Netgear R7800 exploration (IPQ8065, QCA9984)


#1371

As far as I can tell, Marvell does a lot of development to get it's Ethernet driver to be performant. Qualcomm does none.


#1372

from my own experience the R7800 CPU is more powerfull as the one in my WRT1900AC(S) V2. so i'm not so certain that the R7800 would actually be slower.
Wireless driver is definitively more stable as the WRT driver.


#1373

dd-wrt uses the Qualcomm SFE to accelerate traffic flow. With OpenWRT, you’ll have to build your own firmware to incorporate it. Or use the community builds who incorporated it.


#1374

Yes, the IPQ8065 does have a slight edge however it's very minor in most cases (raw CPU power). As far as WIFI goes, I'd say they're equal however there's a lot that can impact on WIFI performance which these radios handles differently. Keep in mind that drivers are just as important if not even more in terms of stability and performance.


#1375

How do you go about building firmware with Qualcomm SFE. I am using hnyman's build environment with his patches. Is it an option in make menuconfig?
Thank you.


#1376

I know MU-MIMO isn't supported (yet). Is there any way to get it working or do I have to wait until someone reverse engineers it?


#1377

Do any of your client devices support MU-MIMO already? Because if not it will not work.
Apparently even the vendor implementations do not really work/offer not real benefit.

With that background i wouldnt get my hopes too high to see it anytime soon in OWRT

Check SNB for some background.


#1378

Good point. I have some fairly new devices but now when I googled, it seems they support MIMO at most.


#1379

Hi, I'm running the OpenWrt 18.06-SNAPSHOT r7297-13dccfc8e4 build, and if I enable both 2.4GHz and 5GHz wifi simultaneously my R7800 just hard resets once clients connect. It just reboots, and then it works fine, until the clients connect again, which causes it to reset again.

I'm running out of things to try to debug this, do any of you maybe have an idea how to investigate this? I've tried all kinds of different configurations, and individually both 5GHz and 2.4GHz run well.

Since it just resets I'm not getting any handle on what might be wrong, I've tried running while true; do dmesg -c; done over a wired connection to try and get any output before it crashes, but nothing in there.

I've even tried using the ath10k-ct firmware, but with it it behaves exactly the same.

Any help would be greatly appreciated.

Config looks like this:

config wifi-device 'radio0'
	option type 'mac80211'
	option hwmode '11a'
	option path 'soc/1b500000.pci/pci0000:00/0000:00:00.0/0000:01:00.0'
	option htmode 'VHT80'
	option channel '52'
	option country 'US'
	option legacy_rates '1'

config wifi-iface 'default_radio0'
	option device 'radio0'
	option network 'lan'
	option mode 'ap'
	option encryption 'psk-mixed'
	option key 'secret'
	option wps_pushbutton '0'
	option ssid 'OpenWrt5'
	option disabled '1'

config wifi-device 'radio1'
	option type 'mac80211'
	option channel '11'
	option hwmode '11g'
	option path 'soc/1b700000.pci/pci0001:00/0001:00:00.0/0001:01:00.0'
	option htmode 'HT20'
	option country 'US'
	option legacy_rates '1'

config wifi-iface 'default_radio1'
	option device 'radio1'
	option network 'lan'
	option mode 'ap'
	option ssid 'OpenWrt'
	option encryption 'psk-mixed'
	option key 'secret'
	option wps_pushbutton '0'

dmesg | grep ath looks like this, maybe you can spot something in there?

[   18.377486] ath10k_pci 0000:01:00.0: assign IRQ: got 67
[   18.377864] ath10k_pci 0000:01:00.0: enabling device (0140 -> 0142)
[   18.377942] ath10k_pci 0000:01:00.0: enabling bus mastering
[   18.378389] ath10k_pci 0000:01:00.0: pci irq msi oper_irq_mode 2 irq_mode 0 reset_mode 0
[   18.568383] ath10k_pci 0000:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/firmware-6.bin failed with error -2
[   18.568430] ath10k_pci 0000:01:00.0: Falling back to user helper
[   18.595136] firmware ath10k!QCA9984!hw1.0!firmware-6.bin: firmware_loading_store: map pages failed
[   18.708116] ath10k_pci 0000:01:00.0: Unknown FW IE: 30
[   18.708143] ath10k_pci 0000:01:00.0: qca9984/qca9994 hw1.0 target 0x01000000 chip_id 0x00000000 sub 168c:cafe
[   18.712145] ath10k_pci 0000:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 1
[   18.723700] ath10k_pci 0000:01:00.0: firmware ver 10.4-ct-9984-xtW-010-868495e api 5 features peer-flow-ctrl crc32 b68bff6b
[   21.042564] ath10k_pci 0000:01:00.0: board_file api 2 bmi_id 0:1 crc32 dd6d039c
[   26.865910] ath10k_pci 0000:01:00.0: htt-ver 2.2 wmi-op 6 htt-op 4 cal pre-cal-file max-sta 32 raw 0 hwcrypto 1
[   26.956529] ath: EEPROM regdomain: 0x0
[   26.956542] ath: EEPROM indicates default country code should be used
[   26.956552] ath: doing EEPROM country->regdmn map search
[   26.956571] ath: country maps to regdmn code: 0x3a
[   26.956582] ath: Country alpha2 being used: US
[   26.956592] ath: Regpair used: 0x3a
[   26.961759] ath10k_pci 0001:01:00.0: assign IRQ: got 100
[   26.962811] ath10k_pci 0001:01:00.0: enabling device (0140 -> 0142)
[   26.962946] ath10k_pci 0001:01:00.0: enabling bus mastering
[   26.963622] ath10k_pci 0001:01:00.0: pci irq msi oper_irq_mode 2 irq_mode 0 reset_mode 0
[   27.135905] ath10k_pci 0001:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/firmware-6.bin failed with error -2
[   27.135936] ath10k_pci 0001:01:00.0: Falling back to user helper
[   27.255579] firmware ath10k!QCA9984!hw1.0!firmware-6.bin: firmware_loading_store: map pages failed
[   27.255729] ath10k_pci 0001:01:00.0: Unknown FW IE: 30
[   27.263544] ath10k_pci 0001:01:00.0: qca9984/qca9994 hw1.0 target 0x01000000 chip_id 0x00000000 sub 168c:cafe
[   27.268557] ath10k_pci 0001:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 1
[   27.280507] ath10k_pci 0001:01:00.0: firmware ver 10.4-ct-9984-xtW-010-868495e api 5 features peer-flow-ctrl crc32 b68bff6b
[   29.596422] ath10k_pci 0001:01:00.0: board_file api 2 bmi_id 0:2 crc32 dd6d039c
[   35.409898] ath10k_pci 0001:01:00.0: htt-ver 2.2 wmi-op 6 htt-op 4 cal pre-cal-file max-sta 32 raw 0 hwcrypto 1
[   35.495752] ath: EEPROM regdomain: 0x0
[   35.495763] ath: EEPROM indicates default country code should be used
[   35.495773] ath: doing EEPROM country->regdmn map search
[   35.495789] ath: country maps to regdmn code: 0x3a
[   35.495801] ath: Country alpha2 being used: US
[   35.495810] ath: Regpair used: 0x3a

#1380

Hi, I’m new here, recently moved from stock due to need for multi-wan for my IPTV. Stocks firmware qos doing very great. However, using openwrt 18.06.1 r7258-5eb055306f, I can’t even ping 8.8.8.8 (request time-out) if there’s heavy download traffic eg. Torrent, IDM. Playing normal YouTube doesn’t introduce any packet loss (request time out), only if there’s heavy download traffic, then packet loss happen. My gaming network is fully affected (spikes and freezes). All this happen when sqm is enabled using cake > piece of cake qos script. My advertised bandwidth is 20/5 Mbps. Download/Upload set up in sqm 17/3.7 down/up. Link layer adaptation set to none(default). Advance conf: none/default.

Any reply would be appreciated


#1381

Here, I will post my sqm config and my tc -d qdisc

tc -d qdisc
qdisc noqueue 0: dev lo root refcnt 2 
qdisc mq 0: dev eth0 root 
qdisc fq_codel 0: dev eth0 parent :1 limit 10240p flows 1024 quantum 1514 target 5.0ms interval 100.0ms memory_limit 4Mb ecn 
qdisc mq 0: dev eth1 root 
qdisc fq_codel 0: dev eth1 parent :1 limit 10240p flows 1024 quantum 1514 target 5.0ms interval 100.0ms memory_limit 4Mb ecn 
qdisc noqueue 0: dev br-lan root refcnt 2 
qdisc noqueue 0: dev eth1.1 root refcnt 2 
qdisc noqueue 0: dev eth0.3 root refcnt 2 
qdisc noqueue 0: dev wlan1 root refcnt 2 
qdisc noqueue 0: dev wlan0 root refcnt 2 
qdisc noqueue 0: dev eth0.2 root refcnt 2 
qdisc cake 8031: dev pppoe-wan root refcnt 2 bandwidth 3700Kbit besteffort dual-srchost nat split-gso rtt 100.0ms raw overhead 0 
qdisc ingress ffff: dev pppoe-wan parent ffff:fff1 ---------------- 
qdisc fq_codel 0: dev tun0 root refcnt 2 limit 10240p flows 1024 quantum 1500 target 5.0ms interval 100.0ms memory_limit 4Mb ecn 
qdisc cake 8032: dev ifb4pppoe-wan root refcnt 2 bandwidth 18100Kbit besteffort dual-dsthost nat wash split-gso rtt 100.0ms raw overhead 0
/etc/config/sqm
config queue 'eth0'
        option linklayer 'none'
        option verbosity '5'
        option debug_logging '0'
        option qdisc 'cake'
        option upload '3700'
        option interface 'pppoe-wan'
        option enabled '1'
        option download '18100'
        option script 'piece_of_cake.qos'
        option qdisc_advanced '1'
        option squash_dscp '1'
        option squash_ingress '1'
        option ingress_ecn 'ECN'
        option egress_ecn 'NOECN'
        option qdisc_really_really_advanced '1'
        option iqdisc_opts 'nat dual-dsthost'
        option eqdisc_opts 'nat dual-srchost'

#1382

config queue 'eth0' should be pppoe and it is important to select the link layer correctly as i told you. For fiber should be 38 or 39 of overhead.


#1383

I thought pppoe should be put in the “interface” option?

But according to the wiki, overhead for fiber is hard to determined.

fiber and ethernet it is much harder to figure out the exact overhead to specify… (the question is typically how is the ISP's upstream traffic shaper configured

Src: https://openwrt.org/docs/guide-user/network/traffic-shaping/sqm-details

I’m currently out of town, will try to specify the overhead as soon as I get back and update u


#1384

I'm running fiber 300/300. When i was tweaking the shaper parameters i tested with many options. I've asked @moeller0 some questions about this and followed his guidance.
Maybe this thread can help you out:

About the queue interface, i've also read (in the openwrt forums) that if you're using pppoe it is better to set the queue in the wan interface itself, but i cannot tell which thread it was


#1385

Ok, I tried to set the line with pppoe as u instructed from

config queue 'eth0'
        option linklayer 'none'
        option verbosity '5'
        option debug_logging '0'
        ...

to

config queue 'pppoe'
        option linklayer 'none'
        option verbosity '5'
        option debug_logging '0'
        ...

However, after changing that, the config in the luci gone (as it has been deleted, literally no config at all) and starting the sqm script, doesn't produce any log (maybe it detect config based on luci, so if luci sqm config missing, it treated the config as missing), changing back to eth0, the config showed up as before. trying to delete sqm config and make the config using luci interface doesn't list config queue, it was just like this

config queue
        option linklayer 'none'
        option verbosity '5'
        option debug_logging '0'
        ...

and the sqm produce log saying it's started again.


#1386

I took a glimpse about the post, but I don't think it is relevant with my issue. I indeed see improvement in my network, almost no spike in the latency. Regular browser download, streaming 1080p doesn't give any packet loss or spikes, however sometimes torrenting and downloading games data (fortnite iOS [ I suspect it uses multiple stream connection]) give packet loss. It even doesn't give any spikes, just pure packet loss with not a single packet received during download (literally 100% packet loss). However, browsing still continue perform well, streaming too doesn't introduce any buffering (IT'S FABULOUS). BUT, PACKET LOSS happened. It doesn't make sense to me at all.

UPDATE: I noticed just recently something strange happened. Streaming video 1080p during games download that causes pure packet loss in my internet, give packet response (rather good latency about 30 ms, no timed-out request at all). However, once the streaming stop, it continue to give packet loss again.
I will try to record my screen in a next few days. I'm so busy :sleepy:


#1387

Sorry! the pppoe-wan queue was in my 17.01 settings. Now in the 18.06 the queue is eth0 and and you specify the interface with the option interface pppoe-wan. You're correct. As for my sqm settings:

config queue 'eth0'
	option debug_logging '0'
	option qdisc 'cake'
	option verbosity '5'
	option qdisc_advanced '1'
	option squash_dscp '1'
	option squash_ingress '1'
	option ingress_ecn 'ECN'
	option egress_ecn 'NOECN'
	option linklayer 'ethernet'
	option qdisc_really_really_advanced '1'
	option iqdisc_opts 'nat dual-dsthost'
	option eqdisc_opts 'nat dual-srchost'
	option script 'piece_of_cake.qos'
	option interface 'pppoe-wan'
	option overhead '38'
	option enabled '1'
	option download '298000'
	option upload '298000'

#1388

Do I have to use -Os when compiling the kernel or will -O2 or -O3 work, was wondering because of the size issue.
When using -Ofast the vmlinux.elf size was slightly above 6MB and bzImage was 2.2MB, I guess it's too much? Not sure which of those files it uses.

EDIT:
-Ofast worked but dropbear had issues, not sure if it's because of that flag but i tried -O2 instead and all was good.


#1389

I seem to be losing 5GHz regularly. I've tried a few different builds; as of this moment I am running my own build off master from today, and I see this:

[ 6473.681862] ath10k_pci 0000:01:00.0: firmware crashed! (guid n/a)
[ 6473.681907] ath10k_pci 0000:01:00.0: qca9984/qca9994 hw1.0 target 0x01000000 chip_id 0x00000000 sub 168c:cafe
[ 6473.686951] ath10k_pci 0000:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 0
[ 6473.698720] ath10k_pci 0000:01:00.0: firmware ver 10.4-ct-9984-fW-011-cf79c7f api 5 features peer-flow-ctrl,txstatus-noack,wmi-10.x-CT,ratemask-CT,regdump-CT,txrate-CT,flush-all-CT,pingpong-CT,ch-regs-CT,nop-CT,set-special-CT,cust-stats-CT crc32 25783e66
[ 6473.705588] ath10k_pci 0000:01:00.0: board_file api 2 bmi_id 0:1 crc32 dd6d039c
[ 6473.727475] ath10k_pci 0000:01:00.0: htt-ver 2.2 wmi-op 6 htt-op 4 cal pre-cal-file max-sta 32 raw 0 hwcrypto 1

as well as lots of messages like this:

[ 6055.825542] ath10k_pci 0000:01:00.0: failed to submit frame: -19
[ 6055.825596] ath10k_pci 0000:01:00.0: failed to push frame: -19
[ 6055.837842] ath10k_pci 0000:01:00.0: failed to submit frame: -19
[ 6055.837902] ath10k_pci 0000:01:00.0: failed to push frame: -19
[ 6055.889570] ath10k_pci 0000:01:00.0: failed to submit frame: -19
[ 6055.889649] ath10k_pci 0000:01:00.0: failed to push frame: -19
[ 6055.923108] ath10k_pci 0000:01:00.0: Firmware lacks feature flag indicating a retry limit of > 2 is OK, requested limit: 4
[ 6474.262001] ieee80211 phy0: Hardware restart was requested

Has anyone else seen this?


#1390

you should report this to ath10k-ct