- Reboot --
Feb 15 05:23:02 localhost systemd-sleep[579]: Suspending system...
Feb 15 05:23:02 localhost polkitd[571]: Unregistered Authentication Agent for unix-process:568:16215 (system bu
Feb 15 05:23:02 localhost systemd[1]: Starting Suspend...
Feb 15 05:23:02 localhost systemd[1]: Reached target Sleep.
Feb 15 05:23:02 localhost systemd[1]: Starting Sleep.
...

Dmesg from a pm_trace call.
https://wiki.ubuntu.com/DebuggingKernelSuspend
https://www.kernel.org/doc/Documentatio … /s2ram.txt

# sync && echo 1 > /sys/power/pm_trace && echo mem > /sys/power/state
...
[    1.581677]   Magic number: 3:925:156
[    1.581741] acpi device:23: hash matches
[    1.581773] acpi PNP0A08:00: hash matches
...

I cannot figure out the identity of device:23 or PNP0A08:00.  I can find them in /sys, but cat-ing any of the files in device:23, if produces text, is just as cryptic.

cat of path file in device:23

\_SB_.PCI0.EUSB.RHUB.USBC

A USB hub?