Alexander Rössler has put together a page showing how to create a Machinekit image for the BeagleBone based on Debian Packages.
The current packages still require absolute paths to the PRU code in your configuration files (sadly, a different path than is needed for the previous run-in-place builds). I hope to have enough time to address that soon by embedding the PRU code directly into the HAL driver. This will also prevent any possible confusion caused by the HAL and PRU code versions not matching (which I can confirm is a nasty issue to track down if you’re not looking for it!).
Package installs also make it much easier to update both Machinekit and the kernel, both of which are “difficult” to update on all of my currently available images.
If you feel up to the task, I strongly suggest switching to a package based install. It should make it much easier to track ongoing progress with both Machinekit and newer kernels.
Showing posts with label BeagleBone. Show all posts
Showing posts with label BeagleBone. Show all posts
Thursday, September 11, 2014
Wednesday, August 6, 2014
New Toy!
OK, “toy” might not be the right word (it’s really a powerful tool), but it’s just so much fun to play with! 🙂
UPS delivered a new controller box for my Probotix Fireball Comet CNC router yesterday. Unlike the PC version that came with the Comet, this one has a BeagleBone inside and will run stand-alone!
The PC configuration for the Comet is a great example of things you can do with HAL. The Y axis gantry is is controlled by a single step/dir generator using some HAL logic to split the drive pulses into signals for two separate stepper motors with two independent home switches. But this approach only works with software stepgen, so something else is needed for “outboard” stepgen that is not directly part of HAL (ie: hm2 drivers and the BeagleBone PRU driver).
I have written a gantry HAL component that mimics the behavior of the Probotix HAL file but works with outboard step generation and up to a seven joint gantry axis. Testing this with a three-joint gantry axis (my linear-delta 3D printer, temporarily re-purposed as a test-bed) was a lot of fun, but I really look forward to getting this working with real hardware and making some chips. Look for some example configurations using the gantry HAL component in the near future!
NOTE: In addition to working with standard gantry routers (Fireball, Shapeoko, etc), the gantry component will also work well for other multi-motor axis (like the two-motor Z axis found on many Mendel-style 3D printers) as long as each joint (motor) has it’s own home switch.
UPS delivered a new controller box for my Probotix Fireball Comet CNC router yesterday. Unlike the PC version that came with the Comet, this one has a BeagleBone inside and will run stand-alone!
The PC configuration for the Comet is a great example of things you can do with HAL. The Y axis gantry is is controlled by a single step/dir generator using some HAL logic to split the drive pulses into signals for two separate stepper motors with two independent home switches. But this approach only works with software stepgen, so something else is needed for “outboard” stepgen that is not directly part of HAL (ie: hm2 drivers and the BeagleBone PRU driver).
I have written a gantry HAL component that mimics the behavior of the Probotix HAL file but works with outboard step generation and up to a seven joint gantry axis. Testing this with a three-joint gantry axis (my linear-delta 3D printer, temporarily re-purposed as a test-bed) was a lot of fun, but I really look forward to getting this working with real hardware and making some chips. Look for some example configurations using the gantry HAL component in the near future!
NOTE: In addition to working with standard gantry routers (Fireball, Shapeoko, etc), the gantry component will also work well for other multi-motor axis (like the two-motor Z axis found on many Mendel-style 3D printers) as long as each joint (motor) has it’s own home switch.
Wednesday, June 18, 2014
Packages (not tied up with strings)
If you don’t monitor the Machinekit group (you should), over the weekend John Morris announced the availability of binary packages for Machinekit and it’s required dependencies (at least the ones that are not already in the Debian archives).
If you’re trying to build a minimal system (particularly helpful on the BealgeBone with it’s limited eMMC rootfs), this is welcome news indeed!
I will be updating my build scripts and the resulting images to use the pre-built packages soon, meaning you’ll be able to simply “apt-get update” and “apt-get upgrade” to get the latest official Machinekit release. Until then, feel free to add the pre-built packages to your existing install, the run-in-place development version and the installed packages should work side-by-side and should not interfere with each other. Instructions can be found at the package repository website.
If you run into any problems, let us know on the Machinekit list.
If you’re trying to build a minimal system (particularly helpful on the BealgeBone with it’s limited eMMC rootfs), this is welcome news indeed!
I will be updating my build scripts and the resulting images to use the pre-built packages soon, meaning you’ll be able to simply “apt-get update” and “apt-get upgrade” to get the latest official Machinekit release. Until then, feel free to add the pre-built packages to your existing install, the run-in-place development version and the installed packages should work side-by-side and should not interfere with each other. Instructions can be found at the package repository website.
If you run into any problems, let us know on the Machinekit list.
Thursday, June 5, 2014
CRAMPS Kits Available!
I am pleased to announce that I have a limited number of CRAMPS V2.1 DIY Kits available for sale. This includes a PCB and all the parts required to build a functioning board. Assembly details are on the RepRap wiki, and the design files are on github. Please review the known issues, and be sure that you are comfortable soldering 0603 surface-mount chip components prior to ordering a kit.
I apologize in advance for the somewhat clunky shopping cart, but it is the best open-source solution I could find that dynamically calculates shipping rates based on destination (which is required, since the rates for international package shipments vary dramatically). Yes, you will have to create yet another web account you will likely never use again, but it keeps me from having to charge everyone some obscene amount like $25 for shipping.
I currently have about 20 kits available, and I do not expect to make any more (it is a LOT of work!), so act quickly if you want one. Also, I am doing this to help with the adoption of the CRAMPS design, so there is a one-kit-per-person limit.
Wednesday, May 21, 2014
Updated Machinekit Images
After receiving a BeagleBone RevC with 4GB on-board eMMC, I noticed the latest images have grown to the point they are essentially unusable on a 4GB device.
I have updated the Machinekit images to exclude the previously required documentation build dependencies, which saves over 1GB of space on the images! There is still a lot more room to be recovered once ARM binary packages become available, but for now the space savings means you can comfortably run Machinekit on a 4GB uSD card or the built-in 4GB eMMC on a RevC ‘Bone.
I have updated the Machinekit images to exclude the previously required documentation build dependencies, which saves over 1GB of space on the images! There is still a lot more room to be recovered once ARM binary packages become available, but for now the space savings means you can comfortably run Machinekit on a 4GB uSD card or the built-in 4GB eMMC on a RevC ‘Bone.
Thursday, May 15, 2014
CRAMPS Kits?
The CRAMPS V2.1 circuit boards are fabricated and heading my way. I am planning to sell the bare boards for a nominal cost (under $5), but I may get motivated and sell some DIY kits that include all the parts needed to assemble your board. The kit would probably be around $50-60, which is pretty much my cost for the parts in low volume, and I’d only be able to make about 10-20 kits total (before I run out of circuit boards).
So a quick informal poll: If you are interested in buying a bare PCB or a DIY kit for the new CRAMPS board, send me a direct email and if I get enough interest kits I’ll try to put some together. Sending me an email places you under no obligation to actually buy anything, nor does it guarantee you’ll actually get a kit should I decide to make some.
I am also working on getting someone to build and sell assembled boards, but that is a process that will likely take another couple months. The kits would probably be available in 2-4 weeks, depending on how long it takes the PCBs to arrive and if I find any problems with the first prototypes.
So a quick informal poll: If you are interested in buying a bare PCB or a DIY kit for the new CRAMPS board, send me a direct email and if I get enough interest kits I’ll try to put some together. Sending me an email places you under no obligation to actually buy anything, nor does it guarantee you’ll actually get a kit should I decide to make some.
I am also working on getting someone to build and sell assembled boards, but that is a process that will likely take another couple months. The kits would probably be available in 2-4 weeks, depending on how long it takes the PCBs to arrive and if I find any problems with the first prototypes.
Friday, May 9, 2014
Machinekit BeagleBone Debian Images Released
After what has been far too long, I am please to announce a new official release of the Machinekit image for the BeagleBone. The release of the RevC BeagleBone Black with it’s 4GB eMMC and switch to Debian makes the BeagleBone an even better choice for machine control!
There are lots of significant changes with this release, so I recommend everyone using Machinekit updates when practical. The biggest change is the switch from my “enhanced” version of Robert Nelson’s minimal Debian build to generating images based on the official BeagleBone Debian. That means all the BeagleBone specific features like node.js, USB networking, etc. should work just like they would on a factory BeagleBone. Of course there are a LOT of changes included with this, like switching from xfce to lxde, changing from sysvinit to systemd, and many others. On the plus side, the Machinekit images are now very similar to the shipping BeagleBone release, so it should be easier to find documentation and ask for help.
The other really big change is a switch from LinuxCNC to the Machinekit project. Machinekit enables LinuxCNC’s real-time core to operate on a wide variety of platforms, making it possible to run on the BeagleBone. I had been hoping the required real-time changes would become part of mainline LinuxCNC (and they still might), but they have been passed over for the upcoming 2.6 release.
If you are lucky enough to have a BeagleBone RevC, there is an (untested) eMMC flasher image available as well. If you would like a tested and verified working eMMC flasher image, feel free to FedEx me your RevC ‘Bone! Otherwise, you’ll have to wait until one of the two I’ve ordered actually arrives for me to test. 😉
Thursday, May 8, 2014
CRAMPS V2.1
I finally got around to finishing the updated version of the CRAMPS board. This version focused on simplifying the design as much as possible, and I really like the end result. It feels much closer to the original RAMPS for the Arduino Mega in spirit than the previous 1.0 version and the RAMPS-FD.
Changes in this version include:
- No fine-pitch surface mount parts, so the board is much easier to hand assemble
- No servo outputs
- SPI expansion header changed to allow use as extra GPIO if needed
- Expansion headers modified to support an additional 3-axis add-on board (the CRAMP3)
While the design is fully open, not everyone wants to build their own boards from scratch. I am working on getting someone to build and sell completed boards, so stay tuned for updates!
UPDATE: 2014.05.12 – The first batch of 20 boards has shipped, now the 0-30 day wait for a “registered air parcel” begins.
Tuesday, April 22, 2014
Machinekit BeagleBone Debian Images
If you’ve been following developments around the BeagleBone, you know the new RevC boards with 4Gig eMMC are coming soon, and these boards will ship with Debian instead of Angstrom. I have been working to get my Machinekit binary images tracking the new BeagleBone Debian vs. the generic minimal Debian install used previously. This means you can run Machinekit without loosing any of the BeagleBone specific add-ons like node.js, and switching between the Machinekit image and the official BeagleBone software should be fairly painless.
I am pleased to announce the first public Machinekit test image based on the new BeagleBone Debian release. Download the image file, then check the md5sum. This is a raw image file, so just uncompress it and dd it straight onto a 4 GB or larger uSD card, pop it into your BeagleBone, and boot. If you need help writing the uSD card, follow the BeagleBone instructions, just use the image file above instead of the official BeagleBone files.
If you encounter any problems, please report them via the Machinekit issue tracker on github, or on the Machinekit Mailing list. While this is the first public release, there have been several test images for limited release (check the Machinekit list to keep updated!), and it looks like things are pretty solid. As soon as there is a final BeagleBone Debian build, I’ll make a matching Machinekit version and post it on the Machinekit page. There will even be an eMMC flasher version for those of you lucky enough to get a RevC board. I’ve got mine on order, and am hoping that I’m near the front of the line!
Tuesday, March 25, 2014
CRAMPS Update
I have been building up the second CRAMPS board to verify the low-cost PCBs from off-shore really work as expected. Since I’m doing another build anyway, I have been documenting the process with pictures and have created a CRAMPS 1.0 page with assembly details, ECOs, and other board details on the RepRap Wiki. Follow along there if you’re interested in tracking progress or are building your own board. If you are interested in building up a board and can hand-solder the six fine-pitch SMT parts, send me an email and I’ll try to get you a bare PCB.
Meanwhile, work is progressing on the next version of the CRAMPS board thanks to Murry Lindeblom. Murry has been doing some great with KiCAD getting the schematic whipped into shape while I’ve been busy with other projects. The main changes are to simplify the design, reducing parts count and making the board cheaper to produce. See the list of ECOs on the RepRap Wiki for hints about the next version.
Monday, March 24, 2014
tBone Cape
The tBone 3D printer / CNC cape has been announced for the BeagleBone. It looks like an interesting board with some unique features. I like the use of more sophisticated stepper driver chips from Trinamic, and the support for higher-current stepper motors. This board seems to have an Arduino on-board to handle real-time, so it will likely take a bit of coding to get it working with MachineKit which is expecting to control real-time motion with the PRU.
For more details, see the tBone website and their Indiegogo campaign.
Sunday, March 2, 2014
CRAMPS Board Working
I have the first CRAMPS board assembled and mostly checked out. There are a few minor glitches (see the ToDo file on github), but so far no serious problems or anything that makes the current PCB design useless. I have an example LinuxCNC configuration available if you want to try things out on your own. The configuration requires you install my new Universal BeagleBone I/O Overlay, which can be used to switch between most “interesting” hardware options for the various I/O pins without having to load multiple overlays or possibly merge several incompatible overlay files by hand.
If you’re a fan of scary movies, I even made a video:
If you’re a fan of scary movies, I even made a video:
Thursday, February 20, 2014
Linear Delta Printing with LinuxCNC
Reviewing my posts, it looks like I somehow missed sharing Brandon Heller’s incredibly awesome video of his Kossel linear-delta printing parts for a GUS Simpson (a grounded delta). I printed up a few sets of the GUS Parts on my MendelMax and will vouch for how hard they are to print. I’ve gotten my Kossel based printer to move via LinuxCNC, but it’s no where near as far along as Branon’s. Great work!
Monday, February 10, 2014
CRAMPS Board
I have finished the design for the first version of the CRAMPS 3D print controller board for the BeagleBone. This “new” design is a rework of the RAMPS-FD board for the Arduino Due, which in turn is a rework of the RAMPS board for the Arduino Mega (isn’t open-source great!). The PCB files are queued at OSH Park meaning I will hopefully have boards around the end of February.
Features & Details:
- 12-24V operation
- 6 Pololu drivers
- 15A driver for heated bed, with dedicated power
- 3 high-current drivers for extruder heaters
- 2 medium-current drivers for fans/lights, always 12V
- 4 Thermistor inputs
- 6 Limit switches (5V tolerant)
- 4 Servo outputs
- SPI and I2C expansion headers (5V tolerant)
- Flexible power options for the BeagleBone (use ‘Bone 5V jack, 5V screw terminals, or add a regulator to generate 5V from 12-24V system power)
- Hardware ESTOP chain
- Stack-through design allows for expansion
Currently the BOM pretty much doesn’t exist, and please don’t judge me by the lack of assembly drawings in the PCB files. KiCAD apparently doesn’t have assembly layers or allow creating arbitrary mechanical layers so I can’t make assembly drawings the way I’m used to for the last 15 years. I’ll get it all sorted out, and add a full BOM once I’ve sourced and ordered all the bits and pieces needed to build the board.
Saturday, February 1, 2014
New -bone39 Kernel Available
I’ve been building a lot of kernels lately, testing various options recommended by the Xenomai folks to improve latency. Nothing has been a “magic silver bullet”, but the overall combination of options has dramatically reduced typical latency, even though worst-case values haven’t dropped much.
Also, some folks testing with recent kernels noticed the CPU wasn’t running at full speed. I have re-enabled the frequency scaling support in the kernel (the Xenomai folks recommend disabling this), which has restored the expected 1 GHz operation. Tests so far show no impact on latency figures, since the CPU frequency isn’t actually changing (only the performance scheduler is compiled into the kernel, which keeps the CPU running at the highest possible speed and doesn’t try to throttle the CPU to save power). AFAIK, the latency hit is from the CPU shutting down while the PLLs relock to a new frequency, so if the CPU never changes it’s operating frequency there shouldn’t be any issues.
You can download the new kernels here. If you try it out, please let me know how it works for you. I’ll probably try to make new images with just the kernel update soon, but I’m in the middle of a bunch of other coding and updates, so pester me if you really want a new image and it hasn’t shown up on the MachineKit page.
In other news, work is in progress to develop interfaces geared towards the BeagleBone (lower resolution screens with less 3D intensive “eye candy”), I’ve started work on the CRAMPS 3D Printer interface board, and I hope to have encoder support (both hardware and PRU software based) working soon. Rigid tapping could be coming soon to a ‘Bone near you! 🙂
I am making good progress on the CRAMPS and encoder tasks. If you would like to assist with interface design for the ‘Bone, please post on the LinuxCNC developer list. I don’t do well designing GUI interfaces, and am looking for as much help as I can get to provide usable and attractive interfaces for the BeagleBone, for both CNC machining and for 3D printing. Thanks!
Also, some folks testing with recent kernels noticed the CPU wasn’t running at full speed. I have re-enabled the frequency scaling support in the kernel (the Xenomai folks recommend disabling this), which has restored the expected 1 GHz operation. Tests so far show no impact on latency figures, since the CPU frequency isn’t actually changing (only the performance scheduler is compiled into the kernel, which keeps the CPU running at the highest possible speed and doesn’t try to throttle the CPU to save power). AFAIK, the latency hit is from the CPU shutting down while the PLLs relock to a new frequency, so if the CPU never changes it’s operating frequency there shouldn’t be any issues.
You can download the new kernels here. If you try it out, please let me know how it works for you. I’ll probably try to make new images with just the kernel update soon, but I’m in the middle of a bunch of other coding and updates, so pester me if you really want a new image and it hasn’t shown up on the MachineKit page.
In other news, work is in progress to develop interfaces geared towards the BeagleBone (lower resolution screens with less 3D intensive “eye candy”), I’ve started work on the CRAMPS 3D Printer interface board, and I hope to have encoder support (both hardware and PRU software based) working soon. Rigid tapping could be coming soon to a ‘Bone near you! 🙂
I am making good progress on the CRAMPS and encoder tasks. If you would like to assist with interface design for the ‘Bone, please post on the LinuxCNC developer list. I don’t do well designing GUI interfaces, and am looking for as much help as I can get to provide usable and attractive interfaces for the BeagleBone, for both CNC machining and for 3D printing. Thanks!
Friday, January 24, 2014
Replicape Available
The Replicape, winner of the BeagleBone cape design contest, is now available for purchase from Thing-Printer. Congrats to Elias for his cape contest win, and especially for having the follow-through to get the product designed and built!
The Replicape looks like a good board to use for 3D printing, but won’t immediately work with the LinuxCNC/MachineKit images currently available. SPI and I2C setup is required, which means some coding is needed to get everything working properly. I have a board ordered, and I am currently reworking the PRU code (to make pin numbering hopefully understandable). I hope to add Replicape support to the next release of my images, but if you’re willing to try and tackle any of the SPI or I2C configuration tasks, please let me know. The setup routines can probably be done as user-mode HAL modules written in Python, so it shouldn’t be too hard (unless you’re like me, and grok assembly language better than things like Python and Ruby!).
Don’t forget there is an ever-growing list of capes to do CNC and 3D printing with the BeagleBone. Choice is a good thing, and I’m excited the Replicape is finally shipping, adding another great choice to the list of available boards. I can’t wait until mine arrives!
The Replicape looks like a good board to use for 3D printing, but won’t immediately work with the LinuxCNC/MachineKit images currently available. SPI and I2C setup is required, which means some coding is needed to get everything working properly. I have a board ordered, and I am currently reworking the PRU code (to make pin numbering hopefully understandable). I hope to add Replicape support to the next release of my images, but if you’re willing to try and tackle any of the SPI or I2C configuration tasks, please let me know. The setup routines can probably be done as user-mode HAL modules written in Python, so it shouldn’t be too hard (unless you’re like me, and grok assembly language better than things like Python and Ruby!).
Don’t forget there is an ever-growing list of capes to do CNC and 3D printing with the BeagleBone. Choice is a good thing, and I’m excited the Replicape is finally shipping, adding another great choice to the list of available boards. I can’t wait until mine arrives!
Sunday, December 29, 2013
Updated MachineKit Image Available
A new MachineKit image is now available. While there are no dramatic changes with this release, pretty much everything got updated. There’s a new kernel (-bone33), the latest Debian (7.3), and continued updates to LinuxCNC. One somewhat major change is switching from the xdm display manager to lightdm, which plays nice with consolekit so the shutdown/reboot buttons work, user-mode USB mounting works, etc.
There’s no dramatic reason to update if your current system is working well. This is primarily a baseline build I wanted to make available before I head off and start seriously working with the joints-axis code on the BeagleBone for controlling the non-Cartesian 3D printers I’m building (a Mini-Kossel–Pro mashup, and Nicholas Stewart’s Wally and GUS Simpson).
There’s no dramatic reason to update if your current system is working well. This is primarily a baseline build I wanted to make available before I head off and start seriously working with the joints-axis code on the BeagleBone for controlling the non-Cartesian 3D printers I’m building (a Mini-Kossel–Pro mashup, and Nicholas Stewart’s Wally and GUS Simpson).
Monday, December 2, 2013
New BeagleBone Cape Available
Alex Joni has finished up his bbb_parport cape and has them available for sale!
This cape provides support for up to 6 axis, and has an optional ADC for analog inputs. There is a DB25 connector that should allow direct connection of an existing LPT based CNC machine to a BeagleBone. If you’re more interested in cutting metal than printing with plastic, this is the cape for you!
Great work, Alex!
This cape provides support for up to 6 axis, and has an optional ADC for analog inputs. There is a DB25 connector that should allow direct connection of an existing LPT based CNC machine to a BeagleBone. If you’re more interested in cutting metal than printing with plastic, this is the cape for you!
Great work, Alex!
Tuesday, November 5, 2013
BeBoPr-Bridge Purchase Page
To make it even easier to purchase a BeBoPr-Bridge board set, I have added a purchase page complete with nifty PayPal “add-to-cart” buttons.
I have stock levels linked to the buttons, so if PayPal lets you send me money, I have the PCBs and connectors in-hand and ready to ship to you (saving weeks vs. ordering direct from OSH Park).
I also have another BeBoPr and BeagleBone Black coming, so I’ll be putting together my second Bridge board set soon and I promise this time I’ll take lots of pictures and put up some assembly instructions. When I built the first Bridge board I didn’t realize I would eventually be selling them! 🙂
I have stock levels linked to the buttons, so if PayPal lets you send me money, I have the PCBs and connectors in-hand and ready to ship to you (saving weeks vs. ordering direct from OSH Park).
I also have another BeBoPr and BeagleBone Black coming, so I’ll be putting together my second Bridge board set soon and I promise this time I’ll take lots of pictures and put up some assembly instructions. When I built the first Bridge board I didn’t realize I would eventually be selling them! 🙂
Saturday, October 19, 2013
Hung Task Bug in Xenomai Kernel
Thanks to Ralf Roesch, some glitches I had been writing off to bad SD cards have been shown to be a significant bug in the Xenomai patched kernel for the BeagleBone. The symptom is the kernel hangs and prints a message similar to the following:
To ‘tickle’ this bug, all you have to do is boot, get to a console prompt, and run:
There’s a thread over on the Xenomai list about this issue, and it looks like it’s a problem with the interrupt code for the am335x MMC or DMA driver (the MMC code makes heavy use of DMA). I’m not sure how long it will take to track down this bug, or if I’ll be able to recruit any help from TI or elsewhere.
NOTE: You can still use LinuxCNC on the BeagleBone, as this issue doesn’t really show up that often (I’ve done overnight prints and had systems run fine for a couple weeks straight), but it is definitely something that needs to get fixed.
UPDATE: I initially figured this was probably an issue with the TI code for the AM335x mmc controller, but it looks like it could possibly be an actual kernel bug based on the following very similar issue:
[26160.894920] INFO: task mmcqd/0:74 blocked for more than 60 seconds. [26160.901577] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. [26160.909811] mmcqd/0 D c0699e08 0 74 2 0x00000000 [26160.916607] [] (__schedule+0x5b8/0x774) from [ ] (schedule_timeout+0x1c/0x21c) [26160.925943] [ ] (schedule_timeout+0x1c/0x21c) from [ ] (wait_for_common+0x130/0x170) [26160.935718] [ ] (wait_for_common+0x130/0x170) from [ ] (mmc_wait_for_req_done+0x1c/0x74) [26160.945851] [ ] (mmc_wait_for_req_done+0x1c/0x74) from [ ] (mmc_start_req+0x50/0x158) [26160.955712] [ ] (mmc_start_req+0x50/0x158) from [ ] (mmc_blk_issue_rw_rq+0xa4/0x348) [26160.965485] [ ] (mmc_blk_issue_rw_rq+0xa4/0x348) from [ ] (mmc_blk_issue_rq+0x3fc/0x450) [26160.975624] [ ] (mmc_blk_issue_rq+0x3fc/0x450) from [ ] (mmc_queue_thread+0xa0/0x104) [26160.985481] [ ] (mmc_queue_thread+0xa0/0x104) from [ ] (kthread+0xa0/0xb0) [26160.994345] [ ] (kthread+0xa0/0xb0) from [ ] (ret_from_fork+0x18/0x38) [26161.002875] Kernel panic - not syncing: hung_task: blocked tasks [26161.009184] [ ] (unwind_backtrace+0x0/0xe0) from [ ] (panic+0x84/0x1e0) [26161.017756] [ ] (panic+0x84/0x1e0) from [ ] (watchdog+0x1d4/0x234) [26161.025866] [ ] (watchdog+0x1d4/0x234) from [ ] (kthread+0xa0/0xb0) [26161.034065] [ ] (kthread+0xa0/0xb0) from [ ] (ret_from_fork+0x18/0x38) [26161.042529] drm_kms_helper: panic occurred, switching back to text console
To ‘tickle’ this bug, all you have to do is boot, get to a console prompt, and run:
grep TestConfig /usr -r
There’s a thread over on the Xenomai list about this issue, and it looks like it’s a problem with the interrupt code for the am335x MMC or DMA driver (the MMC code makes heavy use of DMA). I’m not sure how long it will take to track down this bug, or if I’ll be able to recruit any help from TI or elsewhere.
NOTE: You can still use LinuxCNC on the BeagleBone, as this issue doesn’t really show up that often (I’ve done overnight prints and had systems run fine for a couple weeks straight), but it is definitely something that needs to get fixed.
UPDATE: I initially figured this was probably an issue with the TI code for the AM335x mmc controller, but it looks like it could possibly be an actual kernel bug based on the following very similar issue:
- MMC locking: mmc_request accesses from dw_mmc driver ok?
- [PATCH 0/7] mmc: core: cleanup and locking patches description
Subscribe to:
Posts (Atom)








