piControl i/o: missed x cycles - kernel log
-
- Posts: 7
- Joined: 26 Jul 2021, 06:54
piControl i/o: missed x cycles - kernel log
Hi all,
I am using RevPi Compact. When I look at the /var/log/kern.log, i am seeing a flood of messages "piControl i/o: missed xx cycles". Why is this error? How can we stop this error?
-Christian
I am using RevPi Compact. When I look at the /var/log/kern.log, i am seeing a flood of messages "piControl i/o: missed xx cycles". Why is this error? How can we stop this error?
-Christian
-
- Posts: 7
- Joined: 26 Jul 2021, 06:54
Re: piControl i/o: missed x cycles - kernel log
anyone else is also having this error/warning in their kernel log? The file is in /var/log/kern.log
I notice that at idle, piControl driver is taking up CPU about 20-25% - which i think it's kind of high.
I notice that at idle, piControl driver is taking up CPU about 20-25% - which i think it's kind of high.
Re: piControl i/o: missed x cycles - kernel log
CPU load is kinda normal because there is a lot of traffic on the RS485 and the DMA / FIFO HW of the Broadcom chip has limitations. I would also like to know what the message means, is this related to the piBridge Messages gotten corrupt on my test setup from almost 2 years ago? Was there more improvements made on the driver?
- p.rosenberger
- Posts: 91
- Joined: 03 Jul 2020, 11:07
Re: piControl i/o: missed x cycles - kernel log
Hi Christian,
this a known issue with RevPi Compact and the 4.19 kernel. There seems to be a regression with the 4.19 kernel. We can't currently narrow it down. It is a warning that the PiControl driver can't run the cycles as fast as it tries. From our current point of view this should not impact you. This problem is also exclusive to the RevPi Compact. As only the Compact is using a pretty short cycle time.
If you don't experience any other problems, it should be safe to ignore the message for now. We are still investigating the issue. But it will likely take a longer time to find a solution for this.
The CPU load is normal, as PiCOntrol runs a loop to read the data from the H/W and write it to the processimage.
Best Regards
Philipp
this a known issue with RevPi Compact and the 4.19 kernel. There seems to be a regression with the 4.19 kernel. We can't currently narrow it down. It is a warning that the PiControl driver can't run the cycles as fast as it tries. From our current point of view this should not impact you. This problem is also exclusive to the RevPi Compact. As only the Compact is using a pretty short cycle time.
If you don't experience any other problems, it should be safe to ignore the message for now. We are still investigating the issue. But it will likely take a longer time to find a solution for this.
The CPU load is normal, as PiCOntrol runs a loop to read the data from the H/W and write it to the processimage.
Best Regards
Philipp
-
- Posts: 7
- Joined: 26 Jul 2021, 06:54
Re: piControl i/o: missed x cycles - kernel log
Hi Philipp,
Thanks for the reply. I do have issue reported as another topic - which may or may not be related https://revolution.kunbus.de/forum/view ... f=6&t=2997. It is the fact that analog input read from piTest intermittently fail. I am stuck there in that topic. (Currently i had to implement a "retry" to read several times with some sleep - until the read is successful - when the failure happens - but this is really not ideal solution).
Do you think these are possibly related, or they would not be connected?
-Christian
Thanks for the reply. I do have issue reported as another topic - which may or may not be related https://revolution.kunbus.de/forum/view ... f=6&t=2997. It is the fact that analog input read from piTest intermittently fail. I am stuck there in that topic. (Currently i had to implement a "retry" to read several times with some sleep - until the read is successful - when the failure happens - but this is really not ideal solution).
Do you think these are possibly related, or they would not be connected?
-Christian
- p.rosenberger
- Posts: 91
- Joined: 03 Jul 2020, 11:07
Re: piControl i/o: missed x cycles - kernel log
Hi Christian,
I don't think they are connected. The missing cycle issue is related to the digital I/Os. The analog I/Os are sampled not as fast. We didn't see the missing cycle related to these.
Best Regards
Philipp
I don't think they are connected. The missing cycle issue is related to the digital I/Os. The analog I/Os are sampled not as fast. We didn't see the missing cycle related to these.
Best Regards
Philipp
Re: piControl i/o: missed x cycles - kernel log
Hello, I'm having the same on a RevPi compact, but kernel running is 5.10
Code: Select all
uname -a
Linux mp-zhs1 5.10.17-v7l+ #1414 SMP Fri Apr 30 13:20:47 BST 2021 armv7l GNU/Linux
Re: piControl i/o: missed x cycles - kernel log
Hi sherifeid,
we haven't released a kernel 5.10 yet, so I'm guessing that you run a custom build. There might be some issues regarding piControl and kernel 5.10, so please stay with our officially supported versions as long as you need piControl support..
Nicolai
we haven't released a kernel 5.10 yet, so I'm guessing that you run a custom build. There might be some issues regarding piControl and kernel 5.10, so please stay with our officially supported versions as long as you need piControl support..
Nicolai
Re: piControl i/o: missed x cycles - kernel log
apologies, I was running uname on a different raspberry pi
Re: piControl i/o: missed x cycles - kernel log
Hello all,
I am having the same issue, namely a flood of "piControl: piControl I/O: missed x cycles" in the journalctl and dmesg logs.
My hardware is a Revolution Pi and the image I flashed is the "RevPi Bullseye (08/2023) (ID: 7324)".
Kernel is 5.10.
I'd like to get rid of those error messages, but thankfully they don't affect the Anlog I/Os or the Digital ones operations in our setup.
Has there been any progress on this matter?
Thanks!
I am having the same issue, namely a flood of "piControl: piControl I/O: missed x cycles" in the journalctl and dmesg logs.
My hardware is a Revolution Pi and the image I flashed is the "RevPi Bullseye (08/2023) (ID: 7324)".
Kernel is 5.10.
I'd like to get rid of those error messages, but thankfully they don't affect the Anlog I/Os or the Digital ones operations in our setup.
Has there been any progress on this matter?
Thanks!