- Posts: 529
- Thank you received: 0
LinuxCNC 2.8.1 configuration doesn't run after upgrading from 2.7.15
Are all your pyvcp components in custom_gvcp.hal? That's why I had very confusing reports.
Please Log in to join the conversation.
It is possible that the HAL pins take up more space on a 64-bit OS than on a 32-bit OS.
There is certainly no reason to expect any difference in behaviour here between 2.7.15 and 2.8.1 caused by LinuxCNC changes alone.
Please Log in to join the conversation.
hello chris,chris@cnc wrote: Just an idea ...
Are all your pyvcp components in custom_gvcp.hal? That's why I had very confusing reports.
You can organize your configuration with multiple successive hal files, you just have to declare
them in the [hal] section of the ini boot file. Exemple base.hal, custom.hal, test.hal etc
At the very end you have a last hal file which is read after the GUI (axis for exemple) is loaded.
it loads the hal connections with the Pyvcp interface add-in which is designed with the xml file
Please Log in to join the conversation.
andypugh wrote: Is your 2.8.1 configuration running on the same OS and machine as the 2.7.15 configuration?
It is possible that the HAL pins take up more space on a 64-bit OS than on a 32-bit OS.
There is certainly no reason to expect any difference in behaviour here between 2.7.15 and 2.8.1 caused by LinuxCNC changes alone.
No,
2.7.15 Debian Wheezy is under i386 so in 32 bit because with RTAI
2.8.1 Debian 10 buster therefore under 64 bit with Prempt-RT
the 2 OS runs on the same hardware, each on an SSD hard drive
Please Log in to join the conversation.
HALFILE = 1-ini.hal
HALFILE = 2-io.hal
HALFILE = 3-axis.hal
HALFILE = 4-spindle.hal
POSTGUI_HALFILE = postgui_call_list.hal
in file simulator-3 the content from pyvcp_options is in custom_gvcp.hal.
ini section.
EMBED_TAB_NAME = GladeVCP
EMBED_TAB_COMMAND = halcmd loadusr -Wn gladevcp gladevcp -c gladevcp -H gvcp_call_list.hal -x {XID} gvcp-panel.ui
I would try both if one didn't work

Please Log in to join the conversation.
With a 64-bit kernel:
So, 13 pins on the 32 bit kernel consumes 1664 bytes, and with the 64-bit kernel 1760 (creating float pins)
Creating bit pins the numbers are the same. (which makes some sense as each "pin" is really a memory pointer.
Breaking it down:
32 bits is 256 bytes for the mux function, 168 bytes for the parameters and 1144 bytes for the pins.
64 bits is 336 bytes for the mux function, 176 bytes for the parameters and 1152 bytes for the pins.
It's not a big difference, but there is a difference.
Have you considered just updating the 2.7.15 installation to 2.8.1? There is no need to change the kernel and OD just to run the latest version of LinuxCNC.
Please Log in to join the conversation.
I just decided that it would be a good idea to be able to set the rtapi messaging level from HAL, so set about looking at how to do that.
And found it has already been done.
Last year.
By me.
So, to check that you really are running out of memory:
At the top of the HAL file where the problem is found add
Please Log in to join the conversation.
I did not need to go to 2.8 under wheezy
my concern is wheezy's obsolescence with respect to the hardware.
For example wheezy does not recognize an active usb3 cable of 10 meters, buster yes
debug 5 in the hal file or in the ini file ?
with debug 5 in the ini file :
linuxcnc -d -v
Attachment not found
Please Log in to join the conversation.
gaston48 wrote: debug 5 in the hal file or in the ini file ?
In the HAL file.
Please Log in to join the conversation.
andypugh wrote:
gaston48 wrote: debug 5 in the hal file or in the ini file ?
In the HAL file.
linuxcnc launched in the console with linuxcnc -d -v etc ...
Attachment not found
Please Log in to join the conversation.
