A-MSDU reception not working?
Janusz Dziedzic
janusz.dziedzic at tieto.com
Wed Jul 9 00:39:21 PDT 2014
On 9 July 2014 08:09, Denton Gentry <denton.gentry at gmail.com> wrote:
> I ran iperf on an 802.11ac MacBook Air, through my ath10 AP, and to an
> Ubuntu system connected to the AP via Ethernet. The AP was running the
> reorder buffer patches, the patch to make A-MSDU bypass the reorder
> buffer, and the patch to make amsdu aggregation configurable via
> debugfs.
>
> http://lists.infradead.org/pipermail/ath10k/2014-June/002551.html
> http://lists.infradead.org/pipermail/ath10k/2014-June/002553.html
> http://lists.infradead.org/pipermail/ath10k/2014-July/002597.html
> http://permalink.gmane.org/gmane.linux.kernel.wireless.general/124128
>
> It was notably not running the patch from earlier in this thread to
> copy A-MSDU subframes into one skb.
>
>
> I used a second ath10k AP configured in monitor mode as a Wifi
> sniffer. I also gathered pcaps on the MacBook Air and on the Ubuntu
> system. Wireshark on the MacBook cannot gather radiotap headers nor
> put the interface into monitor mode, so we see only Ethernet frames
> and we don't see retransmissions on the Wifi link.
>
> I've attached decodes of the same two packets as captured by the
> MacBook, by the sniffer, and by the Ubuntu server.
>
> The MacBook sends two frames, TCP sequence numbers 3065441 and 3066889
> (the Wireshark "relative sequence number"). In the Wifi sniffer we see
> these being aggregated as two subframes in an A-MSDU frame. The first
> time this A-MSDU is sent it is corrupted in the air. We see the TCP
> checksum of one of the subframes is wrong, and a few bytes of the
> payload replaced with ">%'g9(1$".
>
> The A-MSDU is retransmitted a short time later, and the second time
> the TCP checksum of both subframes is correct.
>
> This is all fine so far.
>
> However in the pcap taken from the Ubuntu server, we see TCP sequence
> number 3065441 being delivered *twice.* The first time, the TCP
> checksum is wrong. The second time the TCP checksum is correct.
>
> To me, it looks like A-MSDU frames with bad FCS are not being
> discarded after ieee80211_rx_monitor(). The corrupted frames are being
> delivered. Delivering the corrupted frames results in sending more TCP
> Dup ACKs for the same sequence number back to the MacBook, and I think
> this is what causes the MacBook to decide there is congestion and slow
> down.
>
OK, to be sure this is the main issue we can just drop frames with wrong FCS.
Could you check this:
@@ -1267,10 +1267,10 @@ static void ath10k_htt_rx_handler(struct
ath10k_htt *htt,
continue;
}
- if (attention & RX_ATTENTION_FLAGS_FCS_ERR)
- rx_status->flag |= RX_FLAG_FAILED_FCS_CRC;
- else
- rx_status->flag &= ~RX_FLAG_FAILED_FCS_CRC;
+ if (attention & RX_ATTENTION_FLAGS_FCS_ERR) {
+ ath10k_htt_rx_free_msdu_chain(msdu_head);
+ continue;
+ }
if (attention & RX_ATTENTION_FLAGS_TKIP_MIC_ERR)
If will help we will have to check how to handle monitor + ap case
correctly, one of idea is to not drop frames with wrong FCS only in
case of standalone monitor mode.
But, please try this patch to be sure.
BR
Janusz
> One note: the wifi sniffer does not show the same corruption to the
> same packet as the pcap from the Ubuntu system shows. I think that is
> normal: the sniffer won't see precisely the same noise, won't have
> precisely the same receive sensitivity, and its antennas are not
> pointing in the same direction as the primary AP. If I do this test
> again, I'll gather a pcap on the primary AP as well to compare to what
> we see on the Ubuntu system.
>
>
> On Tue, Jul 8, 2014 at 12:29 AM, Janusz Dziedzic
> <janusz.dziedzic at tieto.com> wrote:
>> On 8 July 2014 09:02, Janusz Dziedzic <janusz.dziedzic at tieto.com> wrote:
>>> On 8 July 2014 08:50, Janusz Dziedzic <janusz.dziedzic at tieto.com> wrote:
>>>> On 8 July 2014 08:43, Denton Gentry <denton.gentry at gmail.com> wrote:
>>>>> I think I know what is happening now, though I've no idea why. The
>>>>> throughput is low because we have many TCP retransmissions. We have
>>>>> retransmissions because the TCP checksum is wrong on a number of
>>>>> frames, and I do find data corruption in the payload so the checksum
>>>>> definitely should be wrong. All of the corrupted frames were
>>>>> originally one of the subframes in an A-MSDU packet.
>>>>>
>>>>> An example follows at the end of this message, as dissected by
>>>>> Wireshark. iperf sends a very regular data pattern of "0123456789..."
>>>>> over and over. Note how in subframe #2 offset 0x1e0 the bytes "32 33
>>>>> 34" have been replaced by "72 36 35"
>>>>>
>>>>> 01e0 32 33 34 35 36 37 38 39 30 31 72 36 35 35 36 37 2345678901r65567
>>>>>
>>>>> I added printks at the bottom of ath10k_htt_rx_amsdu immediately
>>>>> before the call to ath10k_process_rx. I found this same packet, and we
>>>>> see the "72 36 35" corruption in the printk. So I think it happened in
>>>>> ath10k_process_rx or before, not anything weird after passing it up to
>>>>> mac80211.
>>>>>
>>>>> [ 101.863712] ath10k: 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30
>>>>> [ 101.863727] ath10k: 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36
>>>>> [ 101.863742] ath10k: 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32
>>>>> [ 101.863757] ath10k: 33 34 35 36 37 38 39 30 31 72 36 35 35 36 37 38
>>>>> [ 101.863773] ath10k: 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34
>>>>> [ 101.863788] ath10k: 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30
>>>>> [ 101.863803] ath10k: 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36
>>>>>
>>>>>
>>>>> I've found a number of examples of similar corruption, always with
>>>>> between one and four bytes replaced.
>>>>>
>>>>> 35363738 -> e52c6e07
>>>>> 3435 -> b43f
>>>>> 3839 -> c238
>>>>> 31 -> 7f
>>>>> 3435 -> 7436
>>>>> 30 -> 50
>>>>> 3233 -> bc37
>>>>>
>>>> Seems this could be because of:
>>>>
>>>> + /* cfg80211 expect this padding */
>>>> + padding = (4 - (skb->len + sizeof(subframe_hdr))) & 0x3;
>>>> + skb_put(skb, padding);
>>>>
>>>
>>> BTW, when AP (ath10k) send TCP ACK - are this packets also AMSDU?
>>> I see in my case AP can send 3 x A-MSDU (with total size 304 bytes)
>>> small frames aggregated. Maybe your HW have problems with that.
>>> As I remember correctly someone some time ago report problems with
>>> MacBook pro retina but I am not sure this is the same, while no one
>>> tests the fix.
>>>
>>>>>
>>>>> The packet described above, dissected by Wireshark:
>>>>>
>>>>> No. Time Source Destination
>>>>> Protocol Length Info
>>>>> 2235 18.953349 192.168.144.79 192.168.144.13 TCP
>>>>> 3112 52697 > 5001 [ACK] Seq=1391553 Ack=1 Win=131760 [TCP CHECKSUM
>>>>> INCORRECT] Len=1448 TSval=1298580657 TSecr=4294947481
>>>>>
>>>>> Frame 2235: 3112 bytes on wire (24896 bits), 3112 bytes captured (24896 bits)
>>>>> Encapsulation type: IEEE 802.11 plus radiotap radio header (23)
>>>>> Arrival Time: Jul 7, 2014 23:03:37.763365000 PDT
>>>>> [Time shift for this packet: 0.000000000 seconds]
>>>>> Epoch Time: 1404799417.763365000 seconds
>>>>> [Time delta from previous captured frame: 0.003476000 seconds]
>>>>> [Time delta from previous displayed frame: 0.515641000 seconds]
>>>>> [Time since reference or first frame: 18.953349000 seconds]
>>>>> Frame Number: 2235
>>>>> Frame Length: 3112 bytes (24896 bits)
>>>>> Capture Length: 3112 bytes (24896 bits)
>>>>> [Frame is marked: False]
>>>>> [Frame is ignored: False]
>>>>> [Protocols in frame: radiotap:wlan:llc:ip:tcp:data:llc:ip:tcp:data]
>>>>> [Coloring Rule Name: TCP]
>>>>> [Coloring Rule String: tcp]
>>>>> Radiotap Header v0, Length 38
>>>>> Header revision: 0
>>>>> Header pad: 0
>>>>> Header length: 38
>>>>> Present flags
>>>>> .... .... .... .... .... .... .... ...1 = TSFT: True
>>>>> .... .... .... .... .... .... .... ..1. = Flags: True
>>>>> .... .... .... .... .... .... .... .0.. = Rate: False
>>>>> .... .... .... .... .... .... .... 1... = Channel: True
>>>>> .... .... .... .... .... .... ...0 .... = FHSS: False
>>>>> .... .... .... .... .... .... ..1. .... = dBm Antenna Signal: True
>>>>> .... .... .... .... .... .... .0.. .... = dBm Antenna Noise: False
>>>>> .... .... .... .... .... .... 0... .... = Lock Quality: False
>>>>> .... .... .... .... .... ...0 .... .... = TX Attenuation: False
>>>>> .... .... .... .... .... ..0. .... .... = dB TX Attenuation: False
>>>>> .... .... .... .... .... .0.. .... .... = dBm TX Power: False
>>>>> .... .... .... .... .... 1... .... .... = Antenna: True
>>>>> .... .... .... .... ...0 .... .... .... = dB Antenna Signal: False
>>>>> .... .... .... .... ..0. .... .... .... = dB Antenna Noise: False
>>>>> .... .... .... .... .1.. .... .... .... = RX flags: True
>>>>> .... .... .... .0.. .... .... .... .... = Channel+: False
>>>>> .... .... .... 0... .... .... .... .... = HT information: False
>>>>> .... .... ...0 .... .... .... .... .... = A-MPDU Status: False
>>>>> .... .... ..1. .... .... .... .... .... = VHT information: True
>>>>> ...0 0000 00.. .... .... .... .... .... = Reserved: 0x00000000
>>>>> ..0. .... .... .... .... .... .... .... = Radiotap NS next: False
>>>>> .0.. .... .... .... .... .... .... .... = Vendor NS next: False
>>>>> 0... .... .... .... .... .... .... .... = Ext: False
>>>>> MAC timestamp: 78051063
>>>>> Flags: 0x00
>>>>> .... ...0 = CFP: False
>>>>> .... ..0. = Preamble: Long
>>>>> .... .0.. = WEP: False
>>>>> .... 0... = Fragmentation: False
>>>>> ...0 .... = FCS at end: False
>>>>> ..0. .... = Data Pad: False
>>>>> .0.. .... = Bad FCS: False
>>>>> 0... .... = Short GI: False
>>>>> Channel frequency: 5745 [A 149]
>>>>> Channel type: 802.11a (0x0140)
>>>>> .... .... ...0 .... = Turbo: False
>>>>> .... .... ..0. .... = Complementary Code Keying (CCK): False
>>>>> .... .... .1.. .... = Orthogonal Frequency-Division
>>>>> Multiplexing (OFDM): True
>>>>> .... .... 0... .... = 2 GHz spectrum: False
>>>>> .... ...1 .... .... = 5 GHz spectrum: True
>>>>> .... ..0. .... .... = Passive: False
>>>>> .... .0.. .... .... = Dynamic CCK-OFDM: False
>>>>> .... 0... .... .... = Gaussian Frequency Shift Keying (GFSK): False
>>>>> ...0 .... .... .... = GSM (900MHz): False
>>>>> ..0. .... .... .... = Static Turbo: False
>>>>> .0.. .... .... .... = Half Rate Channel (10MHz Channel Width): False
>>>>> 0... .... .... .... = Quarter Rate Channel (5MHz Channel Width): False
>>>>> SSI Signal: -53 dBm
>>>>> Antenna: 0
>>>>> RX flags: 0x0000
>>>>> .... .... .... .... .... ..0. = Bad PLCP: False
>>>>> VHT information
>>>>> Known VHT information: 0x44
>>>>> .... .... .... ...0 = STBC: False
>>>>> .... .... .... ..0. = TXOP_PS_NOT_ALLOWED: False
>>>>> .... .... .... .1.. = Guard interval: True
>>>>> .... .... .... 0... = SGI Nsym disambiguation: False
>>>>> .... .... ...0 .... = LDPC extra OFDM symbol: False
>>>>> .... .... ..0. .... = Beamformed: False
>>>>> .... .... .1.. .... = Bandwidth: True
>>>>> .... .... 0... .... = Group ID: False
>>>>> .... ...0 .... .... = Partial AID: False
>>>>> .... .0.. = Guard interval: long (0)
>>>>> Bandwidth: 80 MHz (4)
>>>>> User 0: MCS 8
>>>>> 1000 .... = MCS index 0: 8 (256-QAM 3/4)
>>>>> .... 0010 = Spatial streams 0: 2
>>>>> Space-time streams 0: 2
>>>>> Coding 0: BCC (0)
>>>>> [Data Rate: 702.0 Mb/s]
>>>>> IEEE 802.11 QoS Data, Flags: .......T
>>>>> Type/Subtype: QoS Data (0x28)
>>>>> Frame Control Field: 0x8801
>>>>> .... ..00 = Version: 0
>>>>> .... 10.. = Type: Data frame (2)
>>>>> 1000 .... = Subtype: 8
>>>>> Flags: 0x01
>>>>> .... ..01 = DS status: Frame from STA to DS via an AP (To
>>>>> DS: 1 From DS: 0) (0x01)
>>>>> .... .0.. = More Fragments: This is the last fragment
>>>>> .... 0... = Retry: Frame is not being retransmitted
>>>>> ...0 .... = PWR MGT: STA will stay up
>>>>> ..0. .... = More Data: No data buffered
>>>>> .0.. .... = Protected flag: Data is not protected
>>>>> 0... .... = Order flag: Not strictly ordered
>>>>> .000 0000 0011 0000 = Duration: 48 microseconds
>>>>> Receiver address: SenaoNet_18:a8:00 (88:dc:96:18:a8:00)
>>>>> BSS Id: SenaoNet_18:a8:00 (88:dc:96:18:a8:00)
>>>>> Transmitter address: Apple_67:24:54 (84:38:35:67:24:54)
>>>>> Source address: Apple_67:24:54 (84:38:35:67:24:54)
>>>>> Destination address: SenaoNet_18:a8:00 (88:dc:96:18:a8:00)
>>>>> Fragment number: 0
>>>>> Sequence number: 1021
>>>>> Qos Control: 0x0080
>>>>> .... .... .... 0000 = TID: 0
>>>>> [.... .... .... .000 = Priority: Best Effort (Best Effort) (0)]
>>>>> .... .... ...0 .... = QoS bit 4: Bits 8-15 of QoS Control
>>>>> field are TXOP Duration Requested
>>>>> .... .... .00. .... = Ack Policy: Normal Ack (0x0000)
>>>>> .... .... 1... .... = Payload Type: A-MSDU
>>>>> 0000 0000 .... .... = TXOP Duration Requested: 0 (no TXOP requested)
>>>>> IEEE 802.11 Aggregate MSDU
>>>>> A-MSDU Subframe #1
>>>>> Destination address: GoogleFi_00:14:cd (f8:8f:ca:00:14:cd)
>>>>> Source address: Apple_67:24:54 (84:38:35:67:24:54)
>>>>> A-MSDU Length: 1510
>>>>> Logical-Link Control
>>>>> DSAP: SNAP (0xaa)
>>>>> IG Bit: Individual
>>>>> SSAP: SNAP (0xaa)
>>>>> CR Bit: Command
>>>>> Control field: U, func=UI (0x03)
>>>>> 000. 00.. = Command: Unnumbered Information (0x00)
>>>>> .... ..11 = Frame type: Unnumbered frame (0x03)
>>>>> Organization Code: Encapsulated Ethernet (0x000000)
>>>>> Type: IP (0x0800)
>>>>> Internet Protocol Version 4, Src: 192.168.144.79
>>>>> (192.168.144.79), Dst: 192.168.144.13 (192.168.144.13)
>>>>> Version: 4
>>>>> Header length: 20 bytes
>>>>> Differentiated Services Field: 0x00 (DSCP 0x00: Default;
>>>>> ECN: 0x00: Not-ECT (Not ECN-Capable Transport))
>>>>> 0000 00.. = Differentiated Services Codepoint: Default (0x00)
>>>>> .... ..00 = Explicit Congestion Notification: Not-ECT
>>>>> (Not ECN-Capable Transport) (0x00)
>>>>> Total Length: 1500
>>>>> Identification: 0xc622 (50722)
>>>>> Flags: 0x00
>>>>> 0... .... = Reserved bit: Not set
>>>>> .0.. .... = Don't fragment: Not set
>>>>> ..0. .... = More fragments: Not set
>>>>> Fragment offset: 0
>>>>> Time to live: 64
>>>>> Protocol: TCP (6)
>>>>> Header checksum: 0x0d4c [correct]
>>>>> [Good: True]
>>>>> [Bad: False]
>>>>> Source: 192.168.144.79 (192.168.144.79)
>>>>> Destination: 192.168.144.13 (192.168.144.13)
>>>>> [Source GeoIP: Unknown]
>>>>> [Destination GeoIP: Unknown]
>>>>> Transmission Control Protocol, Src Port: 52697 (52697), Dst
>>>>> Port: 5001 (5001), Seq: 1390105, Ack: 1, Len: 1448
>>>>> Source port: 52697 (52697)
>>>>> Destination port: 5001 (5001)
>>>>> [Stream index: 0]
>>>>> Sequence number: 1390105 (relative sequence number)
>>>>> [Next sequence number: 1391553 (relative sequence number)]
>>>>> Acknowledgment number: 1 (relative ack number)
>>>>> Header length: 32 bytes
>>>>> Flags: 0x010 (ACK)
>>>>> 000. .... .... = Reserved: Not set
>>>>> ...0 .... .... = Nonce: Not set
>>>>> .... 0... .... = Congestion Window Reduced (CWR): Not set
>>>>> .... .0.. .... = ECN-Echo: Not set
>>>>> .... ..0. .... = Urgent: Not set
>>>>> .... ...1 .... = Acknowledgment: Set
>>>>> .... .... 0... = Push: Not set
>>>>> .... .... .0.. = Reset: Not set
>>>>> .... .... ..0. = Syn: Not set
>>>>> .... .... ...0 = Fin: Not set
>>>>> Window size value: 8235
>>>>> [Calculated window size: 131760]
>>>>> [Window size scaling factor: 16]
>>>>> Checksum: 0xa1c0 [correct]
>>>>> [Good Checksum: True]
>>>>> [Bad Checksum: False]
>>>>> Options: (12 bytes), No-Operation (NOP), No-Operation
>>>>> (NOP), Timestamps
>>>>> No-Operation (NOP)
>>>>> Type: 1
>>>>> 0... .... = Copy on fragmentation: No
>>>>> .00. .... = Class: Control (0)
>>>>> ...0 0001 = Number: No-Operation (NOP) (1)
>>>>> No-Operation (NOP)
>>>>> Type: 1
>>>>> 0... .... = Copy on fragmentation: No
>>>>> .00. .... = Class: Control (0)
>>>>> ...0 0001 = Number: No-Operation (NOP) (1)
>>>>> Timestamps: TSval 1298580657, TSecr 4294947481
>>>>> Kind: Timestamp (8)
>>>>> Length: 10
>>>>> Timestamp value: 1298580657
>>>>> Timestamp echo reply: 4294947481
>>>>> Data (1448 bytes)
>>>>>
>>>>> 0000 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0010 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0020 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0030 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0040 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0050 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0060 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0070 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0080 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0090 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 00a0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 00b0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 00c0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 00d0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 00e0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 00f0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0100 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0110 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0120 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0130 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0140 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0150 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0160 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0170 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0180 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0190 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 01a0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 01b0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 01c0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 01d0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 01e0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 01f0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0200 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0210 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0220 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0230 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0240 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0250 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0260 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0270 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0280 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0290 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 02a0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 02b0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 02c0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 02d0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 02e0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 02f0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0300 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0310 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0320 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0330 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0340 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0350 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0360 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0370 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0380 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0390 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 03a0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 03b0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 03c0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 03d0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 03e0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 03f0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0400 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0410 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0420 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0430 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0440 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0450 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0460 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0470 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0480 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0490 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 04a0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 04b0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 04c0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 04d0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 04e0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 04f0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0500 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0510 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0520 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0530 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0540 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0550 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0560 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0570 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0580 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0590 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 05a0 34 35 36 37 38 39 30 31 45678901
>>>>> Data: 343536373839303132333435363738393031323334353637...
>>>>> [Length: 1448]
>>>>> A-MSDU Subframe #2
>>>>> Destination address: GoogleFi_00:14:cd (f8:8f:ca:00:14:cd)
>>>>> Source address: Apple_67:24:54 (84:38:35:67:24:54)
>>>>> A-MSDU Length: 1510
>>>>> Logical-Link Control
>>>>> DSAP: SNAP (0xaa)
>>>>> IG Bit: Individual
>>>>> SSAP: SNAP (0xaa)
>>>>> CR Bit: Command
>>>>> Control field: U, func=UI (0x03)
>>>>> 000. 00.. = Command: Unnumbered Information (0x00)
>>>>> .... ..11 = Frame type: Unnumbered frame (0x03)
>>>>> Organization Code: Encapsulated Ethernet (0x000000)
>>>>> Type: IP (0x0800)
>>>>> Internet Protocol Version 4, Src: 192.168.144.79
>>>>> (192.168.144.79), Dst: 192.168.144.13 (192.168.144.13)
>>>>> Version: 4
>>>>> Header length: 20 bytes
>>>>> Differentiated Services Field: 0x00 (DSCP 0x00: Default;
>>>>> ECN: 0x00: Not-ECT (Not ECN-Capable Transport))
>>>>> 0000 00.. = Differentiated Services Codepoint: Default (0x00)
>>>>> .... ..00 = Explicit Congestion Notification: Not-ECT
>>>>> (Not ECN-Capable Transport) (0x00)
>>>>> Total Length: 1500
>>>>> Identification: 0xda09 (55817)
>>>>> Flags: 0x00
>>>>> 0... .... = Reserved bit: Not set
>>>>> .0.. .... = Don't fragment: Not set
>>>>> ..0. .... = More fragments: Not set
>>>>> Fragment offset: 0
>>>>> Time to live: 64
>>>>> Protocol: TCP (6)
>>>>> Header checksum: 0xf964 [correct]
>>>>> [Good: True]
>>>>> [Bad: False]
>>>>> Source: 192.168.144.79 (192.168.144.79)
>>>>> Destination: 192.168.144.13 (192.168.144.13)
>>>>> [Source GeoIP: Unknown]
>>>>> [Destination GeoIP: Unknown]
>>>>> Transmission Control Protocol, Src Port: 52697 (52697), Dst
>>>>> Port: 5001 (5001), Seq: 1391553, Ack: 1, Len: 1448
>>>>> Source port: 52697 (52697)
>>>>> Destination port: 5001 (5001)
>>>>> [Stream index: 0]
>>>>> Sequence number: 1391553 (relative sequence number)
>>>>> [Next sequence number: 1393001 (relative sequence number)]
>>>>> Acknowledgment number: 1 (relative ack number)
>>>>> Header length: 32 bytes
>>>>> Flags: 0x010 (ACK)
>>>>> 000. .... .... = Reserved: Not set
>>>>> ...0 .... .... = Nonce: Not set
>>>>> .... 0... .... = Congestion Window Reduced (CWR): Not set
>>>>> .... .0.. .... = ECN-Echo: Not set
>>>>> .... ..0. .... = Urgent: Not set
>>>>> .... ...1 .... = Acknowledgment: Set
>>>>> .... .... 0... = Push: Not set
>>>>> .... .... .0.. = Reset: Not set
>>>>> .... .... ..0. = Syn: Not set
>>>>> .... .... ...0 = Fin: Not set
>>>>> Window size value: 8235
>>>>> [Calculated window size: 131760]
>>>>> [Window size scaling factor: 16]
>>>>> Checksum: 0x9a16 [incorrect, should be 0x5913 (maybe
>>>>> caused by "TCP checksum offload"?)]
>>
>> And yes we are using checksum offload in ath10k.
>> Best is using standalone 80211ac sniffer for that case to be sure
>> about checksum.
>>
>>>>> [Good Checksum: False]
>>>>> [Bad Checksum: True]
>>>>> [Expert Info (Error/Checksum): Bad checksum]
>>>>> [Message: Bad checksum]
>>>>> [Severity level: Error]
>>>>> [Group: Checksum]
>>>>> Options: (12 bytes), No-Operation (NOP), No-Operation
>>>>> (NOP), Timestamps
>>>>> No-Operation (NOP)
>>>>> Type: 1
>>>>> 0... .... = Copy on fragmentation: No
>>>>> .00. .... = Class: Control (0)
>>>>> ...0 0001 = Number: No-Operation (NOP) (1)
>>>>> No-Operation (NOP)
>>>>> Type: 1
>>>>> 0... .... = Copy on fragmentation: No
>>>>> .00. .... = Class: Control (0)
>>>>> ...0 0001 = Number: No-Operation (NOP) (1)
>>>>> Timestamps: TSval 1298580657, TSecr 4294947481
>>>>> Kind: Timestamp (8)
>>>>> Length: 10
>>>>> Timestamp value: 1298580657
>>>>> Timestamp echo reply: 4294947481
>>>>> Data (1448 bytes)
>>>>>
>>>>> 0000 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0010 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0020 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0030 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0040 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0050 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0060 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0070 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0080 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0090 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 00a0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 00b0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 00c0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 00d0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 00e0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 00f0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0100 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0110 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0120 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0130 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0140 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0150 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0160 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0170 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0180 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0190 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 01a0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 01b0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 01c0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 01d0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 01e0 32 33 34 35 36 37 38 39 30 31 72 36 35 35 36 37 2345678901r65567
>>>>> 01f0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0200 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0210 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0220 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0230 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0240 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0250 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0260 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0270 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0280 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0290 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 02a0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 02b0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 02c0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 02d0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 02e0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 02f0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0300 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0310 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0320 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0330 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0340 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0350 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0360 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0370 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0380 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0390 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 03a0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 03b0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 03c0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 03d0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 03e0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 03f0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0400 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0410 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0420 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0430 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0440 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0450 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0460 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0470 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0480 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0490 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 04a0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 04b0 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 04c0 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 04d0 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 04e0 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 04f0 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0500 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0510 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0520 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0530 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0540 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 0550 32 33 34 35 36 37 38 39 30 31 32 33 34 35 36 37 2345678901234567
>>>>> 0560 38 39 30 31 32 33 34 35 36 37 38 39 30 31 32 33 8901234567890123
>>>>> 0570 34 35 36 37 38 39 30 31 32 33 34 35 36 37 38 39 4567890123456789
>>>>> 0580 30 31 32 33 34 35 36 37 38 39 30 31 32 33 34 35 0123456789012345
>>>>> 0590 36 37 38 39 30 31 32 33 34 35 36 37 38 39 30 31 6789012345678901
>>>>> 05a0 32 33 34 35 36 37 38 39 23456789
>>>>> Data: 323334353637383930313233343536373839303132333435...
>>>>> [Length: 1448]
>>>>>
>>>>> On Mon, Jul 7, 2014 at 12:26 PM, Denton Gentry <denton.gentry at gmail.com> wrote:
>>>>>> The initial results are not promising: a MacOS 802.11ac client gets
>>>>>> between 0-2 Mbps with this change, where it was getting about 8 Mbps
>>>>>> prior to this change and ~170 Mbps prior to the reordering fix. A pcap
>>>>>> from the receiving system shows a very large number of out of order
>>>>>> frames, likely due to TCP retransmission.
>>>>>>
>>>>>> An 802.11n MacBook gets very good throughput, only the 802.11ac
>>>>>> MacBook shows the very poor result. I'm trying to figure out why.
>>>>>>
>>>>>>
>>>>>> One specific note (and probably not related to the throughput): I
>>>>>> believe ath10k_htt_rx_amsdu runs in the tasklet, which means it would
>>>>>> need to use GFP_ATOMIC instead of GFP_KERNEL. Do I understand it
>>>>>> correctly?
>>>>>>
>>>>>>
>>>>>> On Mon, Jul 7, 2014 at 1:30 AM, Janusz Dziedzic
>>>>>> <janusz.dziedzic at tieto.com> wrote:
>>>>>>> On 6 July 2014 04:27, Adrian Chadd <adrian at freebsd.org> wrote:
>>>>>>>> I think you may have to tell mac80211 that it's okay and not to drop the frames.
>>>>>>>>
>>>>>>>> The earlier atheros chips would just give you the AMSDU frames as
>>>>>>>> deaggregated A-MPDU sub-frames - you'd just pass the A-MSDU up to
>>>>>>>> net80211 and it'd ull it apart. But if the driver is doing it (or,
>>>>>>>> well, the chip is doing it) then mac80211 needs to know not to drop
>>>>>>>> those sub-frames.
>>>>>>>>
>>>>>>>> I wonder if you'll ever get notified before the complete A-MPDU has
>>>>>>>> been received. That happens on previous chips. eg, you have an A-MPDU
>>>>>>>> of 16 frames with 4 MSDUs in each MPDU. If you get notified and handle
>>>>>>>> half of one MPDU before you hit EOL, the next notification you process
>>>>>>>> will be the next MSDU in the same MPDU - and then you'll hit the
>>>>>>>> reordering hilarity again.
>>>>>>>>
>>>>>>>> So hm, i will face the same issue in FreeBSD at some point, so I'd
>>>>>>>> likely do what you're thinking of doing - pass up a chain of mbufs
>>>>>>>> representing the current MPDU and treat the whole lot as the frame(s)
>>>>>>>> to care about. Honestly though, I'm kind of wondering whether I should
>>>>>>>> mostly just do what the Atheros reference driver does and treat it as
>>>>>>>> ethernet payload frames (ie, it's already de-encapsulated and the
>>>>>>>> reordering is already done) and just tack on the wifi header bit via
>>>>>>>> another of the DMA rings.
>>>>>>>>
>>>>>>>> Ugh, I really should sit down and write the FreeBSD version of this.
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> -a
>>>>>>>>
>>>>>>>> (I'm still having flashbacks from working on this firmware at QCA. Aiee.)
>>>>>>>>
>>>>>>>>
>>>>>>>> On 5 July 2014 06:55, Denton Gentry <denton.gentry at gmail.com> wrote:
>>>>>>>>> There are two issues in handling the dis-aggregated A-MSDU subframes
>>>>>>>>> in ieee80211_sta_manage_reorder_buf:
>>>>>>>>>
>>>>>>>>> 1. The out-of-date check:
>>>>>>>>>
>>>>>>>>> /* frame with out of date sequence number */
>>>>>>>>> if (ieee80211_sn_less(mpdu_seq_num, head_seq_num)) {
>>>>>>>>> dev_kfree_skb(skb);
>>>>>>>>> goto out;
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> As all of the subframes carry the same sequence number, the first
>>>>>>>>> subframe will be delivered and increment head_seq_num and then all
>>>>>>>>> subsequent subframes will be discarded.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> 2. The duplicates check a bit later in the routine:
>>>>>>>>>
>>>>>>>>> /* check if we already stored this frame */
>>>>>>>>> if (tid_agg_rx->reorder_buf[index]) {
>>>>>>>>> dev_kfree_skb(skb);
>>>>>>>>> goto out;
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> If there is enough packet loss that packets are queued in the reorder
>>>>>>>>> buffer and not immediately released, then only the first subframe will
>>>>>>>>> be stored. Subsequent subframes will be discarded as duplicates.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> An 802.11ac MacBook is able to get about 170 Mbps with iperf prior to
>>>>>>>>> the reordering buffer changes, and dropped to about 8 Mbps with the
>>>>>>>>> reorder buffer. Hacking around the out-of-date sequence number check
>>>>>>>>> to allow all subframes to egress restores it to 170 Mbps.
>>>>>>>>>
>>>>>>>>> In the area where I'm testing there isn't enough 5 GHz noise to make
>>>>>>>>> the duplicates-check issue happen very often. By adding a printk I can
>>>>>>>>> see that it does happen, but it doesn't impact the throughput and I
>>>>>>>>> can't report the impact of fixing it.
>>>>>>>>>
>>>>>>>>> ----
>>>>>>>>>
>>>>>>>>> I do wonder if both of these are symptoms of dis-aggregating the
>>>>>>>>> A-MSDU too early. mac80211 expects to be dealing with the whole MPDU
>>>>>>>>> at a time, and the ath10k A-MSDU case is sending it subframes instead.
>>>>>>>>> Trying to make the ath10k code send up all of the subframes as a chain
>>>>>>>>> of skbs didn't immediately work, but I do wonder if that would better
>>>>>>>>> match the mac80211 expectations.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Fri, Jul 4, 2014 at 11:58 AM, Denton Gentry <denton.gentry at gmail.com> wrote:
>>>>>>>>>> Yes, after some more testing it does look like an unfortunate
>>>>>>>>>> interaction between the reorder buffer and A-MSDU. The disaggregated
>>>>>>>>>> subframes all carry the same sequence number. The first subframe is
>>>>>>>>>> released from the reorder buffer and increments the head_seq_num.
>>>>>>>>>> Subsequent subframes are all discarded as being out of date.
>>>>>>>>>>
>>>>>>>>>> [ 308.514021] ieee80211_sta_manage_reorder_buf: out of date seq=0xb05
>>>>>>>>>> head=0xb06
>>>>>>>>>> [ 308.520577] ieee80211_sta_manage_reorder_buf: out of date seq=0xb0a
>>>>>>>>>> head=0xb0b
>>>>>>>>>> [ 308.527198] ieee80211_sta_manage_reorder_buf: out of date seq=0xb0f
>>>>>>>>>> head=0xb10
>>>>>>>>>> [ 308.533857] ieee80211_sta_manage_reorder_buf: out of date seq=0xb14
>>>>>>>>>> head=0xb15
>>>>>>>>>> [ 308.540480] ieee80211_sta_manage_reorder_buf: out of date seq=0xb19
>>>>>>>>>> head=0xb1a
>>>>>>>>>> [ 308.547730] ieee80211_sta_manage_reorder_buf: out of date seq=0xb1e
>>>>>>>>>> head=0xb1f
>>>>>>>>>> [ 308.554069] ieee80211_sta_manage_reorder_buf: out of date seq=0xb23
>>>>>>>>>> head=0xb24
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Wed, Jul 2, 2014 at 9:49 AM, Michal Kazior <michal.kazior at tieto.com> wrote:
>>>>>>>>>>> On 30 June 2014 22:15, Denton Gentry <denton.gentry at gmail.com> wrote:
>>>>>>>>>>>> In iperf tests using a MacBook STA bridging through an ath10k AP to an
>>>>>>>>>>>> Ethernet server, I'm noticing very selective packet loss. The second
>>>>>>>>>>>> and subsequent frames in an A-MSDU packet appear to be dropped.
>>>>>>>>>>>>
>>>>>>>>>>>> The AP sets the A-MSDU size to 3839 bytes, and the MacBook frequently
>>>>>>>>>>>> sends A-MSDU packets containing two TCP frames. So far as I can tell,
>>>>>>>>>>>> the first TCP frame from an A-MSDU aggregate is delivered and the
>>>>>>>>>>>> second is consistently lost. The MacBook generally retransmits the
>>>>>>>>>>>> lost frame as a singleton with no aggregation, and the retransmitted
>>>>>>>>>>>> frame makes it through.
>>>>>>>>>>>>
>>>>>>>>>>>> This became more noticeable after the reordering fixes in
>>>>>>>>>>>> http://lists.infradead.org/pipermail/ath10k/2014-June/002552.html
>>>>>>>>>>>>
>>>>>>>>>>>> I see this A-MSDU packet loss behavior both with and without the
>>>>>>>>>>>> reordering fixes, the first packet in an A-MSDU is delivered while the
>>>>>>>>>>>> second is dropped. However, *without* the reordering fixes (and
>>>>>>>>>>>> therefore with packets delivered out of order) the MacBook sends
>>>>>>>>>>>> relatively few A-MSDU frames. *With* the reordering fixes, so all
>>>>>>>>>>>> packets are delivered in order, the MacBook keeps sending A-MSDU and
>>>>>>>>>>>> therefore has to deal with more packet loss. I suspect it is an
>>>>>>>>>>>> interaction with the MacOS TCP congestion window which I'm likely
>>>>>>>>>>>> never going to fully understand, its stuck in a region of the
>>>>>>>>>>>> congestion window where the Wifi driver keeps choosing to using
>>>>>>>>>>>> A-MSDU.
>>>>>>>>>>>
>>>>>>>>>>> I was actually worried about A-MSDU once A-MPDU re-ordering issue was raised.
>>>>>>>>>>>
>>>>>>>>>>> ath10k fw reports A-MSDU subframes separately. To avoid memory
>>>>>>>>>>> copying/allocation overhead each subframe is reported as a singly
>>>>>>>>>>> A-MSDU MSDU to mac80211 with an extra rx_flag AMSDU_MORE. Perhaps
>>>>>>>>>>> A-MPDU reordering intereferes with A-MSDU now?
>>>>>>>>>>>
>>>>>>> Denton could you try attached patch: report amsdu in one big frame.
>>>>>>> If help, we can add amsdu skb list support to mac80211/cfg80211 - to
>>>>>>> improve performance and reduce memcpy, while currently we have skb
>>>>>>> frames, put them in one big skb and next cfg80211 split them again
>>>>>>> into msdus and report to stack.
>>>>>>>
>>>>>>> BR
>>>>>>> Janusz
More information about the ath10k
mailing list