Since Friday, BengalOS gained support for DisplayPort on SHIFT 6MQ.

This is a new addition to the recently added USB-C role switching (OTG).

Kernel support included from the `sdm845-next-20260820-15` tag, if you want to integrate it into your project!

#SHIFT #BengalOS #SHIFT6mq #DisplayPort #sdm845 #LinuxMobile #MobileLinux

Goodbye #LinuxDays2026, warm welcome at #OpenSourceSummit in #Prague. The #openSUSE Tumbleweed installation on #MILKV Pioneer 64-core #RISCV system, presented at #LinuxDays running our #QtRvSim educational simulator (Sv32/Sv39 visual emulation on Sv39), #pysmiCoder, and CTU CAN FD-based #CANBUS controller on CycloneIV PCIe board, will be available to play with on the @opensuse booth at #OpenSourceSummit from this Wednesday till Friday. The graphical #Rust application for control and visualization of a CAN FD-connected motor controller running on the MCU in #Zephyr has been implemented @remus32. I have bootstrapped (through Canadian cross from my #Debian) git version #FreePascal and built the #Lazarus libraries, which allowed me to build our Open Chromatography Station CHROMuLAN software. Even though this child #CHROMuLAN chronically starves for more love and maintenance time, it is being actively used with our HPLC and amino acid analyzer systems for decades. It is one of the few open systems available for chromatogram evaluation, even for other instruments in general; see references at Scholar for its use in application and primary research. But compared to these achievements, the #SuperTUX and #SuperTuxKart have been the most appreciated applications on this 64-core RISCV box by #LinuxDays children visitors.

Great stuff you get to see at #linuxdays2026 : #opensuse tumbleweed running on a #riscv board, the MilkV thanks to the work by @ppisa and colleagues 🥳

Come to #linuxdays tomorrow to see it live!

#QtRvSim 0.9.9 release adds Sv32 and Sv39 #RISCV standard MMU, S and U modes, PC position in source code visualization, and direct fetch and submission into the evaluation site. Online version at #CVUT central computer architecture education site https://comparch.edu.cvut.cz/ is updated as well as GitHub release page, where you can find a little more detailed changelog with contributors’ attributions. The related lectures sources are available at https://github.com/cvut/comparch-slides;

We send our big thanks to SUSE.com, who conceived and support the openSUSE Build Service (https://build.opensuse.org/), which allowed us to start QtMips and then QtRvSim packaging for a broad range of Linux distributions in the past, and motivates us to prepare it for openSUSE as well.

We will present advances in QtRvSim and the RVapo VHDL equivalent with a new GDB interface at https://www.linuxdays.cz/2026/ on Sunday, October 4. We will have the Open Technologies Research Education and Exchange Services booth there with our whole hardware ZOO and #QtRvSim running on Milk-V Pioneer 64-core RISC-V system with an openSUSE Tumbleweed installation. Thanks to #SUSE’s sponsorship, we will continue demos at the #openSUSE prepared booth, focused on cooperation with #CVUT, #CVUTFEL, and other academic institutions at Open Source Summit Europe in Prague from October 7 to October 9.

Introduced QtRvSim MMU, which allows running MIT PDOS XV6 (slowly and slightly modified for QtRvSim).

We will present other RISC-V hardware, as well as our #CANbus projects, including the CTU CAN FD IP core on PCIe card in #MilkV Pioneer, industrial control projects, and educational projects, e.g., pysimCoder.

Thanks to @remus32 for helping with #MilkV Pioneer bringup as to Han Gao RevySR. We have used his RevyOS to bringup system and boot it with his #Debian derivative kernel and EFI boot. We have switched to Linux his 7.3.0-rc4+ #Linux kernel built on openSUSE. The mainline #Grub EFI has been used as the second stage loader.

Hello @ChuckMcManis we have worked on #RTEMS framebuffer support in past, I lead work on #Microwindows port for RTEMS year ago, see microwindows.bset in official RSB. But it need more work to be updated. We have worked on Cyrrus FBsupport to test it under QEMU on x86 and more. This year I have led GSoC to port Microwindows to #NuttX, the work progress tracking can be seen https://github.com/apache/nuttx/issues/18566. I am attaching example from QEMU x86_64 flat memory space NuttX build. My colleague tested result even on STM32 with LCD connected by parallel memory mapped bus. I am not sure how well this will fit into your graphic stack. But we have run even serial production of medical infusion pump based on RTEMS and this graphic stack with own simple GUT/widget set SuTk. I am quite busy by other things these days, but in longer term I have interest to join forces on RTEMS graphics support or look for students for our more GSoCs or final thesis works at university.

I will attend the ESA Industry Space Days next week. I would be happy to meet and discuss our #RTEMS, #Linux, #NuttX, communication, #canbus, #SpaceWire, IP cores related work. Some examples

• for RTEMS CAN/CAN FD, there is near to merge Ti AM335x suport, result of this year Ning Zhang #GSoC
• motion control at #PiKRON (used by ESA for #satellite tracking and for LISA ground experiment), and at #CVUT (SaMoCon)
• for rapid prototyping CTU Simulink™ support for Linux and NuttX which can be easily extended for RTEMS and open-source pysimCoder which already target Linux, NuttX and RTEMS (ongoing preparation for Ahmad Hisham’ s thesis)
• the CTU CAN FD IP core is already available on mass produced MCU COTS silicon in addition to multiple FPGA applications including space ones
• RTEMS LwIP port started as our experiment during Premysl Houdek’s thesis and #GSoC is getting traction and was used at docking computer during Artems II mission (by the way, there will be presentation about the mission at #cvutfel by Milan Halousek at Sptember 24) and is used on TMS570 and ESP32C3 RTEMS ports
• our SumInt ADC conversion principle used in our HPLC instruments and tested at #nanoXplore for ESA which can have interesting potential for space silicon
• our NuttX support for IEEE 802.3cg 10BASE−T1S Open Alliance SPI MACPHY can be ported to RTEMS etc.
• I offer support for CAN/CAN FD emulation on #QEMU, #FlexCAN added this year
• lot of our other real-time work for Linux, NuttX and RTEMS, se my university Open Technologies Research Education and Exchange Services topics and theses lists
• we invest lot into #RISCV education at CTU, see our https://comparch.edu.cvut.cz/ pages with #QtRvSim emulator and our HDL education and sometimes even practically used cores and experiments

I arrive to Noordwijk around the lunch time on Tuesday September 15 and stay there to about 2 PM on Friday so there is lot of options to meet.

Mobile #Linux Hackday #8 hit record attendance in #SUSE's newly renovated Prague office! Several attendees hacking on #Mesa #GPU drivers, #BengalOS, Snapdragon 845 #kernel debugging, and #AI developer tools. The community keeps growing! 📱🐧 https://news.opensuse.org/2026/09/01/mobile-hackday-8/

Some people still believe in Moore's law. Well, things changed around 2010. https://www.youtube.com/watch?v=Pn4ZwlEh5nw

Ukraine received 10 ALTO NG light trainer aircraft for future combat pilot training. Five were provided by the Czech government, while five more were purchased with support from the Czech charity fund Dárek pro Putina. #Ukraine

The processed recordings of the three #CompArch B0B35LSP lectures I presented have been updated/reprocessed. I have spent considerable time fixing the blur, which makes the #QtRvSim #RISCV simulation sequences especially less readable. There have been more problems, mainly in the professional #BlackMagic setup at our university. Even after some reports and related fixes, there seems to be at least one frame missing in every 4k recordings, so there is no match between the number of frames, recording time, and frame rate. This results in the detection of the file as a variable frame rate one by #kdenlive and even other tools (i.e. commercial on colleague’s MacOS). This results in transcoding to proper 25fps during import. There is an option to choose that as lossless, but when I already have 50GB from the original 1.5-hour recording, then I cannot imagine what the size would be. So there is degradation, and when the project is FullHD, and I want to expand one quarter of the original https://youtu.be/pqW9_1t5Y5U 4k video, which is holding my presentation monitor and projector output to the whole FullHD screen, then it leads to practical half resolution. Even though I have spent time with full 4k processing, then final scale down to FullHD loads to interpolation and blur, even for the nearest pixel select option. I have solved the problem by extracting the top-right quarter of the input using ffmpeg with high-quality encoding, then I have imported this new screen file, added it to the timeline after the already edited lecture, and then done precise manual replacement/switch of both audio and video chains in the “*.kdenlive” file to place this extracted screen to all already edited chunks in the correct playlist. I have then even replaced the converted 25 fps sources with the problematic original ones and re-run rendering, with confirmation of a VBR warning about possible skew between audio and video. Quite a lot of time was spent on incorrect output from a professional and probably very expensive setup. I have never had such problems with #opensource #obsstudio when using my old Intel HD graphics-based laptop (I have objected to signing the NVIDIA NDA at the university and do not want to use their HW if at least a little possible) to capture and process FullHD 50 Hz and sometimes even 4k 25 Hz videos from abandoned 4k #DataVideo camera with uncompressed stream over HDMI and USB CamLink. I would expect that if the expensive HW producers provide full resources on how to update, correct, enhance, and replace SW and firmware in their products to the community, then it would save a lot of time for everybody.

Xiaomi Mi 8 (codename dipper) landed on my desk today.

Inserted SIM card, logged in into Xiaomi service a voila, unlocked.

I'll try to introduce some basic support, but without active community don't expect comparable experience to OnePlus 6 or Xiaomi Poco F1.

Seriously, we're happy that people use our work, but we need active devs.

We working on phones which aren't "cool" today. Phones based on Snapdragon 845 are phones which people actually daily-drive.

#sdm845 #MobileLinux #LinuxMobile

The #SpaceWire ( #satellite oriented wired communication standard) #OpenSource #GNU/ #Linux software support (C library, CLI, Python) for Remote Memory Access Protocol ( #RMAP ) initiator and target realized by STAR-Dundee RMAP IP Cores integrated into custom variant of TEDIA PCD-8006 PCIe expansion card developed by Kyaw Htet Lin at #CVUT #CVUTFEL and his related thesis have been published.

The project page https://gitlab.fel.cvut.cz/otrees/spacewire/spw-pcie-tedia, Wiki, Thesis PDF, Presentation PDF and complete record with review reports https://dspace.cvut.cz/entities/publication/490357d1-8609-414e-91ad-87d4cc0e9b7e.

Longer term plan is to continue with support of more SpW IP cores, #RTEMS and #Zynq and #BeagleV/ #BeagleVFire with potential to be used even for space qualified #PolarFire #SoC.

@Oskar456 Yes, I am fully aware that there is no additional security when I use Eduroam network and in general that each connection, may it be except my local wired home network behind NAT and with limited external ports open, has to be considered as the wild environment. Eduroam federation and authentication have big advantage, that I can connect at lot of research organizations worldwide without negotiation with local admins. So yes, the referred Czech TV warning and claims are useless. The devices should be updated and do not allow any brute force and malformed packets and connections to go in. For public chargers, the USB-A to C cable with power wires only could be solution for slow but relatively safe charging.

@Oskar456 I have never let run Eduroam installer on my computer. Sane entities, like #CVUT FEL and CESNET use well established Root CA certificates which are available on every GNU/Linux system which could be selected directly in teh GUI tools or WPA supplicant configuration. Problem si that who is not careful usually does not setup required CA pro authentication and other problem is if

 altsubject_match="DNS:radius.fel.cvut.cz"

is not set. As I understand it, if the radius DNS match and specific CA are not set, then the weak MS-CHAPV2 challenge-response cryptography with MitM is susceptible to brute force attacks. We probably can again tanks to Microsoft for this standard and problem to switch to something better because it would not work with their systems at Eduroam scale.

linux-image-7.1-sdm845 is ready for testing!

This release brings major improvements across the sdm845 device family:

✅ Rear camera on SHIFT 6MQ
✅ Both front cameras on Pixel 3
✅ NFC on SHIFT 6MQ and Pixel 3
✅ Better overall stability
✅ Improved OnePlus 6 display driver
✅ Rewritten OnePlus 6T driver
✅ Pixel 3 touchscreen support moved from downstream to upstream patches
✅ Many patches improved and upstreamed

And that's only part of the changelog.

#LinuxMobile #MobileLinux #sdm845 #OnePlus6

@chiptronCZ In your ESP32 list and comparison, TWAI/CAN/CAN FD #canbus support is missing completely. The most of the #Espressif #ESP32 family members have standard CAN support. ESP32C5 is the first with CAN FD (2x), ESP32S31 CAN FD (2x) and ESP32H4 CAN FD (1x).

Thanks to the great work done by @wrenix, @valpackett, and many others (certainly not me), I've finally gotten calls working on my Fairphone 5.

While there's some kinks left to be ironed out, call quality is excellent out of the box (certainly better than my PinePhone (Pro)).

I've pushed a `linux-sc7280-calls` kernel to my arch repo for anyone who wants to try it. The following PR discussion should help set it all up:

https://github.com/sc7280-mainline/linux/pull/27

#linuxmobile #fairphone #archlinux #arch

@donhawkins I am happy that I did not like LinkedIn's closed nature from the start, so I have never registered, even though the elites of our university suggested that all teachers in the Open Informatics program should join their LinkedIn group. I am a long-time registered user on OpenHub and even GitHub, which has been bought by Microsoft as well. And I login (almost one year already) from my GNU/Linux system to educational sites which we build for students and local git services through Microsoft Entra ID due to decisions of our Microsoft-favoring university's central IT department. So each day there is a risk that somebody from US intelligence decides to impersonate me in study systems or grab all mine, our researchers and students data, or deny us access to our systems. Big shame to every IT person who pushes this direction at our university and elsewhere. It seems that this is not enough and Chat Control law will push us all to Peter Thiel's Palantir and Persona forcibly.

I wanted the blue checkmark on LinkedIn. The one that says “this person is real.” In a sea of fake recruiters, bot accounts, and AI-generated headshots, it seemed like a smart thing to do.

So I tapped “verify.” I scanned my passport. I took a selfie. Three minutes later — done. Badge acquired. I felt a tiny dopamine hit of legitimacy.

Then I did what apparently nobody does. I went and read the privacy policy and terms of service.

Not LinkedIn’s. The other company’s.

https://thelocalstack.eu/posts/linkedin-identity-verification-privacy/

UK, welcome back to Erasmus+ 🇪🇺🇬🇧

The EU and the United Kingdom sealed the deal to bring the UK back into Erasmus+ from 2027.

We are rebuilding opportunities for the next generation to study, work and grow together, giving people the chance to connect across borders once again.

It will also help create and deepen partnerships between our academic institutions.

Together again through learning, exchange and connection 🤝

More: https://link.europa.eu/MHc9JP