I am using the latest 18.06 (Oct 12) build and it is rock solid. Just like the ones before. Running it on two R7800’s in two locations. Samsung phones/tablets and iPhones/iPads have no issues whatsoever.
actual master is (imho) unusable with samsung devices
i've reverted back to a 75something
So 18.06 is ok but master has bugs with WiFi on certain devices?
The "ath10k-CT" WiFi firmware version in master is not playing nicely with all devices.
18.06 still has the old regular ath10k.
Standard master still has the regular ath10k and kmod-ath10k packages available and they've just updated to the latest firmware.
I ran CT for a while and didn't have much issues (maybe iPhone was a bit slow) but now when master changed to CT I feel that wireless is much slower, AppleTV took some time to start movies etc, even longer to resume if you paused during watching.
When I first ran the CT firmware i don't think i had kmod-ath10k-ct tho', i ran kmod-ath10k with the CT firmware so maybe that'S why it was "working" decent for me.
Anyway, now I run master with the normal kmod-ath10k and the 3.6.0.1 firmware and everything seems much faster.
Has anybody noticed lately any slowness in the compilation process?
I was traveling for a week and after returning home, I think that compilling master suddenly takes twice as long as earlier. So I wonder if something has changed in my buildhost (Ubuntu 18.04 in Virtualbox) or something more generic in Openwrt build process.
I did notice significant build time increases with meltdown/ spectre fixes being applied to the (buildhost) kernel and the jump to newer gcc versions, additionally enabling LTO for selected packages also increases their build time, but neither of those are as recent as a couple of weeks.
1 Like
Fresh Building (incl make download) master with your scripts takes 73 minutes on a Lenovo i5 notebook running Ubuntu 18.10.
I cannot confirm a slowness.
Can't really say, compiling always took time on my core2duo server.
But like i said earlier, updating the telephony feed times out so I removed this line from updatenmake.sh:
[ "$?" -ne 0 ] && echo "Updating the feeds failed." && exit 1
Interesting, what CPU?
I noticed that build time of my R7800 build (after make clean but using the existing toolchain, so just the plain firmware), grew from some 25 minutes (2 weeks ago) to some 80 minutes (now). In addition to Openwrt source commits, there have been both some minor updates to Ubuntu 18.04 and also the jump from VirtualBox 5.2.18 to 5.2.20.
So I have wondered if this is just something caused by new Virtualbox version or some Ubuntu updates, or something in the Openwrt.
Based on the feedback so far, I think that this is likely something related to Virtualbox and Ubuntu. Will have to test vith older versions.
I think that I have found the culprit for my recent compiling slowness:
It is likely not Virtualbox or Ubuntu, but Windows 10 trying to be secure.
There is a new Windows 10 patch KB4100347 for Intel processor microcode to tackle the Spectre v2 vulnerability CVE 2017-5715. Windows 10 installed that patch for me on 12.10.2018. The build jobs before that were quick, and the builds after that have been a nightmare.
Reverting that patch decreased the Openwrt compile time from 80+ minutes back to 25-30 minutes on Ubuntu 18.04 in Virtualbox 5.2.18.
Not sure if that is due to specifically Ubuntu & Virtualbox & Windows10 combination with my hardware, but the patch kills the compiling performance in the Ubuntu guest in my use case. I will likely need to update the 4-year old Haswell motherboard & CPU to Spectre-proof new hardware versions in intermediate future (once Intel really gets the.
I am guessing, but likely the same microcode update will sooner or later hit the Linux distros, so this is nor just a Windows 10 issue. (but of course the extreme slowness might be due to Windows 10 - Virtualbox - Ubuntu combination)
Ps. I have blocked the update for now, so I am vulnerable but speedy 
1 Like
think the problem is with virtual box + win 10
with bash on windows i don't have such problems
1 Like
Interesting.
Have you checked if you have that patch installed in Windows? (Windows has likely installed it automatically, so likely yes, but just inquiring... ) And what is your CPU's age/family?
@ansuel
I tested the bash in Windows in 2016 (see https://forum.archive.openwrt.org/viewtopic.php?id=67204&p=1#p336161 ), but I did not like it as at that time it forced to use the user profile dir in the main Windows drive for Openwrt.
Are you using that dir? or have you somehow formatted a separate ext2/3/4 drive?
use a separate drive with ext2/3/4 is not possible as windows doesn't support that file system
But afaik now we are able to save the linux dist in a different directory, will search about it.
it looks like it's possible with this https://stackoverflow.com/questions/38779801/move-wsl-bash-on-windows-root-filesystem-to-another-hard-drive
or the official way https://github.com/Microsoft/WSL-distrolauncher
I have been losing internet on the 5ghz every few minutes. Tried 18.06 and master, and also going back to old versions that I had. Manually overwriting the ath10k firmware-5.bin file mentioned above fixed it.
I have a lot of errors with the ath10k though. These errors were here before and after overwriting the file.
[ 16.260569] ath10k_pci 0000:01:00.0: assign IRQ: got 67
[ 16.260596] ath10k 4.16 driver, optimized for CT firmware, probing pci device: 0x46.
[ 16.261444] ath10k_pci 0000:01:00.0: enabling device (0140 -> 0142)
[ 16.267631] ath10k_pci 0000:01:00.0: enabling bus mastering
[ 16.268140] ath10k_pci 0000:01:00.0: pci irq msi oper_irq_mode 2 irq_mode 0 reset_mode 0
[ 16.448788] ath10k_pci 0000:01:00.0: Direct firmware load for ath10k/fwcfg-pci-0000:01:00.0.txt failed with error -2
[ 16.448823] ath10k_pci 0000:01:00.0: Falling back to user helper
[ 17.226929] firmware ath10k!fwcfg-pci-0000:01:00.0.txt: firmware_loading_store: map pages failed
[ 17.235340] ath10k_pci 0000:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/ct-firmware-5.bin failed with error -2
[ 17.235377] ath10k_pci 0000:01:00.0: Falling back to user helper
[ 17.281069] firmware ath10k!QCA9984!hw1.0!ct-firmware-5.bin: firmware_loading_store: map pages failed
[ 17.281406] ath10k_pci 0000:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/ct-firmware-2.bin failed with error -2
[ 17.289389] ath10k_pci 0000:01:00.0: Falling back to user helper
[ 17.329343] firmware ath10k!QCA9984!hw1.0!ct-firmware-2.bin: firmware_loading_store: map pages failed
[ 17.329671] ath10k_pci 0000:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/firmware-6.bin failed with error -2
[ 17.337658] ath10k_pci 0000:01:00.0: Falling back to user helper
[ 17.378097] firmware ath10k!QCA9984!hw1.0!firmware-6.bin: firmware_loading_store: map pages failed
[ 17.618511] ath10k_pci 0000:01:00.0: qca9984/qca9994 hw1.0 target 0x01000000 chip_id 0x00000000 sub 168c:cafe
[ 17.618548] ath10k_pci 0000:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 0
[ 17.629823] ath10k_pci 0000:01:00.0: firmware ver 10.4-3.6.0.1-00003 api 5 features no-p2p,mfp,peer-flow-ctrl,btcoex-param,allows-mesh-bcast,no-ps crc32 266d70b0
[ 19.908985] ath10k_pci 0000:01:00.0: board_file api 2 bmi_id 0:1 crc32 dd6d039c
[ 25.743693] ath10k_pci 0000:01:00.0: 10.4 wmi init: vdevs: 16 peers: 528 tid: 102
[ 25.743728] ath10k_pci 0000:01:00.0: msdu-desc: 2500 skid: 32
[ 25.827727] ath10k_pci 0000:01:00.0: htt-ver 2.2 wmi-op 6 htt-op 4 cal pre-cal-file max-sta 512 raw 0 hwcrypto 1
[ 25.924235] ath: EEPROM regdomain: 0x0
[ 25.924251] ath: EEPROM indicates default country code should be used
[ 25.924260] ath: doing EEPROM country->regdmn map search
[ 25.924279] ath: country maps to regdmn code: 0x3a
[ 25.924292] ath: Country alpha2 being used: US
[ 25.924302] ath: Regpair used: 0x3a
[ 25.929826] ath10k_pci 0001:01:00.0: assign IRQ: got 100
[ 25.929875] ath10k 4.16 driver, optimized for CT firmware, probing pci device: 0x46.
[ 25.930949] ath10k_pci 0001:01:00.0: enabling device (0140 -> 0142)
[ 25.936947] ath10k_pci 0001:01:00.0: enabling bus mastering
[ 25.937640] ath10k_pci 0001:01:00.0: pci irq msi oper_irq_mode 2 irq_mode 0 reset_mode 0
[ 26.105280] ath10k_pci 0001:01:00.0: Direct firmware load for ath10k/fwcfg-pci-0001:01:00.0.txt failed with error -2
[ 26.105321] ath10k_pci 0001:01:00.0: Falling back to user helper
[ 26.355072] firmware ath10k!fwcfg-pci-0001:01:00.0.txt: firmware_loading_store: map pages failed
[ 26.357661] ath10k_pci 0001:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/ct-firmware-5.bin failed with error -2
[ 26.362938] ath10k_pci 0001:01:00.0: Falling back to user helper
[ 26.409877] firmware ath10k!QCA9984!hw1.0!ct-firmware-5.bin: firmware_loading_store: map pages failed
[ 26.410117] ath10k_pci 0001:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/ct-firmware-2.bin failed with error -2
[ 26.418220] ath10k_pci 0001:01:00.0: Falling back to user helper
[ 26.465555] firmware ath10k!QCA9984!hw1.0!ct-firmware-2.bin: firmware_loading_store: map pages failed
[ 26.465691] ath10k_pci 0001:01:00.0: Direct firmware load for ath10k/QCA9984/hw1.0/firmware-6.bin failed with error -2
[ 26.473820] ath10k_pci 0001:01:00.0: Falling back to user helper
[ 26.523042] firmware ath10k!QCA9984!hw1.0!firmware-6.bin: firmware_loading_store: map pages failed
[ 26.523222] ath10k_pci 0001:01:00.0: qca9984/qca9994 hw1.0 target 0x01000000 chip_id 0x00000000 sub 168c:cafe
[ 26.531056] ath10k_pci 0001:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 0
[ 26.544851] ath10k_pci 0001:01:00.0: firmware ver 10.4-3.6.0.1-00003 api 5 features no-p2p,mfp,peer-flow-ctrl,btcoex-param,allows-mesh-bcast,no-ps crc32 266d70b0
[ 28.817032] ath10k_pci 0001:01:00.0: board_file api 2 bmi_id 0:2 crc32 dd6d039c
[ 34.656476] ath10k_pci 0001:01:00.0: 10.4 wmi init: vdevs: 16 peers: 528 tid: 102
[ 34.656507] ath10k_pci 0001:01:00.0: msdu-desc: 2500 skid: 32
[ 34.742463] ath10k_pci 0001:01:00.0: htt-ver 2.2 wmi-op 6 htt-op 4 cal pre-cal-file max-sta 512 raw 0 hwcrypto 1
Maybe also this one is related to the r7800 master build?
On iOS and Ubuntu I get sporadically (but if, then continuously) a loop of
"deauthenticated due to local deauth request" and "associated"
Sometimes for 1-2 minute, sometimes for 10-20 minutes.
Other clients being connected for a longer time are not affected, but if, all devices coming into the reach of the radio are affected. I thought it is a local issue, but I realised that Ubuntu and iOS was affected at the same time.
My cat really likes this router (and firmware 
), always sleeps next to it. 
I'm considering wallmounting it tho', because the cat pushes the router off the table. And it might not the be the most ideal place, behind 2 computer screens. Sorry for OT. 
2 Likes
still just getting 300 of 400 mbit/s with the latest master. WRT1900ACSv2 with Davids Owrt gives me full speed. Any clue how to solve this?