| Message ID | 1357668645-5101-4-git-send-email-seth.forshee@canonical.com (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
Hi On Tue, Jan 08, 2013 at 12:10:45PM -0600, Seth Forshee wrote: > Leaving powersave enabled while the scan is suspended requires the use > of PS-Poll to retrieve frames buffered at the AP. PS-Poll isn't a very > efficient means of data transfer and may not even work with drivers not > supporting powersave. These problems can largely be avoided by simply > disabling off-channel powersave when the scan is suspended and enabling > it again when returning to off-channel operation. > > The results of this change are fantastic. Previously when using iperf I > would see greatly decreased throughput (often dropping to 0 bits/sec for > several seconds) and very high packet loss during software scans. This > patch completely eliminates the packet loss and reduces the loss in > throughput to very reasonable levels. I posted a patch, which remove offchannel_ps_disable argument from ieee80211_offchannel_return, since there were possible AP and STA power save state mismash: http://marc.info/?l=linux-wireless&m=135601089321822&w=2 It should give the same effect on performance as this patch. Thanks Stanislaw -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Jan 09, 2013 at 12:03:44PM +0100, Stanislaw Gruszka wrote: > Hi > > On Tue, Jan 08, 2013 at 12:10:45PM -0600, Seth Forshee wrote: > > Leaving powersave enabled while the scan is suspended requires the use > > of PS-Poll to retrieve frames buffered at the AP. PS-Poll isn't a very > > efficient means of data transfer and may not even work with drivers not > > supporting powersave. These problems can largely be avoided by simply > > disabling off-channel powersave when the scan is suspended and enabling > > it again when returning to off-channel operation. > > > > The results of this change are fantastic. Previously when using iperf I > > would see greatly decreased throughput (often dropping to 0 bits/sec for > > several seconds) and very high packet loss during software scans. This > > patch completely eliminates the packet loss and reduces the loss in > > throughput to very reasonable levels. > > I posted a patch, which remove offchannel_ps_disable argument from > ieee80211_offchannel_return, since there were possible AP and STA power > save state mismash: > > http://marc.info/?l=linux-wireless&m=135601089321822&w=2 > > It should give the same effect on performance as this patch. I guess I missed that one in my pile of post-holiday emails. I agree, it ought to have the same effect as this patch. Seth -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/net/mac80211/scan.c b/net/mac80211/scan.c index a875f74..f982dda 100644 --- a/net/mac80211/scan.c +++ b/net/mac80211/scan.c @@ -699,11 +699,11 @@ static void ieee80211_scan_state_suspend(struct ieee80211_local *local, ieee80211_hw_config(local, IEEE80211_CONF_CHANGE_CHANNEL); /* - * Re-enable vifs and beaconing. Leave PS - * in off-channel state..will put that back - * on-channel at the end of scanning. + * Re-enable vifs and beaconing. Disable PS + * while the scan is suspended for more + * efficient frame rx than with PS-Poll. */ - ieee80211_offchannel_return(local, false); + ieee80211_offchannel_return(local, true); *next_delay = HZ / 5; /* afterwards, resume scan & go to next channel */ @@ -713,8 +713,8 @@ static void ieee80211_scan_state_suspend(struct ieee80211_local *local, static void ieee80211_scan_state_resume(struct ieee80211_local *local, unsigned long *next_delay) { - /* PS already is in off-channel mode */ - ieee80211_offchannel_stop_vifs(local, false); + /* Put PS back in off-channel mode */ + ieee80211_offchannel_stop_vifs(local, true); if (local->ops->flush) { drv_flush(local, false);
Leaving powersave enabled while the scan is suspended requires the use of PS-Poll to retrieve frames buffered at the AP. PS-Poll isn't a very efficient means of data transfer and may not even work with drivers not supporting powersave. These problems can largely be avoided by simply disabling off-channel powersave when the scan is suspended and enabling it again when returning to off-channel operation. The results of this change are fantastic. Previously when using iperf I would see greatly decreased throughput (often dropping to 0 bits/sec for several seconds) and very high packet loss during software scans. This patch completely eliminates the packet loss and reduces the loss in throughput to very reasonable levels. Signed-off-by: Seth Forshee <seth.forshee@canonical.com> --- net/mac80211/scan.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)