Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
Well, confoundable. I have recompiled both cfg80211 and mac80211 with the DEBUGFS flags set to yes, and I'm still not seeing anything under the phy0 directory. At this point, I'm stumped. I cannot locate anything on the internet describing my problem. I'll keep digging, but any help is appreciated. Adam On Fri, Jun 21, 2013 at 11:58 AM, Felix Fietkau n...@openwrt.org wrote: On 2013-06-21 5:41 PM, Adam Allred wrote: I think it's already on in the debian 3.9 kernel .config (CONFIG_MAC80211_DEBUGFS=y), and I do not see a parameter to turn it on via modinfo. Sorry to run down the rabbit hole... Maybe you need CONFIG_CFG80211_DEBUGFS=y as well. There's no runtime part that needs to be enabled, as long as debugfs is mounted. - Felix ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
* On 24.06.2013 04:17 PM, Adam Allred wrote: Well, confoundable. I have recompiled both cfg80211 and mac80211 with the DEBUGFS flags set to yes, and I'm still not seeing anything under the phy0 directory. I do. Albeit with a WPEA-127NI card. Check: CONFIG_CFG80211_DEBUGFS=y CONFIG_MAC80211_DEBUGFS=y CONFIG_ATH9K_DEBUGFS=y smime.p7s Description: S/MIME Cryptographic Signature ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
I've checked all of those. They are all set. Adam On Mon, Jun 24, 2013 at 10:53 AM, Mihai Moldovan io...@ionic.de wrote: * On 24.06.2013 04:17 PM, Adam Allred wrote: Well, confoundable. I have recompiled both cfg80211 and mac80211 with the DEBUGFS flags set to yes, and I'm still not seeing anything under the phy0 directory. I do. Albeit with a WPEA-127NI card. Check: CONFIG_CFG80211_DEBUGFS=y CONFIG_MAC80211_DEBUGFS=y CONFIG_ATH9K_DEBUGFS=y ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
* On 24.06.2013 05:31 PM, Adam Allred wrote: I've checked all of those. They are all set. I do. Albeit with a WPEA-127NI card. Scratch that, it's WPEA-127N exactly. AR9380. Very same card you have. ;) What about... CONFIG_ATH_DEBUG And just for good measure: CONFIG_MAC80211_DEBUG_MENU CONFIG_MAC80211_VERBOSE_DEBUG CONFIG_MAC80211_STA_DEBUG CONFIG_MAC80211_HT_DEBUG CONFIG_MAC80211_IBSS_DEBUG (A word of warning, though: that's noisy!) Can you paste iw list, please? Are you sure phy0 is your ath9k card? smime.p7s Description: S/MIME Cryptographic Signature ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
root@aero:/sys/kernel/debug/ieee80211/phy0# ls -la total 0 drwxr-xr-x 2 root root 0 Jun 24 15:29 . drwxr-xr-x 3 root root 0 Jun 24 15:29 .. root@aero:/sys/kernel/debug/ieee80211/phy0# mount ... debugfs on /sys/kernel/debug type debugfs (rw,relatime) ... root@aero:/sys/kernel/debug/ieee80211/phy0# iw list Wiphy phy0 Band 1: Capabilities: 0x11ef RX LDPC HT20/HT40 SM Power Save disabled RX HT20 SGI ... Yeah, it's the right card (and the only wireless card). I'll jam more debug bits in there and see what happens. On Mon, Jun 24, 2013 at 11:52 AM, Mihai Moldovan io...@ionic.de wrote: * On 24.06.2013 05:31 PM, Adam Allred wrote: I've checked all of those. They are all set. I do. Albeit with a WPEA-127NI card. Scratch that, it's WPEA-127N exactly. AR9380. Very same card you have. ;) What about... CONFIG_ATH_DEBUG And just for good measure: CONFIG_MAC80211_DEBUG_MENU CONFIG_MAC80211_VERBOSE_DEBUG CONFIG_MAC80211_STA_DEBUG CONFIG_MAC80211_HT_DEBUG CONFIG_MAC80211_IBSS_DEBUG (A word of warning, though: that's noisy!) Can you paste iw list, please? Are you sure phy0 is your ath9k card? ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
See my hostapd.conf at the beginning of this thread. I had trouble with video streaming until I dropped in the various wmm_ and tx_ options. Admittedly, I do not know what they all mean, but the internet claims they affect video and audio streams for the better. On Fri, Jun 21, 2013 at 11:03 PM, Paul Farrow at...@thefabfarrows.com wrote: I run a similar set up using a sparklan WPEA-127N on fedora 17 and I am see my stats in /sys/kernel/debug/ieee80211/phy0/netdev:wlan1/stations/54:26:96:cf:1e:a3 but rc_stats is empty. I do have a file called current_tx_rate and the value in there is 117.0 MBit/s After having trouble with my 92xx card I upgraded to the sparklan but have yet to get it going very fast. What I found was that once I enabled the HT config parts all my ios devices couldnt stream youtube but everything else was fine. Perhaps someone could post their hostapd config file so we can compare. Thanks Paul On 2013-06-21 11:58, Felix Fietkau wrote: On 2013-06-21 5:41 PM, Adam Allred wrote: I think it's already on in the debian 3.9 kernel .config (CONFIG_MAC80211_DEBUGFS=y), and I do not see a parameter to turn it on via modinfo. Sorry to run down the rabbit hole... Maybe you need CONFIG_CFG80211_DEBUGFS=y as well. There's no runtime part that needs to be enabled, as long as debugfs is mounted. - Felix ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
[ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
Hello, I am attempting to build an AP that can support airplay streaming between various mac devices. My current test bed is as follows: Macbook Pro 10,2 (13 Retina), with an Airport Extreme (0x14E4, 0x10F) card, which is using the BCM43xx driver (according to system report). Apple TV Gen 3 (no idea what the wireless hardware in it is) A Jetway Mini-top: http://www.newegg.com/Product/Product.aspx?Item=N82E16856107081 It's running fully patched debian stable (wheezy), amd64. I have replaced the stock PCI Express wireless card with a SparkLan WPEA-127N: http://www.sparklan.com/product.php?func=viewprod_id=181 and added two antennas to the top of the case utilizing these: http://www.embeddedworks.net/pdetail.php?mn=assyprod=assy085 http://www.embeddedworks.net/pdetail.php?mn=anteprod=ante542 The darn WPEA-127N came world-roaming burned in to the EEPROM, so I modified the ath9k driver from the stock debian wheezy driver using: https://patchwork.kernel.org/patch/2249511/ (thanks Ben) and for good measure, also updated regulatory.bin for crda to map regdomain 0 to the US (where I'm at). Finally, my modprobe settings for ath9k are: options ath9k override_eeprom_regdomain=0 options ath9k nohwcrypt=1 I have successfully connected to the AP over a 5GHz channel utilizing HT40, and have working network connectivity. My hostapd.conf is as follows: interface=wlan0 driver=nl80211 ssid=aero hw_mode=a channel=44 ieee80211n=1 wmm_enabled=1 wmm_ac_bk_cwmin=4 wmm_ac_bk_cwmax=10 wmm_ac_bk_aifs=7 wmm_ac_bk_txop_limit=0 wmm_ac_bk_acm=0 wmm_ac_be_aifs=3 wmm_ac_be_cwmin=4 wmm_ac_be_cwmax=10 wmm_ac_be_txop_limit=0 wmm_ac_be_acm=0 wmm_ac_vi_aifs=2 wmm_ac_vi_cwmin=3 wmm_ac_vi_cwmax=4 wmm_ac_vi_txop_limit=94 wmm_ac_vi_acm=0 wmm_ac_vo_aifs=2 wmm_ac_vo_cwmin=2 wmm_ac_vo_cwmax=3 wmm_ac_vo_txop_limit=47 wmm_ac_vo_acm=0 tx_queue_data3_aifs=7 tx_queue_data3_cwmin=15 tx_queue_data3_cwmax=1023 tx_queue_data3_burst=0 tx_queue_data2_aifs=3 tx_queue_data2_cwmin=15 tx_queue_data2_cwmax=63 tx_queue_data2_burst=0 tx_queue_data1_aifs=1 tx_queue_data1_cwmin=7 tx_queue_data1_cwmax=15 tx_queue_data1_burst=3.0 tx_queue_data0_aifs=1 tx_queue_data0_cwmin=3 tx_queue_data0_cwmax=7 tx_queue_data0_burst=1.5 ht_capab=[LDPC][HT40+][SHORT-GI-20][SHORT-GI-40][TX-STBC][RX-STBC1][DSSS_CCK-40] #ht_capab=[LDPC][SHORT-GI-20][TX-STBC][RX-STBC1] #ht_capab=[LDPC][TX-STBC][RX-STBC1] macaddr_acl=0 auth_algs=1 ignore_broadcast_ssid=0 wpa=3 wpa_passphrase=[REDACTED] wpa_key_mgmt=WPA-PSK wpa_pairwise=CCMP country_code=US the various wmm and tx_queue parts are mostly unknown to me, as I came across them trying to determine why my airplay stream would not start if the encryption scheme was CCMP. After setting the above, it would start when the scheme was set to CCMP. As of right now, I get about 11MB/s (88Mbit/s), and that seems low. I was hoping to see something closer 200-250Mbit, but I don't really know what to expect between a published maximum (which I currently believe to be 450Mbit for this card), and a practical maximum. So, my question is...am I doing it wrong? Also, since the wmm and tx settings are mostly unknown to me, I've been trying to figure out what they mean, but have not located any useful resources. If anyone could point me in that direction, I may find that these settings are completely wrong for my hardware. Finally, if this is the completely wrong forum to ask this question, I apologize. This is my first dive into a linux AP, and I'm mostly still stumbling around blindly. Thanks, Adam ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
On 2013-06-20 6:05 PM, Adam Allred wrote: Hello, I am attempting to build an AP that can support airplay streaming between various mac devices. My current test bed is as follows: Macbook Pro 10,2 (13 Retina), with an Airport Extreme (0x14E4, 0x10F) card, which is using the BCM43xx driver (according to system report). Apple TV Gen 3 (no idea what the wireless hardware in it is) A Jetway Mini-top: http://www.newegg.com/Product/Product.aspx?Item=N82E16856107081 It's running fully patched debian stable (wheezy), amd64. I have replaced the stock PCI Express wireless card with a SparkLan WPEA-127N: http://www.sparklan.com/product.php?func=viewprod_id=181 and added two antennas to the top of the case utilizing these: http://www.embeddedworks.net/pdetail.php?mn=assyprod=assy085 http://www.embeddedworks.net/pdetail.php?mn=anteprod=ante542 The darn WPEA-127N came world-roaming burned in to the EEPROM, so I modified the ath9k driver from the stock debian wheezy driver using: https://patchwork.kernel.org/patch/2249511/ (thanks Ben) and for good measure, also updated regulatory.bin for crda to map regdomain 0 to the US (where I'm at). Finally, my modprobe settings for ath9k are: options ath9k override_eeprom_regdomain=0 options ath9k nohwcrypt=1 Using nohwcrypt=1 is a bad idea and might explain your performance issues. Where did you get that from? - Felix ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
I picked it up off of http://ubuntuforums.org/showthread.php?t=1746326 and only tried it after I had a working connection, just to see if it would make a difference, since it's ubuntu and a AR92xx series card in that post. Performance is the same both with and without this option. Adam On Thu, Jun 20, 2013 at 12:44 PM, Felix Fietkau n...@openwrt.org wrote: On 2013-06-20 6:05 PM, Adam Allred wrote: Hello, I am attempting to build an AP that can support airplay streaming between various mac devices. My current test bed is as follows: Macbook Pro 10,2 (13 Retina), with an Airport Extreme (0x14E4, 0x10F) card, which is using the BCM43xx driver (according to system report). Apple TV Gen 3 (no idea what the wireless hardware in it is) A Jetway Mini-top: http://www.newegg.com/Product/Product.aspx?Item=N82E16856107081 It's running fully patched debian stable (wheezy), amd64. I have replaced the stock PCI Express wireless card with a SparkLan WPEA-127N: http://www.sparklan.com/product.php?func=viewprod_id=181 and added two antennas to the top of the case utilizing these: http://www.embeddedworks.net/pdetail.php?mn=assyprod=assy085 http://www.embeddedworks.net/pdetail.php?mn=anteprod=ante542 The darn WPEA-127N came world-roaming burned in to the EEPROM, so I modified the ath9k driver from the stock debian wheezy driver using: https://patchwork.kernel.org/patch/2249511/ (thanks Ben) and for good measure, also updated regulatory.bin for crda to map regdomain 0 to the US (where I'm at). Finally, my modprobe settings for ath9k are: options ath9k override_eeprom_regdomain=0 options ath9k nohwcrypt=1 Using nohwcrypt=1 is a bad idea and might explain your performance issues. Where did you get that from? - Felix ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] performance with ath9k, sparklan WPEA-127N, and Macbook pro
On 2013-06-20 6:49 PM, Adam Allred wrote: I picked it up off of http://ubuntuforums.org/showthread.php?t=1746326 and only tried it after I had a working connection, just to see if it would make a difference, since it's ubuntu and a AR92xx series card in that post. Performance is the same both with and without this option. OK. Better leave it out from now on, it's not a good idea to use random workarounds for old issues on a different wifi chip :) I think you should start by using more recent drivers, Linux 3.2 is a bit old. Pick up a recent compat-drivers release - https://backports.wiki.kernel.org/index.php/Releases Also, how exactly are you measuring performance? Once you have the new release installed and are running some performance tests, it would be helpful to post the contents of /sys/kernel/debug/ieee80211/phy0/netdev\:wlan0/stations/*/rc_stats It'll show how well the individual 802.11n rates are working. - Felix ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel