Scheduling while atomic while setting the PS mode

[Michal Kazior]

On 1 October 2013 20:11, Kalle Valo <kvalo at qca.qualcomm.com> wrote:
> Hi,
>
> I haven't been testing with all debug tests enabled for few days.
> Enabled them now and see the BUG below. This is with commit 71bcca1ff7f.
>
> Kalle
>
> [ 250.125852] ath10k_pci 0000:02:00.0: irq 49 for MSI/MSI-X
> [ 250.126959] ath10k: MSI interrupt handling
> [ 251.357561] ath10k: UART prints disabled
> [  251.432475] ath10k: firmware 999.999.0.716 booted
> [  251.443440] ath10k: htt target version 3.0
> [  251.447488] BUG: scheduling while atomic: ip/2695/0x00000002
> [  251.447604] INFO: lockdep is turned off.
> [  251.447714] Modules linked in: ath10k_pci ath10k_core ath mac80211 cfg80211 [last unloaded: cfg80211]
> [  251.449514] CPU: 1 PID: 2695 Comm: ip Tainted: G        W    3.11.0-wl-ath+ #77
> [  251.449631] Hardware name: Hewlett-Packard HP ProBook 6540b/1722, BIOS 68CDD Ver. F.04 01/27/2010
> [  251.449743]  00000000 00000000 eea437e8 c17859c9 edf75680 eea43804 c1781f31 c19a5e18
> [  251.451591]  edf75960 00000a87 00000002 f4620000 eea43894 c178c533 00000000 00000002
> [  251.453479]  00000001 00000000 c12f812f 00000000 edf75680 c1bbe000 c1a9aec8 eea43850
> [  251.454499] Call Trace:
> [  251.455469]  [<c17859c9>] dump_stack+0x4b/0x66
> [  251.455597]  [<c1781f31>] __schedule_bug+0x64/0x76
> [  251.455720]  [<c178c533>] __schedule+0x943/0xb30
> [  251.455844]  [<c12f812f>] ? debug_object_activate+0x6f/0x150
> [  251.455970]  [<c10ad8eb>] ? trace_hardirqs_off+0xb/0x10
> [  251.456091]  [<c178e696>] ? _raw_spin_unlock_irqrestore+0x36/0x60
> [  251.456211]  [<c10b3638>] ? trace_hardirqs_on_caller+0x18/0x1c0
> [  251.456352]  [<c178c9d3>] schedule+0x23/0x60
> [  251.456525]  [<c1789003>] schedule_timeout+0x103/0x280
> [  251.457480]  [<f92e6ee0>] ? ieee80211_iterate_active_interfaces+0x130/0x130 [mac80211]
> [  251.457610]  [<c1053140>] ? call_timer_fn+0x2c0/0x2c0
> [  251.457738]  [<f939a38f>] ath10k_wmi_cmd_send+0xaf/0x110 [ath10k_core]
> [  251.457861]  [<c106bdb0>] ? __init_waitqueue_head+0x60/0x60
> [  251.457986]  [<f939c57a>] ath10k_wmi_set_psmode+0x6a/0x80 [ath10k_core]
> [  251.458110]  [<f938e557>] ath10k_ps_iter+0x87/0x140 [ath10k_core]
> [  251.458258]  [<f92e701a>] ieee80211_iterate_active_interfaces_atomic+0x13a/0x2c0 [mac80211]
> [  251.458435]  [<f92e6ee0>] ? ieee80211_iterate_active_interfaces+0x130/0x130 [mac80211]

Ah, this is another thing I've missed with the WMI/HTC changes (along
with the wep key indexes) :-/

I wonder if it's about time to have a per-arvif config worker.


Michał