Skip to contentSkip to conversation

attempting mesh on ath10k

DISCUSSION Bob Copeland 5 participants 15 messages Apr 24, 2015 – May 27, 2015 Qualcomm Atheros ATH10K wireless driver

lore
Share & tools
lore
#0 lore raw

Hi all,

I have played around a little bit with the raw mode patches trying to get mesh working on ath10k. I've hacked the driver slightly to use an AP vif in the firmware in "mesh" mode, which seems to produce proper looking beacons over the air.

However, when I bring up a second mesh node, the ath10k device crashes upon trying to send the first mesh peering action frame. Here's what dmesg looks like:

[   43.860000] ath10k_pci 0000:01:00.0: wmi event debug mesg len 140
[   44.310000] XXX trying to create peer 00:03:7f:10:4e:2d
[   44.310000] ath10k_pci 0000:01:00.0: wmi peer create vdev_id 0 peer_addr 00:03:7f:10:4e:2d
[   44.320000] XXX sending WMI cmd 36892
[   44.320000] XXX sending WMI cmd nowait 36892
[   44.330000] XXX waiting for peer create 00:03:7f:10:4e:2d
[   44.330000] XXX peer 00:03:7f:10:4e:2d added
[   44.340000] ath10k_pci 0000:01:00.0: wmi peer assoc vdev 0 addr 00:03:7f:10:4e:2d (new)
[   44.350000] XXX sending WMI cmd 36896
[   44.350000] XXX sending WMI cmd nowait 36896
[   44.350000] ath10k_pci 0000:01:00.0: received event id 36891 not implemented
[   44.360000] mesh: sending popen to 00:03:7f:10:4e:2d
[   44.380000] ath10k_pci 0000:01:00.0: wmi mgmt tx skb 86fb4e40 len 84 ftype 00 stype d0
[   44.390000] XXX sending WMI cmd 36905
[   44.390000] XXX sending WMI cmd nowait 36905
[   44.390000] ath10k_pci 0000:01:00.0: firmware crashed! (uuid 839037ff-ad03-4442-9114-f99e03204c
[   44.400000] ath10k_pci 0000:01:00.0: qca988x hw2.0 (0x4100016c, 0x043202ff) fw 10.2-00082-4-2 a
[   44.410000] ath10k_pci 0000:01:00.0: debug 1 debugfs 1 tracing 0 dfs 1 testmode 1
[   44.420000] ath10k_pci 0000:01:00.0: firmware register dump:
[   44.430000] ath10k_pci 0000:01:00.0: [00]: 0x4100016C 0x000015B3 0x009A98BE 0x00955B31
[   44.440000] ath10k_pci 0000:01:00.0: [04]: 0x009A98BE 0x00060330 0x0000001E 0x00000B00
[   44.440000] ath10k_pci 0000:01:00.0: [08]: 0x00447730 0x009C3014 0x00446970 0x00000078
[   44.450000] ath10k_pci 0000:01:00.0: [12]: 0x00000009 0xFFFFFFFF 0x00958360 0x0095836B
[   44.460000] ath10k_pci 0000:01:00.0: [16]: 0x00958080 0x009A6CCF 0x00000000 0x00000000
[   44.470000] ath10k_pci 0000:01:00.0: [20]: 0x409A98BE 0x0040AB14 0x0040FAA8 0x0043E677
[   44.480000] ath10k_pci 0000:01:00.0: [24]: 0x809A9958 0x0040AB74 0x0043E7CC 0xC09A98BE
[   44.480000] ath10k_pci 0000:01:00.0: [28]: 0x8099F2E9 0x0040AB94 0x009C3014 0x0040AD04
[   44.490000] ath10k_pci 0000:01:00.0: [32]: 0x8098933F 0x0040ABB4 0x0043D0D8 0x0040AD04
[   44.500000] ath10k_pci 0000:01:00.0: [36]: 0x809B38C0 0x0040ABD4 0x009C3014 0x0043D0D8
[   44.510000] ath10k_pci 0000:01:00.0: [40]: 0x809AFD68 0x0040AD94 0x00400000 0x00416688
[   44.520000] ath10k_pci 0000:01:00.0: [44]: 0x809BDA75 0x0040ADE4 0x0040AE08 0x00411D44
[   44.520000] ath10k_pci 0000:01:00.0: [48]: 0x809486FA 0x0040AE04 0x00000001 0x00000000
[   44.530000] ath10k_pci 0000:01:00.0: [52]: 0x80948E2C 0x0040AEA4 0x0041E424 0x00412350
[   44.540000] ath10k_pci 0000:01:00.0: [56]: 0x809BCF0D 0x0040AEC4 0x0041E424 0x00000001

My hacky compat-wireless snapshot is over here, any suggestions on what to try next to debug further are welcome:

https://github.com/bcopeland/ath10k-mesh

[Also I tried skipping the mgmt tx datapath and use normal datapath, as seen in the final patch, but that doesn't change things, it crashes either way.]

Maybe I'm not setting up peers correctly or something like that?

Signature · 6 lines
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#1 lore raw

Hi, Bob

I have tried converting the "adhoc" mode on the
firmware-2.bin_999.999.0.636 as mesh interface.

You can find my patches here:
https://github.com/chunyeow/ath/commits/mesh

But probably too many workaround and won't be able to interoperate with other mesh interface.

I have also found out the immediate crashes when peer mesh STA is added by the firmware if using "AP" mode.
https://github.com/chunyeow/ath/commit/c09a7ef83a5912d1965756b9ce4c9a22ae1dcf08

Hi, Ben

You have any updates on CT firmware to support the mesh mode?

Please advice. Thanks

---
Chun-Yeow

Quoted 57 lines · Bob Copeland

Hi all,

I have played around a little bit with the raw mode patches trying to get mesh working on ath10k. I've hacked the driver slightly to use an AP vif in the firmware in "mesh" mode, which seems to produce proper looking beacons over the air.

However, when I bring up a second mesh node, the ath10k device crashes upon trying to send the first mesh peering action frame. Here's what dmesg looks like:

[   43.860000] ath10k_pci 0000:01:00.0: wmi event debug mesg len 140
[   44.310000] XXX trying to create peer 00:03:7f:10:4e:2d
[   44.310000] ath10k_pci 0000:01:00.0: wmi peer create vdev_id 0 peer_addr 00:03:7f:10:4e:2d
[   44.320000] XXX sending WMI cmd 36892
[   44.320000] XXX sending WMI cmd nowait 36892
[   44.330000] XXX waiting for peer create 00:03:7f:10:4e:2d
[   44.330000] XXX peer 00:03:7f:10:4e:2d added
[   44.340000] ath10k_pci 0000:01:00.0: wmi peer assoc vdev 0 addr 00:03:7f:10:4e:2d (new)
[   44.350000] XXX sending WMI cmd 36896
[   44.350000] XXX sending WMI cmd nowait 36896
[   44.350000] ath10k_pci 0000:01:00.0: received event id 36891 not implemented
[   44.360000] mesh: sending popen to 00:03:7f:10:4e:2d
[   44.380000] ath10k_pci 0000:01:00.0: wmi mgmt tx skb 86fb4e40 len 84 ftype 00 stype d0
[   44.390000] XXX sending WMI cmd 36905
[   44.390000] XXX sending WMI cmd nowait 36905
[   44.390000] ath10k_pci 0000:01:00.0: firmware crashed! (uuid 839037ff-ad03-4442-9114-f99e03204c
[   44.400000] ath10k_pci 0000:01:00.0: qca988x hw2.0 (0x4100016c, 0x043202ff) fw 10.2-00082-4-2 a
[   44.410000] ath10k_pci 0000:01:00.0: debug 1 debugfs 1 tracing 0 dfs 1 testmode 1
[   44.420000] ath10k_pci 0000:01:00.0: firmware register dump:
[   44.430000] ath10k_pci 0000:01:00.0: [00]: 0x4100016C 0x000015B3 0x009A98BE 0x00955B31
[   44.440000] ath10k_pci 0000:01:00.0: [04]: 0x009A98BE 0x00060330 0x0000001E 0x00000B00
[   44.440000] ath10k_pci 0000:01:00.0: [08]: 0x00447730 0x009C3014 0x00446970 0x00000078
[   44.450000] ath10k_pci 0000:01:00.0: [12]: 0x00000009 0xFFFFFFFF 0x00958360 0x0095836B
[   44.460000] ath10k_pci 0000:01:00.0: [16]: 0x00958080 0x009A6CCF 0x00000000 0x00000000
[   44.470000] ath10k_pci 0000:01:00.0: [20]: 0x409A98BE 0x0040AB14 0x0040FAA8 0x0043E677
[   44.480000] ath10k_pci 0000:01:00.0: [24]: 0x809A9958 0x0040AB74 0x0043E7CC 0xC09A98BE
[   44.480000] ath10k_pci 0000:01:00.0: [28]: 0x8099F2E9 0x0040AB94 0x009C3014 0x0040AD04
[   44.490000] ath10k_pci 0000:01:00.0: [32]: 0x8098933F 0x0040ABB4 0x0043D0D8 0x0040AD04
[   44.500000] ath10k_pci 0000:01:00.0: [36]: 0x809B38C0 0x0040ABD4 0x009C3014 0x0043D0D8
[   44.510000] ath10k_pci 0000:01:00.0: [40]: 0x809AFD68 0x0040AD94 0x00400000 0x00416688
[   44.520000] ath10k_pci 0000:01:00.0: [44]: 0x809BDA75 0x0040ADE4 0x0040AE08 0x00411D44
[   44.520000] ath10k_pci 0000:01:00.0: [48]: 0x809486FA 0x0040AE04 0x00000001 0x00000000
[   44.530000] ath10k_pci 0000:01:00.0: [52]: 0x80948E2C 0x0040AEA4 0x0041E424 0x00412350
[   44.540000] ath10k_pci 0000:01:00.0: [56]: 0x809BCF0D 0x0040AEC4 0x0041E424 0x00000001

My hacky compat-wireless snapshot is over here, any suggestions on what to try next to debug further are welcome:

https://github.com/bcopeland/ath10k-mesh

[Also I tried skipping the mgmt tx datapath and use normal datapath, as seen in the final patch, but that doesn't change things, it crashes either way.]

Maybe I'm not setting up peers correctly or something like that?

Quoted 7 lines

--
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

#2 lore raw
Quoted 15 lines · Yeoh Chun-Yeow

Hi, Bob

I have tried converting the "adhoc" mode on the
firmware-2.bin_999.999.0.636 as mesh interface.

You can find my patches here:
https://github.com/chunyeow/ath/commits/mesh

But probably too many workaround and won't be able to interoperate with other mesh interface.

I have also found out the immediate crashes when peer mesh STA is added by the firmware if using "AP" mode.
https://github.com/chunyeow/ath/commit/c09a7ef83a5912d1965756b9ce4c9a22ae1dcf08

Quoted 5 lines

Hi, Ben

You have any updates on CT firmware to support the mesh mode?

Please advice. Thanks

I have IBSS support working, and at least non-encrypted raw transmit can work in my firmware, though raw tx requires Michal's RFC patch he posted some time ago.

Don't know about mesh, but if you can crash my firmware, please send me the debug info and I'll see if I can figure out why.

I am not sure raw tx with encryption is possible...and at the least, it is probably not something I'll attempt to do for fun anytime soon.

Thanks,
Ben

Quoted 76 lines

---
Chun-Yeow

Quoted 57 lines · Bob Copeland

Hi all,

I have played around a little bit with the raw mode patches trying to get mesh working on ath10k. I've hacked the driver slightly to use an AP vif in the firmware in "mesh" mode, which seems to produce proper looking beacons over the air.

However, when I bring up a second mesh node, the ath10k device crashes upon trying to send the first mesh peering action frame. Here's what dmesg looks like:

[   43.860000] ath10k_pci 0000:01:00.0: wmi event debug mesg len 140
[   44.310000] XXX trying to create peer 00:03:7f:10:4e:2d
[   44.310000] ath10k_pci 0000:01:00.0: wmi peer create vdev_id 0 peer_addr 00:03:7f:10:4e:2d
[   44.320000] XXX sending WMI cmd 36892
[   44.320000] XXX sending WMI cmd nowait 36892
[   44.330000] XXX waiting for peer create 00:03:7f:10:4e:2d
[   44.330000] XXX peer 00:03:7f:10:4e:2d added
[   44.340000] ath10k_pci 0000:01:00.0: wmi peer assoc vdev 0 addr 00:03:7f:10:4e:2d (new)
[   44.350000] XXX sending WMI cmd 36896
[   44.350000] XXX sending WMI cmd nowait 36896
[   44.350000] ath10k_pci 0000:01:00.0: received event id 36891 not implemented
[   44.360000] mesh: sending popen to 00:03:7f:10:4e:2d
[   44.380000] ath10k_pci 0000:01:00.0: wmi mgmt tx skb 86fb4e40 len 84 ftype 00 stype d0
[   44.390000] XXX sending WMI cmd 36905
[   44.390000] XXX sending WMI cmd nowait 36905
[   44.390000] ath10k_pci 0000:01:00.0: firmware crashed! (uuid 839037ff-ad03-4442-9114-f99e03204c
[   44.400000] ath10k_pci 0000:01:00.0: qca988x hw2.0 (0x4100016c, 0x043202ff) fw 10.2-00082-4-2 a
[   44.410000] ath10k_pci 0000:01:00.0: debug 1 debugfs 1 tracing 0 dfs 1 testmode 1
[   44.420000] ath10k_pci 0000:01:00.0: firmware register dump:
[   44.430000] ath10k_pci 0000:01:00.0: [00]: 0x4100016C 0x000015B3 0x009A98BE 0x00955B31
[   44.440000] ath10k_pci 0000:01:00.0: [04]: 0x009A98BE 0x00060330 0x0000001E 0x00000B00
[   44.440000] ath10k_pci 0000:01:00.0: [08]: 0x00447730 0x009C3014 0x00446970 0x00000078
[   44.450000] ath10k_pci 0000:01:00.0: [12]: 0x00000009 0xFFFFFFFF 0x00958360 0x0095836B
[   44.460000] ath10k_pci 0000:01:00.0: [16]: 0x00958080 0x009A6CCF 0x00000000 0x00000000
[   44.470000] ath10k_pci 0000:01:00.0: [20]: 0x409A98BE 0x0040AB14 0x0040FAA8 0x0043E677
[   44.480000] ath10k_pci 0000:01:00.0: [24]: 0x809A9958 0x0040AB74 0x0043E7CC 0xC09A98BE
[   44.480000] ath10k_pci 0000:01:00.0: [28]: 0x8099F2E9 0x0040AB94 0x009C3014 0x0040AD04
[   44.490000] ath10k_pci 0000:01:00.0: [32]: 0x8098933F 0x0040ABB4 0x0043D0D8 0x0040AD04
[   44.500000] ath10k_pci 0000:01:00.0: [36]: 0x809B38C0 0x0040ABD4 0x009C3014 0x0043D0D8
[   44.510000] ath10k_pci 0000:01:00.0: [40]: 0x809AFD68 0x0040AD94 0x00400000 0x00416688
[   44.520000] ath10k_pci 0000:01:00.0: [44]: 0x809BDA75 0x0040ADE4 0x0040AE08 0x00411D44
[   44.520000] ath10k_pci 0000:01:00.0: [48]: 0x809486FA 0x0040AE04 0x00000001 0x00000000
[   44.530000] ath10k_pci 0000:01:00.0: [52]: 0x80948E2C 0x0040AEA4 0x0041E424 0x00412350
[   44.540000] ath10k_pci 0000:01:00.0: [56]: 0x809BCF0D 0x0040AEC4 0x0041E424 0x00000001

My hacky compat-wireless snapshot is over here, any suggestions on what to try next to debug further are welcome:

https://github.com/bcopeland/ath10k-mesh

[Also I tried skipping the mgmt tx datapath and use normal datapath, as seen in the final patch, but that doesn't change things, it crashes either way.]

Maybe I'm not setting up peers correctly or something like that?

Quoted 7 lines

--
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

Signature · 8 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#3 lore raw
Quoted 3 lines · Ben Greear

I have IBSS support working, and at least non-encrypted raw transmit can work in my firmware, though raw tx requires Michal's RFC patch he posted some time ago.

(Michal's RFC patch is in my repo.)

So I got a bit further. I was still getting a crash with CT firmware, but, suspecting that we pass something invalid in assoc message for mesh, I forced peer->aid = 1 and then it no longer crashes shortly after adding the first peer on either CT or QCA firmware.

I pushed that hack into my repo for now -- I think we want to rethink how and when AIDs are generated in mesh.

# iw dev wlan0 station dump
Station 00:03:7f:10:4e:2d (on wlan0)
        inactive time:  400 ms
        rx bytes:       34019
        rx packets:     959
        tx bytes:       152
        tx packets:     3
        tx retries:     0
        tx failed:      0
        signal:         -27 dBm
        signal avg:     -26 dBm
        tx bitrate:     6.0 MBit/s
        rx bitrate:     9.0 MBit/s
        mesh llid:      220
        mesh plid:      455
        mesh plink:     ESTAB
        mesh local PS mode:     ACTIVE
        mesh peer PS mode:      ACTIVE
        mesh non-peer PS mode:  ACTIVE
        authorized:     yes
        authenticated:  yes
        preamble:       long
        WMM/WME:        yes
        MFP:            no
        TDLS peer:      no

So, progress!

There's still an issue with datapath -- while it looks like we can establish plinks, I couldn't get unicast traffic flowing even if manually setting up mpaths and arp table entries. I haven't yet got all of my 5 ghz capable gear in the same room to monitor and see where the problem lies.

Signature · 6 lines
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#4 lore raw
Quoted 11 lines · Bob Copeland
Quoted 3 lines · Ben Greear

I have IBSS support working, and at least non-encrypted raw transmit can work in my firmware, though raw tx requires Michal's RFC patch he posted some time ago.

(Michal's RFC patch is in my repo.)

So I got a bit further. I was still getting a crash with CT firmware, but, suspecting that we pass something invalid in assoc message for mesh, I forced peer->aid = 1 and then it no longer crashes shortly after adding the first peer on either CT or QCA firmware.

Please send me the crash dump...and let me know the firmware version if the info is not in the dump.

Thanks,
Ben

Quoted 37 lines

I pushed that hack into my repo for now -- I think we want to rethink how and when AIDs are generated in mesh.

# iw dev wlan0 station dump
Station 00:03:7f:10:4e:2d (on wlan0)
         inactive time:  400 ms
         rx bytes:       34019
         rx packets:     959
         tx bytes:       152
         tx packets:     3
         tx retries:     0
         tx failed:      0
         signal:         -27 dBm
         signal avg:     -26 dBm
         tx bitrate:     6.0 MBit/s
         rx bitrate:     9.0 MBit/s
         mesh llid:      220
         mesh plid:      455
         mesh plink:     ESTAB
         mesh local PS mode:     ACTIVE
         mesh peer PS mode:      ACTIVE
         mesh non-peer PS mode:  ACTIVE
         authorized:     yes
         authenticated:  yes
         preamble:       long
         WMM/WME:        yes
         MFP:            no
         TDLS peer:      no

So, progress!

There's still an issue with datapath -- while it looks like we can establish plinks, I couldn't get unicast traffic flowing even if manually setting up mpaths and arp table entries. I haven't yet got all of my 5 ghz capable gear in the same room to monitor and see where the problem lies.

Signature · 8 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#5 lore raw
In reply to Bob Copeland
Quoted 6 lines · Bob Copeland

So, progress!

There's still an issue with datapath -- while it looks like we can establish plinks, I couldn't get unicast traffic flowing even if manually setting up mpaths and arp table entries. I haven't yet got all of my 5 ghz capable gear in the same room to monitor and see where the problem lies.

Now I had a chance to monitor it, and in so doing, rediscovered this patch of Chun-Yeow's:

http://lists.infradead.org/pipermail/ath10k/2014-September/003206.html

In fact my broadcasts were truncated and the above change makes the broadcast ARPs look correct over the air. The (non-ath10k) target mesh STA replies with broadcast PREQ but that is not received / processed / replied-to by the ath10k device.

Perhaps there is a BSSID filter enabled or something like that (in mesh there's no BSSID, just individual addresses on group addressed frames).

Signature · 6 lines
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#6 lore raw
Quoted 13 lines · Bob Copeland
Quoted 6 lines · Bob Copeland

So, progress!

There's still an issue with datapath -- while it looks like we can establish plinks, I couldn't get unicast traffic flowing even if manually setting up mpaths and arp table entries. I haven't yet got all of my 5 ghz capable gear in the same room to monitor and see where the problem lies.

Now I had a chance to monitor it, and in so doing, rediscovered this patch of Chun-Yeow's:

http://lists.infradead.org/pipermail/ath10k/2014-September/003206.html

Quoted 7 lines

In fact my broadcasts were truncated and the above change makes the broadcast ARPs look correct over the air. The (non-ath10k) target mesh STA replies with broadcast PREQ but that is not received / processed / replied-to by the ath10k device.

Perhaps there is a BSSID filter enabled or something like that (in mesh there's no BSSID, just individual addresses on group addressed frames).

In that case try creating a monitor interface and bring it up. This should prompt ath10k to create monitor vdev and subsequently disable most Rx filters in firmware.

Michał

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

#7 lore raw
Quoted 6 lines · Michal Kazior
Quoted 2 lines

Perhaps there is a BSSID filter enabled or something like that (in mesh there's no BSSID, just individual addresses on group addressed frames).

In that case try creating a monitor interface and bring it up. This should prompt ath10k to create monitor vdev and subsequently disable most Rx filters in firmware.

Indeed, that did the trick, thanks! So mesh is functional now with raw mode patch and a few hacks:

root@archer:~# iw dev
phy#0
        Interface mon0
                ifindex 8
                wdev 0x2
                addr e8:de:27:d6:9c:d0
                type monitor
        Interface wlan0
                ifindex 4
                wdev 0x1
                addr e8:de:27:d6:9c:d0
                type mesh point
                channel 149 (5745 MHz), width: 20 MHz (no HT), center1: 5745 MHz
root@archer:~# ping 10.10.10.13
PING 10.10.10.13 (10.10.10.13): 56 data bytes
64 bytes from 10.10.10.13: seq=0 ttl=64 time=10.031 ms
64 bytes from 10.10.10.13: seq=1 ttl=64 time=1.074 ms
64 bytes from 10.10.10.13: seq=2 ttl=64 time=1.074 ms
64 bytes from 10.10.10.13: seq=3 ttl=64 time=1.082 ms
64 bytes from 10.10.10.13: seq=4 ttl=64 time=1.032 ms
64 bytes from 10.10.10.13: seq=5 ttl=64 time=2.905 ms
64 bytes from 10.10.10.13: seq=6 ttl=64 time=1.093 ms
64 bytes from 10.10.10.13: seq=7 ttl=64 time=1.205 ms
64 bytes from 10.10.10.13: seq=8 ttl=64 time=1.057 ms
64 bytes from 10.10.10.13: seq=9 ttl=64 time=1.039 ms
64 bytes from 10.10.10.13: seq=10 ttl=64 time=1.107 ms
64 bytes from 10.10.10.13: seq=11 ttl=64 time=1.040 ms
^C
--- 10.10.10.13 ping statistics ---
12 packets transmitted, 12 packets received, 0% packet loss
round-trip min/avg/max = 1.032/1.978/10.031 ms
root@archer:~# iw dev wlan0 mpath dump
DEST ADDR         NEXT HOP          IFACE       SN      METRIC  QLEN
EXPTIME         DTIM    DRET    FLAGS
00:03:7f:10:4e:2d 00:03:7f:10:4e:2d wlan0       2427    1366    0
0       100     0       0x4
Signature · 6 lines
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#8 lore raw
Quoted 7 lines · Bob Copeland
Quoted 6 lines · Michal Kazior
Quoted 2 lines

Perhaps there is a BSSID filter enabled or something like that (in mesh there's no BSSID, just individual addresses on group addressed frames).

In that case try creating a monitor interface and bring it up. This should prompt ath10k to create monitor vdev and subsequently disable most Rx filters in firmware.

Are there any plans to push this work upstream?

Regards,
Bartosz

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

#9 lore raw
Quoted 10 lines · Bartosz Markowski
Quoted 7 lines · Bob Copeland
Quoted 6 lines · Michal Kazior
Quoted 2 lines

Perhaps there is a BSSID filter enabled or something like that (in mesh there's no BSSID, just individual addresses on group addressed frames).

In that case try creating a monitor interface and bring it up. This should prompt ath10k to create monitor vdev and subsequently disable most Rx filters in firmware.

Are there any plans to push this work upstream?

Hi Bartosz,

Yes, I plan to upstream it. I'm waiting on a hardware to arrive so I can implement VHT mesh with two ath10k chips; once that's in and working I'll clean it up and submit.

Signature · 6 lines
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#10 lore raw
Quoted 1 line · Bob Copeland

On Wed, May 27, 2015 at 08:51:05AM +0200, Bartosz Markowski wrote:

[...]

Quoted 7 lines
Quoted 1 line

Are there any plans to push this work upstream?

Hi Bartosz,

Yes, I plan to upstream it. I'm waiting on a hardware to arrive so I can implement VHT mesh with two ath10k chips; once that's in and working I'll clean it up and submit.

Great! Thanks for your effort.

Signature · 6 lines
Bartosz

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#11 lore raw
In reply to Yeoh Chun-Yeow
Quoted 10 lines · Chun-Yeow Yeoh

Hi, Bob

I have tried converting the "adhoc" mode on the
firmware-2.bin_999.999.0.636 as mesh interface.

You can find my patches here:
https://github.com/chunyeow/ath/commits/mesh

But probably too many workaround and won't be able to interoperate with other mesh interface.

Yeah, I think adhoc will not work so well due to TSF sync and the distributed beacon algorithm. So I'm happy that AP approach seems viable; of course dedicated mesh vdev type would be nice :)

Signature · 6 lines
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#12 lore raw

Great works, Bob.

Which AP firmware that you have tested with?

----
Chun-Yeow

Quoted 12 lines · Bob Copeland
Quoted 10 lines · Chun-Yeow Yeoh

Hi, Bob

I have tried converting the "adhoc" mode on the
firmware-2.bin_999.999.0.636 as mesh interface.

You can find my patches here:
https://github.com/chunyeow/ath/commits/mesh

But probably too many workaround and won't be able to interoperate with other mesh interface.

Quoted 6 lines

Yeah, I think adhoc will not work so well due to TSF sync and the distributed beacon algorithm. So I'm happy that AP approach seems viable; of course dedicated mesh vdev type would be nice :)

--
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

#13 lore raw
Quoted 3 lines · Chun-Yeow Yeoh

Great works, Bob.

Which AP firmware that you have tested with?

That is with:

[   16.670000] ath10k_pci 0000:01:00.0: qca988x hw2.0 (0x4100016c,

0x043202ff) fw 10.2.4.45 api 4 htt 2.1 wmi 5 cal otp max_sta 128

I think it should work with CT too. I need to go back and validate that, and also make sure I can repro the AID crash on CT. A confounding factor is that every time I build new drivers for OpenWRT, it overwrites the firmware again...

Signature · 6 lines
Bob Copeland %% http://bobcopeland.com/

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#14 lore raw
Quoted 7 lines · Bob Copeland
Quoted 3 lines · Chun-Yeow Yeoh

Great works, Bob.

Which AP firmware that you have tested with?

That is with:

Quoted 7 lines
[   16.670000] ath10k_pci 0000:01:00.0: qca988x hw2.0 (0x4100016c,

0x043202ff) fw 10.2.4.45 api 4 htt 2.1 wmi 5 cal otp max_sta 128

I think it should work with CT too. I need to go back and validate that, and also make sure I can repro the AID crash on CT. A confounding factor is that every time I build new drivers for OpenWRT, it overwrites the firmware again...

If you feel adventurous, and want to apply some of my ath10k patches, I have a way to dump some of the interesting registers in CT firmware, including the rx-filter register.

Maybe I can figure out what needs to be twiddled w/out having to add a monitor device.

I could also potentially add a vdev type or flag for mesh interface, but that would definitely require driver hacks to enable the new flag and it would not be compatible with other upstream firmware.

Thanks,
Ben

Signature · 7 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k