A lot of random errors from brcmfmac #1289
Comments
|
The WLAN interface is being used as an AP by the way. Here's the diff of the shipped configuration of hostapd and the one I am using: |
|
So I downgraded to 4.19.83 again, have For the record the errors that shows up with .83 and .84 seem to differ. Also the interface seems to have a higher chance of disappearing with .83 (I don't exactly remember if it is ever gone with .84 / always gone with .83). |
|
Thanks - that narrows it down a bit. Unfortunately there are a large number of internal commits in that one revision, and I've not seen the error yet.
|
|
|
It would be useful to know if the errors only occur during throttled so running Another thing to try is setting this on config.txt The HDMI display resolution will also alter the voltage and clock requirements so it would be useful to know that. If you are running 4Kp60 try 4kp30 |
|
Right after boot. And although the errors differ a bit every boot, so far they always show up. I have systemd-networkd brings up the wlan interface on boot (hostapd is started on boot as well). Was going to see if the errors show up if it is not brought up but a miss in file copying stops the Pi from booting now, so I can't test further (maybe until next month). For the record, the linux-raspberrypi4 package in Arch Linux ARM pulls in some extra brcmfmac firmware of unknown origin. Not sure if it is relevant (was going to delete them and see if it helps, but again I can't boot it anymore, so) P.S. It's a headless server. Neither of the HDMI ports was being used. |
|
systemd has been described as "aggressively parallel", starting all the things as concurrently as possible. Current draw is at its maximum during boot, and the WiFi chip has been seen to suffer when the power supply is marginal, so that's something else to bear in mind along with @timg236's suggestions. |
|
Well, I will test with another power supply (even though the current one should be 5V/3A), and see if unplugging the usb 3.0 thumb drive helps (it was being used for the root filesystem), when I can. |
|
I was able to reproduce the issue and resolve it by setting over_voltage=1. If that fails try over_voltage=2 Please could you try just that change in your original config and report the values for 'vcgencmd measure voltage' with and without the config change. N.B. The voltage varies with clock speed and higher resolutions require higher clock speeds so it's important to be consistent there i.e. if my theory is correct then a higher resolution might appear to make WiFi work as a side effect of increasing clocks / voltage. |
|
@tomty89 Please could you try the over_voltage=6240 setting to see if that resolves the WiFi issue? |
|
While I couldn't test it yet anyway, why From https://www.raspberrypi.org/documentation/configuration/config-txt/overclocking.md: Also: Doesn't that mean So should I set |
|
Sorry that should be 6250. It’s possibly board specific but large numbers translate to microvolts and that’s the smallest step that makes sense on your board. |
|
FYI values over 500,000 (used to be 1,000,000) are treated as absolute values in microvolts, e.g. 880000 is 0.88V. |
|
I had this error a couple of days ago, completely filled up the SD card with syslog/kernel errors. If it reoccurs will take a closer look. Looks like the error report, if nothing else, needs to be rate limited. |
kernel: configs:2711: Add V3D clock support on arm64 kernel: config: arm64: Switch to upstream cpufreq driver kernel: arm/dts: 2711: Add 'pcie0' alias kernel: drm/v3d: Set dma_mask as well as coherent_dma_mask firmware: platform: Allow fixed voltage with avs_disable=1 firmware: EMMC: Use PLLD for EMMC for 250MHz host-clock See: #1289 firmware: platform: Round down effective frequencies when they exceed max See: #1290 firmware: arm_loader: Pass video mode via kernel command for composite See: #1285 firmware: Fix lens shading table generation buglet See: https://www.raspberrypi.org/forums/viewtopic.php?f=43&t=190586&start=75#p1534672 firmware: hdmi: Use RB2 timing for 2560x1440@60 if pixel clock is 241.5 MHz firmware: arm_dt: Look for ethernet0 before ethernet firmware: arm_dt: Set PCIe dma-ranges from memory size
kernel: configs:2711: Add V3D clock support on arm64 kernel: config: arm64: Switch to upstream cpufreq driver kernel: arm/dts: 2711: Add 'pcie0' alias kernel: drm/v3d: Set dma_mask as well as coherent_dma_mask firmware: platform: Allow fixed voltage with avs_disable=1 firmware: EMMC: Use PLLD for EMMC for 250MHz host-clock See: raspberrypi/firmware#1289 firmware: platform: Round down effective frequencies when they exceed max See: raspberrypi/firmware#1290 firmware: arm_loader: Pass video mode via kernel command for composite See: raspberrypi/firmware#1285 firmware: Fix lens shading table generation buglet See: https://www.raspberrypi.org/forums/viewtopic.php?f=43&t=190586&start=75#p1534672 firmware: hdmi: Use RB2 timing for 2560x1440@60 if pixel clock is 241.5 MHz firmware: arm_dt: Look for ethernet0 before ethernet firmware: arm_dt: Set PCIe dma-ranges from memory size
|
Can you try with latest rpi-update firmware? |
|
601d36d fixed it and no more error with current master either. Thanks! |
|
Great, thanks for confirming the fix :) |
Describe the bug
Samples:
To reproduce
Once I update the bootloader to 6e4a723, there are errors flooding (in an occasion the wlan interface won't even show up). Downgrading to b9f83b1 eliminates them. I am not sure if it has anything to do with the kernel update, but downgrading to 4.19.83 instead doesn't help.
System
OS is Arch Linux ARM
Device is Raspberry Pi 4 4GB
uname -a:The text was updated successfully, but these errors were encountered: