juliank
October 19, 2022, 2:11am
8660
I'm on the Robimarko's tagged build from Oct 17 and I'm having issue with dhcp client on the WAN interface (I have not used DHCP client prior so I do not know if it ever worked):
the link is up
tcpdump shows DHCP request from dhcpcd
only TX counter on the WAN interface is increasing (indicating DHCP request going out)
RX is at 0
Based on above it would be proper to conclude that the problem is with the DHCP server. However, when I connect a laptop with wireshark monitoring the port- I do not see DHCP requests on that port.
Any ideas?
vit0r
October 19, 2022, 7:34am
8661
I would first try replacing the ethernet cable.
So, I just spun up an image on my AX3600 and I can see DHCP discover packets constantly arriving on my notebook in wireshark.
juliank
October 19, 2022, 10:49am
8663
Cable is fine, I used the same cable with my laptop.
I should have mentioned that I'm using Dynalink WRX36. Thank you for looking into it - I will build a minimum image next.
Well, that's a whole different thing as somebody else reported that WAN/2.5G port is not working in the Dynalink thread.
Can I get the output of cat /sys/kernel/debug/clk/clk_summary
webzit
October 19, 2022, 11:42am
8665
Not only does 802.11s work fine, but both HP Elitebooks (AX200, AX201) only jump between the two AX3600s when needed. The WAF of this build has increased dramatically. Many thanks!
1 Like
are you using the latest build from Robi's repo ? I have an issue with the 2.5G... if I revert back to a previous build it works
rossini
October 19, 2022, 2:43pm
8668
So after around two days uptime of my AX3600 AP i can say, that the build r0-a10ca2b ipq807x-2022-10-17-1510 is working better than anything before.
4 Likes
Any idea of opensource packages that are able to stress test a router? namely on the number of clients they can support ? a quick coursory search on google seem to indicate there is none... i know this is probably not the right place to ask the question but bearing in mind the brain power in this domain here thought to ask the question
hello, the sysupgrade image from here brick my ax3600. It does not boot anymore.
no dhcp request
only LED is only orange
Sorft-brick revover like here seems not possible.
rossini
October 20, 2022, 10:16am
8673
i always test robimarkos builds.
I got to catch the problem that is causing ath11k to crash (I don't know if this helps..)
[55000.824234] qcom-q6v5-wcss-pil cd00000.q6v5_wcss: fatal error received:
[55000.824234] QC Image Version: QC_IMAGE_VERSION_STRING=WLAN.HK.2.5.0.1-01208-QCAHKSWPL_SILICONZ-1
[55000.824234] Image Variant : IMAGE_VARIANT_STRING=8074.wlanfw.eval_v2Q
[55000.824234]
[55000.824234] ar_wal_peer.c:2462 Assertion is_graceful_to_handle failedparam0 :zero, param1 :zero, param2 :zero.
[55000.824234] Thread ID : 0x00000060 Thread name : WLAN RT1 Process ID : 0
[55000.824234] Register:
[55000.824234] SP : 0x4c135128
[55000.824234] FP : 0x4c135130
[55000.824234] PC : 0x4b195a10
[55000.824234] SSR : 0x00000008
[55000.824234] BADVA : 0x00020000
[55000.824234] LR : 0x4b1951ac
[55000.824234]
[55000.824234] Stack Dump
[55000.824234] from : 0x4c135128
[55000.824234] to : 0x4c135980
[55000.824234]
[55000.872903] remoteproc remoteproc0: crash detected in cd00000.q6v5_wcss: type fatal error
[55000.895144] remoteproc remoteproc0: handling crash #1 in cd00000.q6v5_wcss
[55000.903262] remoteproc remoteproc0: recovering cd00000.q6v5_wcss
[55000.935997] remoteproc remoteproc0: stopped remote processor cd00000.q6v5_wcss
[55001.228888] ath11k c000000.wifi: failed to find peer 10:5a:17:48:99:eb on vdev 0 after creation
[55001.228946] ath11k c000000.wifi: failed to find peer vdev_id 0 addr 10:5a:17:48:99:eb in delete
[55001.236411] ath11k c000000.wifi: failed peer 10:5a:17:48:99:eb delete vdev_id 0 fallback ret -22
[55001.245117] ath11k c000000.wifi: Failed to add peer: 10:5a:17:48:99:eb for VDEV: 0
[55001.254129] ath11k c000000.wifi: Failed to add station: 10:5a:17:48:99:eb for VDEV: 0
[55001.348860] ath11k c000000.wifi: failed to send WMI_PEER_DELETE cmd
[55001.348898] ath11k c000000.wifi: failed to delete peer vdev_id 2 addr 10:5a:17:48:99:eb ret -108
[55001.353942] ath11k c000000.wifi: Failed to delete peer: 10:5a:17:48:99:eb for VDEV: 2
[55001.362999] ath11k c000000.wifi: Found peer entry 9e:9d:7e:75:17:62 n vdev 2 after it was supposedly removed
[55001.370765] ------------[ cut here ]------------
[55001.380582] WARNING: CPU: 0 PID: 2653 at sta_set_sinfo+0xba4/0xbc0 [mac80211]
[55001.385186] Modules linked in: pppoe ppp_async nft_fib_inet nf_flow_table_ipv6 nf_flow_table_ipv4 nf_flow_table_inet ath11k_ahb ath11k ath10k_pci ath10k_core ath wireguard pppox ppp_generic nft_reject_ipv6 nft_reject_ipv4 nft_reject_inet nft_reject nft_redir nft_quota nft_objref nft_numgen nft_nat nft_masq nft_log nft_limit nft_hash nft_flow_offload nft_fib_ipv6 nft_fib_ipv4 nft_fib nft_ct nft_counter nft_chain_nat nf_tables nf_nat nf_flow_table nf_conntrack mac80211 libchacha20poly1305 chacha_neon cfg80211 slhc qrtr_smd qrtr qmi_helpers poly1305_neon ns nfnetlink nf_reject_ipv6 nf_reject_ipv4 nf_log_syslog nf_defrag_ipv6 nf_defrag_ipv4 libcurve25519_generic libcrc32c libchacha hwmon crc_ccitt compat ip6_gre ip_gre gre ip6_udp_tunnel udp_tunnel ip6_tunnel tunnel6 ip_tunnel tun seqiv jitterentropy_rng drbg michael_mic hmac cmac leds_gpio xhci_plat_hcd xhci_pci xhci_hcd dwc3 dwc3_qcom qca_nss_dp qca_ssdk gpio_button_hotplug crc32c_generic
[55001.452793] CPU: 0 PID: 2653 Comm: hostapd Not tainted 5.15.72 #0
[55001.475029] Hardware name: Xiaomi AX3600 (DT)
[55001.481189] pstate: 60400005 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[55001.485533] pc : sta_set_sinfo+0xba4/0xbc0 [mac80211]
[55001.492301] lr : sta_set_sinfo+0xba0/0xbc0 [mac80211]
[55001.497509] sp : ffffffc00d383840
[55001.502539] x29: ffffffc00d383840 x28: ffffff8002892340 x27: ffffffc00d383dd0
[55001.505846] x26: 0000000000000000 x25: ffffffc008ba0740 x24: ffffffc000ac2000
[55001.512965] x23: ffffffc00d383a38 x22: ffffff8009498c00 x21: ffffff80099508c0
[55001.520082] x20: ffffff8005f78880 x19: ffffff801022f000 x18: 0000000000000196
[55001.527201] x17: 322076656476206e x16: 2032363a37313a35 x15: ffffffc008aa79c8
[55001.534318] x14: 00000000000004c2 x13: 0000000000000196 x12: ffffffc00d383358
[55001.541437] x11: ffffffc008aff9c8 x10: 00000000fffff000 x9 : ffffffc008aff9c8
[55001.548554] x8 : 0000000000000000 x7 : ffffffc008aa79c8 x6 : 0000000000008e08
[55001.555672] x5 : 00000000d8d39b7c x4 : 0000000000000000 x3 : ffffff8002892340
[55001.562790] x2 : 0000000000000000 x1 : ffffff8002892340 x0 : 00000000ffffff94
[55001.569910] Call trace:
[55001.577018] sta_set_sinfo+0xba4/0xbc0 [mac80211]
[55001.579279] sta_info_destroy_addr_bss+0x50/0x74 [mac80211]
[55001.584142] ieee80211_color_change_finish+0x1d68/0x2030 [mac80211]
[55001.589524] cfg80211_check_station_change+0x1448/0x47d0 [cfg80211]
[55001.595775] genl_family_rcv_msg_doit+0xb8/0x120
[55001.602023] genl_rcv_msg+0xd4/0x1d0
[55001.606880] netlink_rcv_skb+0x5c/0x130
[55001.610439] genl_rcv+0x38/0x50
[55001.613997] netlink_unicast+0x1ec/0x2e4
[55001.617124] netlink_sendmsg+0x1a4/0x3dc
[55001.621292] ____sys_sendmsg+0x280/0x2c4
[55001.625198] ___sys_sendmsg+0x84/0xf0
[55001.629103] __sys_sendmsg+0x48/0x90
[55001.632662] __arm64_sys_sendmsg+0x24/0x30
[55001.636308] invoke_syscall.constprop.0+0x5c/0x104
[55001.640216] do_el0_svc+0x74/0x16c
[55001.644988] el0_svc+0x18/0x54
[55001.648373] el0t_64_sync_handler+0xa4/0x130
[55001.651413] el0t_64_sync+0x184/0x188
[55001.655840] ---[ end trace ece96d9102f846dd ]---
[55004.297911] ath11k c000000.wifi: failed to send WMI_PDEV_BSS_CHAN_INFO_REQUEST cmd
[55006.238871] qcom-q6v5-wcss-pil cd00000.q6v5_wcss: start timed out
[55006.238925] remoteproc remoteproc0: can't start rproc cd00000.q6v5_wcss: -110
[55010.302758] ath11k_warn: 56 callbacks suppressed
[55010.302781] ath11k c000000.wifi: failed to send WMI_PDEV_BSS_CHAN_INFO_REQUEST cmd
[55010.306473] ath11k c000000.wifi: failed to send pdev bss chan info request
[55010.314135] ath11k c000000.wifi: failed to send WMI_PDEV_SET_PARAM cmd
[55010.320727] ath11k c000000.wifi: Failed to set beacon mode for VDEV: 0
[55010.327232] ath11k c000000.wifi: failed to send WMI_BCN_TMPL_CMDID
[55010.333739] ath11k c000000.wifi: failed to submit beacon template command: -108
[55010.339908] ath11k c000000.wifi: failed to update bcn template: -108
[55010.347098] ath11k c000000.wifi: failed to send WMI_VDEV_SET_PARAM_CMDID
[55010.353707] ath11k c000000.wifi: Failed to set dtim period for VDEV 0: -108
[55010.360469] ath11k c000000.wifi: failed to send WMI_VDEV_SET_PARAM_CMDID
[55016.370217] ath11k_warn: 47 callbacks suppressed
I dont have the firmware sources and thus whatever the Q6 FW dumps in its panic trace is really worthless
Ah ok. Sorry for the noise then
No worries, its still usefull for looking later if QCA maybe fixed it
1 Like
juliank
October 20, 2022, 11:16am
8678
Is CONFIG_PACKAGE_irqbalance=y recommended?
I see that default robimarko's build does not enable it. What are benefits and are there any issues if enabled?
In theory irqbalance is a good idea/practice. But IMO it makes mostly sense only if you push a system/router hard to the limits. E. g. you can see/feel an effect on VM hosting systems with irqbalance enabled/disabled. For a router I would enable it if you use e. g. SIP.
But I've enabled it anyway.
For this specific device install irqbalance and execute the following commands after in a shell:
sed -i 's/option enabled '\''0'\''/option enabled '\''1'\''/g' /etc/config/irqbalance
sed -i '/banirq/d' /etc/config/irqbalance
sed -i '/ignore/a \
list banirq '\''50'\'' \
list banirq '\''51'\'' \
list banirq '\''52'\'' \
list banirq '\''53'\'' \
list banirq '\''73'\'' \
list banirq '\''74'\'' \
list banirq '\''75'\'' ' /etc/config/irqbalance
sed -i '/exit 0/i \
sleep 20 \
#assign 4 rx interrupts to each core \
echo 8 > /proc/irq/50/smp_affinity \
echo 4 > /proc/irq/51/smp_affinity \
echo 2 > /proc/irq/52/smp_affinity \
echo 1 > /proc/irq/53/smp_affinity \
\
#assign 3 tcl completions to 3 CPUs \
echo 4 > /proc/irq/73/smp_affinity \
echo 2 > /proc/irq/74/smp_affinity \
echo 1 > /proc/irq/75/smp_affinity \
\
\' /etc/rc.local
/etc/init.d/irqbalance enable
/etc/init.d/irqbalance start
1 Like
juliank
October 20, 2022, 8:13pm
8680
Thank you for the detailed reply. I have installed your script, collectd interrupt stats and be monitoring '/proc/interrupts` Will see what happens
EDIT: looks like on my gateway router most used IRQ are: 13, 78 and 80.
equid0x
October 21, 2022, 1:49am
8681
Hahaha well, I could do that.
anon54515283:
equid0x:
Added strongswan, dawn, adblock, stubby, changed to ct-full-htt firmware, Sqm qos, and compile everything against openssl instead of wolfssl(wolfssl is missing something I need for the IPsec VPN), and collectd.
While I'm doing those things with other routers as well. I did this so far only with a few changes for this specific device. Even tho' I've "managed it" to kill the WAN device for whatever reason (eth0 was just not there). Didn't look into it what package or config was causing this.
Tons of changes makes it really hard to find the real cause of issues.
Well the good news is, since my latest pull and build the device has been pretty stable. The latest run was just shy of a week, with no oom issues and no other weirdness except strongswan which we know is an upstream problem which will have to wait for them. I only rebooted because internet went down which turned out to be an ISP problem.
Also, the qca-full-htt definitely appears to have resolved some of the roaming errors I was getting with ath11k.
Zram seems to have totally stopped any oom problems at present and memory usage looks stable on my config.
1 Like
