1d 6h here, no memory leaks, I didn't see ram below 115MB
Could you share a compiled ssdk_sh binary, please?
Sure, let's do this in PM, to not create offtopic.
Could you send me binary too? Thanks
Returned to latest release of robimarko's firmware, and discovered old annoying problems: some l2 roaming issue: no connectivity for few minutes after switching cable > wifi, and iphone problem, mentioned here earlier: unstable connectivity / disconnects wifi, while radio signal is full.
Are improvements expected in this place?
Know problem. Check this: Adding OpenWrt support for Xiaomi AX3600 - #5879 by sqrwv
1 Like
Hm,
It does load the modules just fine:
Press the [f] key and hit [enter] to enter failsafe mode
Press the [1], [2], [3] or [4] key and hit [enter] to select the debug level
f
[ 3.747594] ssdk_switch_device_num_init[1160]:INFO:ess-switch dts node number: 1
[ 3.747707] ssdk_dt_get_switch_node[971]:INFO:ess-switch DT exist!
[ 3.754089] ssdk_dt_parse_access_mode[857]:INFO:switch_access_mode: local bus
[ 3.760068] ssdk_dt_parse_access_mode[870]:INFO:switchreg_base_addr: 0x3a000000
[ 3.767262] ssdk_dt_parse_access_mode[871]:INFO:switchreg_size: 0x1000000
[ 3.774387] ssdk_dt_parse_mac_mode[295]:INFO:mac mode = 0xb
[ 3.781324] ssdk_dt_parse_mac_mode[304]:INFO:mac mode1 = 0xc
[ 3.786703] ssdk_dt_parse_mac_mode[313]:INFO:mac mode2 = 0xff
[ 3.792621] ssdk_dt_parse_phy_info[659]:INFO:[PORT 5] port_mac_sel = QGMAC_PORT
[ 3.798273] ssdk_dt_parse_uniphy[332]:INFO:ess-uniphy DT exist!
[ 3.805459] ssdk_dt_parse_intf_mac[800]:INFO:dp1 MAC 9c:9d:7e:7b:d2:9a
[ 3.811299] ssdk_dt_parse_intf_mac[800]:INFO:dp2 MAC 9c:9d:7e:7b:d2:9a
[ 3.817892] ssdk_dt_parse_intf_mac[800]:INFO:dp3 MAC 9c:9d:7e:7b:d2:9a
[ 3.824407] ssdk_dt_parse_intf_mac[800]:INFO:dp4 MAC 9c:9d:7e:7b:d2:9a
[ 3.830923] ssdk_dt_parse_intf_mac[800]:INFO:dp5 MAC 9c:9d:7e:b5:c3:a2
[ 3.837434] ssdk_plat_init start
[ 3.980129] ssdk_gcc_clock_init[1033]:INFO:SSDK gcc clock init successfully!
[ 3.980930] HPPE initializing...
[ 3.986463] malibu_phy_api_ops_init[2860]:INFO:qca probe malibu phy driver succeeded!
[ 3.994018] qca808x_phy_api_ops_init[2208]:INFO:qca probe qca808x phy driver succeeded!
[ 3.998277] regi_init[3567]:INFO:Initializing HPPE!!
[ 4.238701] ssdk_ppe_reset_init[1265]:INFO:ppe reset successfully!
[ 4.241913] qca_hppe_tdm_hw_init[684]:INFO:tdm setup num=96
[ 4.244081] qca_hppe_portctrl_hw_init[110]:INFO:Hawkeye PPE port initializing
[ 5.734864] ssdk_switch_register[1718]:INFO:Chip version 0x1500
[ 5.734924] qca_link_polling_select[1315]:INFO:link-polling-required node does not exist
[ 5.739617] ssdk_switch_register[1744]:INFO:polling is selected
[ 5.747933] regi_init[3571]:INFO:Initializing HPPE Done!!
[ 5.753640] regi_init[3631]:INFO:qca-ssdk module init succeeded!
[ 5.761494] EDMA ver 1 hw init
[ 5.765423] EDMA HW Reset completed succesfully
[ 5.768099] Num rings - TxDesc:1 (23-23) TxCmpl:1 (7-7)
[ 5.772513] RxDesc:1 (15-15) RxFill:1 (7-7)
[ 5.982222] **********************************************************
[ 5.982260] * NSS Data Plane driver
[ 5.987638] **********************************************************
- failsafe -
Waiting for kernel randomness to be initialised...
[ 9.908697] random: crng init done
[ 9.908728] random: 3 urandom warning(s) missed due to ratelimiting
Generating 1024 bit rsa key, this may take a while...
Public key portion is:
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAAAgQCXCDSSOZVwFl8ysSJjMEs4T4HFGnwHjD0tsKM2KTpI+JAWV0Lcpw7wD6Br7o47dy2I+MRFe93u7efh6GHGWszKQ3wxufwpc/8IyMcVO6itddFpkazQ8BxkWRDeCzDOl/DbcuFisVvDQKdrYZByq3sSpOoN/Udop8DjC6TWOIVF0w== root@(none)
Fingerprint: sha1!! 47:01:7f:d9:3b:f0:8f:b6:61:a3:8a:a7:d2:74:49:6d:15:43:c5:47
BusyBox v1.35.0 (2022-02-13 19:56:15 UTC) built-in shell (ash)
ash: can't access tty; job control turned off
_______ ________ __
| |.-----.-----.-----.| | | |.----.| |_
| - || _ | -__| || | | || _|| _|
|_______|| __|_____|__|__||________||__| |____|
|__| W I R E L E S S F R E E D O M
-----------------------------------------------------
OpenWrt SNAPSHOT, r18856-d21fd5135c
-----------------------------------------------------
================= FAILSAFE MODE active ================
special commands:
* firstboot reset settings to factory defaults
* mount_root mount root-partition with config files
after mount_root:
* passwd change root's password
* /etc/config directory with config files
for more help see:
https://openwrt.org/docs/guide-user/troubleshooting/
- failsafe_and_factory_reset
- root_password_reset
=======================================================
root@(none):/# ifconfig
However as you can see it does not bring up the interfaces
Just realized that Felix's bridge code improvement is already part of your latest version (he merged it 3 days ago), so that likely also helped besides the ath11k NAPI patches. 
https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=94b4da9b4aaded6ed8a5bd93bf38a7c92753fe68
It's not included since it only patches 5.10 kernel, otherwise, it would have broken bridging completely as he fixed it after that commit and that one is not included yet.
Now we are starting to see a diff between the 5.10 and 5.15 as all of that new stuff isn't being added to 5.15
BTW:
Anybody experienced the issue they described?
1 Like
don't have any router without the patch but this is my /proc/softirqs after 2 days and 15h of uptime:
root@router-2:~# cat /proc/softirqs
CPU0 CPU1 CPU2 CPU3
HI: 0 0 0 0
TIMER: 678656 1130096 5320746 2104677
NET_TX: 10 12 11 7
NET_RX: 130140430 17020536 10867509 34992740
BLOCK: 0 0 0 0
IRQ_POLL: 0 0 0 0
TASKLET: 31806316 3339 3541 2277
SCHED: 1146227 1520741 5563928 2876641
HRTIMER: 0 0 0 0
RCU: 620824 616013 766551 1051306
I am looking using htop and the CPU usage with a radio running but no traffic is at basically 0 so I would call this a non-issue for upstream ath11k.
1 Like
even with wifi traffic (not much, just normal web page use) cores are around 1-2% (and only on 1/2 of them)
Yeah, then its for sure an issue with the mess of 4.4 QCA kernel and whatever custom backports plus 200+ patches on top and then backporting NAPI threading from 5.12 into that.
What file name did you use? I used 01
I prefixed with 05
I mean, it shouldn't really matter as far as I can tell
I don't think so, only if you used something like 50.
But I have no idea, it should work, as mine isn't a lot different
Can you try a IP link instead of ifconfig? Maybe that's the issue
Also, do you see the lights in the ethernet ports light up?
And can you get connectivity on Lan 1?
Interfaces are registered, ip link tells us that but they are left disabled.
Lights mean nothing, it only means that the PHY is alive and auto negotiated something but the interface is disabled.
And If I try enabling them there is a shitstorm of:
WARN: netdev Null src_info_type:0x5. Drop skb:ffffff8005700500
Only the WAN port doesn't throw anything, its clearly lacking networking config and looping or whatever
But how does it work for me then? 
I honestly dont know, I can tell you that /etc/config/network is completely missing in failsafe.
But I have not used failsafe in ages, last time was to reset the admin password for a random QSDK device.
Failsafe has it's own scripts to bring up only the first Lan interface which should be Lan1 aka eth1, that worked for me and I can push the changes I made to my git and you can take a look.
Either way once I can I'll get my router into failsafe mode and checkout what interfaces are up and my dmesg to see if anything pops out
Also, how did you enter failsafe?
I pressed the reset button at boot
1 Like