Bug#672891: brcmsmac wireless brcms_c_wait_for_tx_completion warning

[Touko Korpela <touko.korpela@iki.fi>]

On Thu, May 24, 2012 at 08:36:02AM -0500, Jonathan Nieder wrote:
> Touko Korpela wrote:
> 
> > This patch (?) causes possible regression.
> > On another laptop (Asus 1215B), that doesn't use encrypted wlan and didn't
> > have this warning bug, now when it's updated to 3.2.18-1 kernel have these
> > in kernel log:
> > [ 6403.270153] ieee80211 phy0: AMPDU status: BA Timeout, seq 152, in_transit 0
> > [ 8460.377520] ieee80211 phy0: AMPDU status: BA Timeout, seq 1655, in_transit 0
> 
> These are harmless and suppressed by
> 
> commit 2b0a53d51b5f263bb581bbdb40ebb9f7e09609b1
> Author: Arend van Spriel <arend@broadcom.com>
> Date:   Thu Feb 23 18:38:23 2012 +0100
> 
>     brcm80211: smac: only print block-ack timeout message at trace level
> 
>     In regular use block-ack timeouts can happen so it does not make
>     sense to fill the log with these messages.
> 
> in 3.3.
> 
> [...]
> > [138693.625124] ieee80211 phy0: AMPDU status: BA Timeout, seq 3545, in_transit 0
> > [138876.798913] ieee80211 phy0: wl0: ampdu tx phy error (0x20)
> 
> This is the one that seemed potentially interesting.  Does it happen often?

Not very often, in between dozens "BA Timeout" messages, only two of those
"ampdu tx phy error" messages are printed (40min apart, time 10:00 and 10:40).
Also in time 10:25 is printed "brcms_c_ampdu_dotxstatus_complete:supr_status 0x18"
I have reported also bug #671803
I don't know if it's regression or not but when I rebooted 3.2.18-1, kernel
oopsed during shutdown (it wasn't logged). I don't remember that happened before.