Failed PPR?

[DanNuke86]

Does anyone know what the below logs are that keep flooding my syslog? I've searched on here and google but can't find anything related. I've had these intermittenly since installing the 386.2 release/betas.

Any help would be much appreciated!

kernel: CONSOLE: 072192.429 ppr_get_vht_mcs: Failed ppr_t:0046d728 BW:2 Nss:1 mode:0 TX_Chain:4 rateset:00175f3c

 

[Trent Hubbert]

I have the same in the log files.

 

[Trent Hubbert]

Any one else? I think these are DNS related. Still testing.

 

[ColinTaylor]

It's something to do with WiFi, not DNS.

 

[Trent Hubbert]

Any idea on why this happens. I tried changing various wifi setting but nothing seems to help. What does the error message mean? kernel: ppr_get_vht_mcs: Failed ppr_t

 

[ColinTaylor]

Are you using 160 MHz bandwidth on the router?

Do you have any Apple WiFi clients?

 

[Trent Hubbert]

Yes I have Apple iPhones (2). I have turned off the privacy so the mac address does not change. I also have 160 MHz on 5G. I have 5 AC clients connected at 80MHz and 1 AX client connected at 160MHz on the 5G. the 2.4G is 20Mhz and all N clients and the iPhones as AX clients.

 

[ColinTaylor]

I suggest you turn off the iPhones one at a time to see if one or both of them are the cause.

 

[Trent Hubbert]

No change. I have both turned off and the message still appear.

 

[ColinTaylor]

OK. For expediency of debugging trying temporarily changing the bandwidth from 160MHz to 80MHz. Do the messages disappear now?

 

[Trent Hubbert]

OK. For expediency of debugging trying temporarily changing the bandwidth from 160MHz to 80MHz. Do the messages disappear now?

No. Changed to 80MHz and the messages still appear.

 

[ColinTaylor]

No. Changed to 80MHz and the messages still appear.

OK thanks for testing. That eliminates all the theories I had in mind.

 

[Skiron]

I build my own kernels on my older laptops etc. and was intrigued by this. Doing a google on that symbol reveals that there is a patch set for embedded kernels that introduces PREEMPT RT, and the symbol ppr_get_vht_mcs appears to be part of that. You can turn it off via a kernel boot time argument to test if it is that, but on a router I don't know how you do that (have not investigated).

Further reading seems to link to a flaky Wifi/Ethernet connection (i.e. the connection bounces up/down rapidly), but other than that without getting the patch set and reading it to confirm, I don't know.

Hope that somewhat helps.

 

[ColinTaylor]

I build my own kernels on my older laptops etc. and was intrigued by this. Doing a google on that symbol reveals that there is a patch set for embedded kernels that introduces PREEMPT RT, and the symbol ppr_get_vht_mcs appears to be part of that. You can turn it off via a kernel boot time argument to test if it is that, but on a router I don't know how you do that (have not investigated).

Further reading seems to link to a flaky Wifi/Ethernet connection (i.e. the connection bounces up/down rapidly), but other than that without getting the patch set and reading it to confirm, I don't know.

Hope that somewhat helps.

It's not related to the realtime kernel. It's part of Broadcom's "Power-per-rate API" and comes from their wireless driver.

/* Get the VHT MCS values for the group specified by Nss, with the given bw and tx chains */
int ppr_get_vht_mcs(ppr_t* pprptr, wl_tx_bw_t bw, wl_tx_nss_t Nss, wl_tx_mode_t mode,
 wl_tx_chains_t tx_chains, ppr_vht_mcs_rateset_t* mcs);