Raspberry Pi Forums

mangodan2003

Stop using sudo rpi-update since you clearly don't understand what it delivers (experimental kernel and experimental bootcode) Why do people keep using rpi-update? rpi-update was a recommended step by the mods in at least one related thread. It was worth a try. No other way to get the latest driver...

mangodan2003

It seems to me the problem may have been resolved with the latest from rpi-update 4.14.41-v7+ However it introduces a new problem whereby the kernel often crashes very early in startup. I had previously posted the problem persisted and seemed to be resolved by ethtool --offload eth0 rx off tx off (a...

mangodan2003

since the latest rpi-update my pis frequently crash early on in kernel start up. Succeeds maybe 1 in 4 or 5 tries.

4.14.39-v7+

However when it does boot the problem with my rsync script causing the rpi to lockup seems to be resolved.

mangodan2003

TL;DR rsyncing a system via LAN to local SD card is a good way to show up if the system is vulnerable to locking up. (rpi-update'd only an hour or so ago) I'm also having problems with frequent lockups. I've just run rpi-update and just been able to reproduce a lockup. I'm running a Master and 8 sla...

mangodan2003

I'm not sure what you mean by "No" but reread the entire OP. Yes you're right that was a hasty reaction. I am irked that netbooting (in the classic sense, SD-free) was publicised when it does not work in real life. That cost me plenty of time and rework on a project. If the RPF or whoever wasn't cl...

mangodan2003

I'm not sure what you mean by "No" but reread the entire OP.

mangodan2003

For reference I also posted here where the OP more closely matches the reason for my post. It is now working for us with one caveat.
viewtopic.php?f=29&t=183732&p=1177161#p1177161

mangodan2003

I am sorry i cannot help much but this is now working for us and I have absolutely no idea why. On day one they behaved as I described before but 3 days later they have all been booting fine every time! So far we have had upto 8 slave pis all booting from the one master and they all boot surprisingl...

mangodan2003

Just stumbled across this thread.

It really is quite easy to get a low ish (~100-200mS) latency reasonably decent FPS (10-20) from a DSLR livepreview into a webbrowser using python and libgphoto2.

If anyone is still interested ill try post a small demo when I get a spare mo.

mangodan2003

Any progress with this? We are seeing the same or similar problem. I just posted regarding it in another thread here viewtopic.php?f=29&t=181409&p=1177159&h ... t#p1177159

mangodan2003

Did you get anywhere with this? We are also trying to boot multiple clients from one master. All are RPI3. At present we are using the bootloader.bin on the SD card method in order to avoid potential problems with the baked in bootloader. All the pis are connected to a Netgear GS116 16 port Gigabit ...

mangodan2003

This seems to now be fixed in the latest from rpi-update

mangodan2003

Simple test - at offset 0940 the message was received corrupted. On the AP pi [email protected]:~ $ netcat -lk -p 1234 | hexdump -C 00000000 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 0a |AAAAAAAAAAAAAAA.| * 00000940 41 41 41 41 41 41 77 00 88 ff 73 00 00 01 00 00 |AAAAAAw...s.....| 00000950 41 41 41...

mangodan2003

Re the hexdump - I didn't post this last night as i didn't believe what I saw putting it down to it being late and me being tired - but whilst trying to clear my head of fruity deserts i realised what I saw explains a few things so ill put it here and confirm later. I used tcpdump and its -X option....

mangodan2003

Yes it is indeed very odd - and I can only assume that the combination of bridged interfaces, rpi3 (if relevant) and 4.9 is rare enough that has only been noticed by a few of us to date. Many of the pi AP examples online are not really APs but NAT routers and thus it may not be a very common setup. ...

mangodan2003

Apologies - this will be a bit verbose but i want to paint as clear a picture as possible... I will refer to two RPis, pi1 an pi2 - they are both Rpi3's - and a third "other host" (not a pi - but a wandboard). Thanks to your patch for the "kevent dropped" to 4.9.23 issue pi1 has now been running sta...

mangodan2003

I'm not sure that it is clear that there are (as far as i can see anyway) two separate issues here so just want to clarify - the one I refer to above is the one originally mentioned in this thread : That the last few bytes of data in TCP messages sometimes get corrupted when wlan0 and eth0 are bridg...

mangodan2003

TLDR; 4.4.50 doesn't seem to have this (data corruption) issue 4.9.23 definitely does (from sudo rpi-update 06c104c37348e104d7bc108b8ad19697df93b589) The longer version : I did a clean start with a clean Jessie lite 2017-04-10. I checked that default out of the box setup is fine, then went a head an...

mangodan2003

This was posted a while ago but maybe the following is still useful as i had a similar issue. You have probably inadvertently created a _file_ /dev/sdX while the device was not there, the fs limit for such files being created under /dev seems to be 10 meg. remove the file with rm -f /dev/sdX reconne...

mangodan2003

Seems to me like something todo with GPIO is broken in sysfs kernel module, once the int fires once it fires forever, like the corresponding flag is never cleared. I thought I was doing something wrong in my python code but then I found this C example which is producing the same results : https://de...

mangodan2003

k, that got me vcdbg and I was able to confirm no device tree trailer.

I'm investigating using GPIO via sysfs instead, but interrupts using poll don't appear to work, if I can't get past this ill rebuild kernel and add the device tree and go back to the RPi.gpio method.

Thanks

mangodan2003

I have the PPA already, and most the packages in it show up, but oddly the one I assume contains the vcdbg tool (raspberrypi-vc) does not :/ i've done apt-get update just to check i'm up-to-date but it doesn't show :/.

mangodan2003

Ok, could you confirm if your kernel is / was 3.18.0-20-rpi2?

mangodan2003

The vchiq udev should be installed at /lib/udev/rules.d/10-local-rpi.rules when libraspberrypi0 is installed (though it may take a "udevadm trigger" or reboot to actually apply). Ah, this is correct. When I looked in the rules.d directory earlier, there were only a few files (10-local-rpi.rules not...

Search found 24 matches

Re: Ethernet Problems PI3 B+

Re: Raspberry Pi 3 B+ lockups

Re: Ethernet Problems PI3 B+

Re: Raspberry Pi 3 B+ lockups

Re: Booting multiple pxe clients

Re: Booting multiple pxe clients

Re: Network boot multiple clients

Re: Booting multiple pxe clients

Re: Canon DSLR Gphoto2 liveview

Re: Booting multiple pxe clients

Re: Network boot multiple clients

Re: ethernet bridging / strange packet corruption

Re: ethernet bridging / strange packet corruption

Re: ethernet bridging / strange packet corruption

Re: ethernet bridging / strange packet corruption

Re: ethernet bridging / strange packet corruption

Re: ethernet bridging / strange packet corruption

Re: ethernet bridging / strange packet corruption

Re: Flashing eMMC on Compute Module...rpiboot gives nothing

Re: Ubuntu 14.04 (trusty) Raspberry Pi 2 image

Re: Ubuntu 14.04 (trusty) Raspberry Pi 2 image

Re: Ubuntu 14.04 (trusty) Raspberry Pi 2 image

Re: Ubuntu 14.04 (trusty) Raspberry Pi 2 image

Re: Ubuntu 14.04 (trusty) Raspberry Pi 2 image