Turris OS 7.0 is in rc!

Same here - but different :slight_smile:

Error notifications

Updater execution failed:
INFO:Target Turris OS: 7.0.0
line not found
ERROR:
inconsistent: Requested package kmod-usb3 that is not available.
line not found
line not found
line not found
line not found

1 Like

Let me share experience from today:

  • WiFi seems to be stable. No crash during the night. So non -ct driver works fine.
  • I have 80/20 Mbit Internet and I am not able to reach more than roughly 60 Mbit because of CPU saturation. I played with packet steering, HW/SW offloading and it goes sometimes to 65Mbit but still a lot of CPU usage. I also tried to switch to nftables/firewall4 (which resolves Firewall Status empty page) but performance was not affected.
  • During download there is still about 15% or even 30% of cpu usage by python3 foris-controller which is strange.
1 Like

Arenā€™t you running Transmission on the router? That seems to eat the CPU on 7.0.0. My speed test achieved full 200/100 on Omnia after the update.

No, I donā€™t have any additional app installed.

cat /proc/interrupts
           CPU0       CPU1
 11:    1713742    1807882     GICv3  30 Level     arch_timer
 14:          0          0     GICv3  23 Level     arm-pmu
 15:      28089          0     GICv3  32 Level     d0010600.spi
 16:        145          0     GICv3  33 Level     d0011000.i2c
 18:        184          0     GICv3  44 Level     d0012000.serial
 19:          0          0     GICv3  45 Level     d0012000.serial
 20:    2925898          0     GICv3  74 Level     eth0
 21:          0          0   advk-RP   0 Edge      pciehp
 22:          0          0     GICv3  35 Level     xhci-hcd:usb2
 23:         24          0     GICv3  49 Level     ehci_hcd:usb1
 30:         69          0     GICv3  50 Level     armada-37xx-rwtm-mailbox
 31:     323582          0     GICv3  57 Level     mmc1
 32:       9760          0     GICv3  58 Level     mmc0
 33:    1669348          0     GICv3  61 Level     advk-pcie
 51:          2          0     GICv3  79 Level     d0060900.xor
 52:          2          0     GICv3  80 Level     d0060900.xor
 53:          6          0     GPIO2   5 Edge      moxtet
 54:          5          0  moxtet-peridot.1#0   8 Edge      mv88e6xxx-d0032004.mdio-mii:10
 59:    1779906          0     GICv3  77 Level     eth1
 63:          0          0  mv88e6xxx-g1   3 Edge      mv88e6xxx-d0032004.mdio-mii:10-g1-atu-prob
 65:          0          0  mv88e6xxx-g1   5 Edge      mv88e6xxx-d0032004.mdio-mii:10-g1-vtu-prob
 67:          5          0  mv88e6xxx-g1   7 Edge      mv88e6xxx-d0032004.mdio-mii:10-g2
 70:          2          0  mv88e6xxx-g2   1 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:01
 71:          2          0  mv88e6xxx-g2   2 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:02
 72:          0          0  mv88e6xxx-g2   3 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:03
 73:          0          0  mv88e6xxx-g2   4 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:04
 74:          0          0  mv88e6xxx-g2   5 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:05
 75:          0          0  mv88e6xxx-g2   6 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:06
 76:          0          0  mv88e6xxx-g2   7 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:07
 77:          0          0  mv88e6xxx-g2   8 Edge      !soc!internal-regs@d0000000!mdio@32004!switch0@10!mdio:08
 78:          1          0  mv88e6xxx-g2   9 Edge      mv88e6xxx-d0032004.mdio-mii:10-serdes-9
 84:          0          0  mv88e6xxx-g2  15 Edge      mv88e6xxx-d0032004.mdio-mii:10-watchdog
 85:          0          0     GPIO1  10 Edge      d00d8000.sdhci cd
 86:          0          0     GPIO2  20 Edge      gpio-keys
 87:    1669347          0  advk-MSI 1572864 Edge      ath10k_pci
IPI0:      3793       3855       Rescheduling interrupts
IPI1:    170485    1559102       Function call interrupts
IPI2:         0          0       CPU stop interrupts
IPI3:         0          0       CPU stop (for crash dump) interrupts
IPI4:         0          0       Timer broadcast interrupts
IPI5:         0          0       IRQ work interrupts
IPI6:         0          0       CPU wake-up interrupts
Err:

Interrupts are not spread across two cores so it is more likely to saturate one core - CPU0.

root@turris:~# cat /proc/irq/20/smp_affinity
3
root@turris:~# cat /proc/irq/59/smp_affinity
3
root@turris:~# cat /proc/irq/default_smp_affinity
3

Changing eth0 IRQ to CPU1 seems to be good option - decrease latency about few ms and spread load more evenly.
echo 2 > /proc/irq/20/smp_affinity
(that is probably correct as I donā€™t have irqbalance installed)

1 Like

But that was the default even before the update, right?

Probably yes. I havenā€™t checked details before update and now I am trying to understand it and fine-tune it. I am sorry for confusion.

As I have sore throat and cold I can spent time during weekend with debugging it :smiley:

Hmmm, seems like luci-app-vpn-policy-routing is not available in Turris OS 7.0. You have it installed manually? I donā€™t see it in upstream repo, so maybe it was dropped? Try to find some replacement and uninstall it and you should be able to update.

Hmmm, this might be related to us not switching to the nftables yet. Do you have some special packages installed? What pkglists do you have enabled? My status page show the status. Iā€™ll try to break it somehow.

Were you able to use the 80/20 connection fully with Turris OS 6.x?

Hi,

what device is this?

I had it installed ā€œalwaysā€ and it worked great. The whole Omnia worked so well that I didnā€™t look at it for several years. I donā€™t even remember how I installed and configured it :frowning:

Fā€¦ck
Now Iā€™m devastated and tear my hair out when I think that I have to mess with the router because Iā€™m too thin :frowning:

Even though I know I should follow the release cycle, now Iā€™m wondering whether to leave it all alone and not update.

Can anyone help me find a replacement and tell me how to configure it?

@Tomov Welcome to docs.openwrt.melmac.net! | stangriā€™s OpenWrt Packages Documentation. Just read that pbr replaced vpn-policy-routing

1 Like

Donā€™t you have any schnapps snapshots you can rollback to?

1 Like

Thank you. This is great news for me!

I donā€™t understand what exactly I should do.

I read about migration here , but I would like to make sure I do it correctly:

  1. Remove the vpn-policy-routing package
  2. Update my Omnia
  3. Install the luci-app-pbr package

Then all my vpn-policy-routing settings will be preserved?

Thank you for reminding me about this wonderful tool :slight_smile:

That is related to nftables for sure as it started to work once I enabled them.

Packages

-- This is automatically generated file managed by opkg wrapper script. Please don't edit!
Install("luci-app-sqm", "turris-nor-update")

Uninstall("ath10k-firmware-qca988x-ct", "kmod-ath10k-ct", { priority = 60 })
Install("ath10k-firmware-qca988x", "kmod-ath10k", "firewall4")
Package("firewall", { virtual = true })

Startup script

# Re-route eth0 IRQ to CPU1
echo 2 > /proc/irq/20/smp_affinity

Usual setup

  • two wifis, 2.4GHz SDIO for IoT, 5GHz for phones and laptops (up to 15 devices altogether)
  • 8ports switch, only TV connected right now
  • wan with NAT (no additional or custom FW rules)

Thank you for your effort on TOS 7.0. With this settings I donā€™t have any stability issues. Both wifi are rock solid, no errors in logs.

That was my wrong observation. I have IPTV from my provider and it was excluded from internet speed. Now IPTV also counts so there is some penalty for internet speed once you have your TV on. So If I switched my TV off it is above 70Mbit/s which seems to be fine.

Unfortunately it does not explain heavy cpu usage during download and also cpu usage by forris-controller during download. I also noticed that sometimes kresd peaks up.

1 Like

Hey,

we are running TurrisOS on our Turris Omnia (blue one).

Best regards

SDIO forever.

Driver update not exist X drive update is good idea. Is the truth out there somewhere?