Return-path: Received: from mail-iw0-f174.google.com ([209.85.214.174]:41368 "EHLO mail-iw0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751954Ab1DFAhr (ORCPT ); Tue, 5 Apr 2011 20:37:47 -0400 Received: by iwn34 with SMTP id 34so925662iwn.19 for ; Tue, 05 Apr 2011 17:37:47 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <4D9B5473.6040003@lwfinger.net> References: <4D9B5473.6040003@lwfinger.net> From: Gottfried Haider Date: Wed, 6 Apr 2011 02:37:27 +0200 Message-ID: Subject: Re: rtl8192ce: odd ping behavior To: Larry Finger Cc: linux-wireless@vger.kernel.org Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: Hi Larry, > I repeated your test. My ping results were much better than yours: > > I do not see the deauthentication/authentication results that you see. Those > would definitely cause network delays. > > There are one difference in our systems. I have a 10ec:8178 device, but I > have an 8176 here that I will try now. I have also received a new version of > the driver from Realtek. If my 8176 has the same problems as yours, perhaps > that will help. Thanks for testing. I had hoped that the ping results could make a light go on somewhere.. I am only starting to get into rolling my own kernels again - but if I can test something, please let me know. One thing that appears to me as being broken with this driver is the LPS mode. I was watching the debug messages go by while having another terminal continuously ping my WiFi router: [ 3751.605727] rtl8192c_common:rtl92c_dm_txpower_tracking_callback_thermalmeter():<0-0> Readback Thermal Meter = 0x10 pre thermal meter 0xf eeprom_thermalmeter 0xf delta 0x1 delta_lck 0x0 delta_iqk 0x2 [ 3751.605744] rtl8192c_common:rtl92c_dm_txpower_tracking_callback_thermalmeter():<0-0> <=== [ 3751.605757] rtlwifi:rtl_lps_enter():<0-0> Enter 802.11 power save mode... [ 3751.605767] rtlwifi:rtl_lps_set_psmode():<0-0> FW LPS enter ps_mode:3 [ 3751.605778] rtl8192c_common:rtl92c_set_fw_pwrmode_cmd():<0-0> FW LPS mode = 3 [ 3751.605787] rtl8192c_common: In process "kworker/0:1" (pid 2323):rtl92c_set_fw_rsvdpagepkt(): u1_h2c_set_pwrmode [ 3751.605799] 03 01 05 [ 3751.605806] [ 3751.605812] rtl8192c_common:_rtl92c_fill_h2c_command():<0-0> come in [ 3751.606120] rtl8192c_common:_rtl92c_fill_h2c_command():<0-0> Write element_id box_reg( 1dc) = 1 [ 3751.606132] rtl8192c_common:_rtl92c_fill_h2c_command():<0-0> pHalData->last_hmeboxnum = 0 [ 3751.606143] rtl8192c_common:_rtl92c_fill_h2c_command():<0-0> go out (then every second a) [ 3751.789003] rtl8192ce:rtl92ce_led_control():<200-1> ledaction 4, (until) [ 3856.004637] rtlwifi:rtl_op_sw_scan_start():<0-0> [ 3856.004655] rtlwifi:rtl_lps_leave():<0-0> Busy Traffic,Leave 802.11 power save.. [ 3856.004666] rtlwifi:rtl_lps_set_psmode():<0-0> FW LPS leave ps_mode:0 [ 3856.004725] rtl8192c_common:rtl92c_set_fw_pwrmode_cmd():<0-0> FW LPS mode = 0 and when I afterwards looked at the second terminal I saw that the pings were just being replied to again - and before I had a long list of failed ping requests. Now maybe this is policy-wise totally the expected behavior - waiting for a storm till leaving the power saving mode - but for debugging it would be good to iron this out, as it's a bit hard to track when it kicks in (and it's obviously _not_ completely transparent to the user). On the ThinkPad Wiki, LPS is also mentioned as a source of problems [1] - though I don't know how similar the in-kernel driver is to the version that the text refers to. send trough my USB-tethered phone ;) Gottfried [1] http://www.thinkwiki.org/wiki/ThinkPad_11b/g/n_Wireless_LAN_Mini-PCI_Express_Adapter_II