The first release was made in 30-09-2017 (https://sourceforge.net/projects/rtospharos/files/1.0.0/)
We are currently developing Pharos as a hobby. Hopefully in the future we will be able to support it with more dedication.
The short-term plan is to have fun and do something that we feel is a good product, technically speaking. Creating a good solution (at least in our eyes) is something that drives us to spend some extra hours in front of the screen programming. Right now we are porting to Cortex-M4 CPUs.
Our vision is to provide a full-proof RTOS completely secure to anyone who wants it.
Right now, we are not thinking of Github, but to host it here, on sourceforge, with SVN. We need some time to do this since we are now focused on porting to M4.
Do you have any suggestion/request/whatever?
Best regards
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Pharos Team,
Thanks for your quick reply!
We're looking for a safety critical RTOS for Cortex-M. I checked the description of this project. It seems you guys are familiar with the safety/security stardards of many fields (ARINC653, DO178, MISRA and etc.).
Our target is to build a functional safety system which includes Cortex-A and Cortex-M for safety critical machines, e.g. Robotics, Automotive and etc. The applications are based on ROS2 (Robot Operating System - acutally a middleware). (Our work are all open source and free to use - mainly BSD liscense)
It looks like this project should be a good choice. So I want to spend some time to investigate it. :)
BTW, we're not experts of FuSa (functional safety), so it's much appreciated if you experts have any
suggestions of building FuSa system.
Github is a nice place to get contribution from the developers over the world, but anyway it's up to your resource and plan. :)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes, we have some experience with qualification standards. We can tell you (if you don't already know) that qualificating a piece of software is very demanding. In terms of tools and time.
Do you have a custom board or a COTS? If a COTS we could try to port Pharos to it - as long as its not too expensive, since the money comes from our own pockets.
We were actually thinking of the next board to qualify next, and were thinking about a Cortex-A8 board (BeagleBone Black for example), but we haven't decided yet.
We understand Github is very "hot" now, we certainly considered it. We ended up choosing sourceforge, at least for now.
If we can help you, please don't hesitate to ask
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Currently, our idea is to use a Cortex-A e.g. 96 boards - Hikey960 and a Cortex-M e.g. STM32F7 to set up the basic system for robotics/automotive.
However, as everyone knows security and safety are very important, so we're looking at what we can do in these areas.
For software part, my draft idea is to adapt seL4 on Cortex-A. It's like a hypervisor and we can run multiple ROS2 instances(based on Linux or maybe unikernel) to provide FuSa. But not sure if this is a correct way.
For Cotex-M, current plan is to build on top of Nuttx. But I'm not sure if it can fit the requirements of FuSa. So that's the reason I'm considering pharos.
For the test boards, I suggest to use Armv8 not v7,as v8 has 64bit support which is the main-tream and I think anyway you will move the aarch64 (Arm 64bit) in the future. Raspberry Pi3 is cheap enough for aarch64 development and it has many many developers. BBB is good but a little bit "old". :)
Our focus is to build up the Arm ecosystem on robotics and automotive areas. So our solution will be open source and anyone/any company can build their projects on top of our work (e.g. build a robot with a few Cortex-A and Cortex-M subsystems). Furthermore, this solution (most components) should be certificated by the industrial standards (I think this matches your idea also).
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank you again for your inputs. They are appreciated.
For us it is a little difficult to speak about seL4 and Nuttx since we do not have a neutral point of view and we are very biased towards Pharos. But I think it all comes down to the certification standard that you want to follow and if the technology suffice your needs.
For example, for avionics (DO178B) there are criticality levels defined (DAL-A being the most critical and DAL-E the least critical). The higher you go tn criticality, the most difficult it is to qualify it. You need a lot of artifacts to demonstrate the certification, tests (a lot), requirements, design, criticality analysis, verification, quality assurance, etc. But that's another game.
If you want to classify your product according to one of these levels of criticality then ALL of the software that you run (including the operating system, libraries and your own code) must be qualified to same level or higher. Pharos is certainly not qualified to any level of criticality.
It was our intent to provide Pharos free from external libraries (because those would need to be qualified as well). Of course you could add any external library directly in your application.
The qualification issue is a very big one to discuss so briefly here. There are many standards and ways to do. Even the HW you choose must be planned carefully (e.g. has EDAC? has dual-step core?). If you have already a specific standard in mind maybe we could take a look and help you.
As to the boards themselves, as I told you, we are focused now on a M4, the http://www.ti.com/tool/EK-TM4C129EXL. I remember taking a look at the STM32 boards and found them interesting. We just ended choosing the TI board since we already worked on it on the previous Pharos release (http://www.ti.com/tool/LAUNCHXL2-570LC43). Maybe in the future we can take a look on the STM boards. Is it possible for you to take a look to the http://www.ti.com/tool/EK-TM4C129EXL?
Regarding the port for the Cortex-A family, we need to study a little bit. Just looking at the aarch64 and multi-core architecture seems a big step right now for us. Maybe a simpler port would be a more conscientious step, then go to armv8.
Maybe a suggestion to you, if you have interested in Pharos is to build right now a prototype of yourapplication on top Pharos and thinking about using EK-TM4C129EXL. We are about to release the port to this board. We could then try the concept and better decide what could be the steps forward. It could also allow you to demonstrate your application.
Please let me know, kind regards
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks.
Board is not a big issue. We can use IT's board or port stm32 by ourselves. (most probably move to Armv8m this year)
We're not prototyping a product. Our intention is to build the robotics/automotive ecosystem for Arm. What we understand is that certifications are important to these areas. So thinking about if we can facilitate developers' certifications by helping/prompting Pharos.
I'm still in investigation stage. Need some time to check the details of Pharos.
I will ask for your help if any questions. :)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
I'm new to this project. :)
When it started? Any short-term/long-term target?
And is there any plan to host it on Github?
Thanks.
Regards,
David Wang
Hi David,
Nice to see you ;)
The first release was made in 30-09-2017 (https://sourceforge.net/projects/rtospharos/files/1.0.0/)
We are currently developing Pharos as a hobby. Hopefully in the future we will be able to support it with more dedication.
The short-term plan is to have fun and do something that we feel is a good product, technically speaking. Creating a good solution (at least in our eyes) is something that drives us to spend some extra hours in front of the screen programming. Right now we are porting to Cortex-M4 CPUs.
Our vision is to provide a full-proof RTOS completely secure to anyone who wants it.
Right now, we are not thinking of Github, but to host it here, on sourceforge, with SVN. We need some time to do this since we are now focused on porting to M4.
Do you have any suggestion/request/whatever?
Best regards
Hi Pharos Team,
Thanks for your quick reply!
We're looking for a safety critical RTOS for Cortex-M. I checked the description of this project. It seems you guys are familiar with the safety/security stardards of many fields (ARINC653, DO178, MISRA and etc.).
Our target is to build a functional safety system which includes Cortex-A and Cortex-M for safety critical machines, e.g. Robotics, Automotive and etc. The applications are based on ROS2 (Robot Operating System - acutally a middleware). (Our work are all open source and free to use - mainly BSD liscense)
It looks like this project should be a good choice. So I want to spend some time to investigate it. :)
BTW, we're not experts of FuSa (functional safety), so it's much appreciated if you experts have any
suggestions of building FuSa system.
Github is a nice place to get contribution from the developers over the world, but anyway it's up to your resource and plan. :)
Thank you very much for your interest!
Yes, we have some experience with qualification standards. We can tell you (if you don't already know) that qualificating a piece of software is very demanding. In terms of tools and time.
Do you have a custom board or a COTS? If a COTS we could try to port Pharos to it - as long as its not too expensive, since the money comes from our own pockets.
We were actually thinking of the next board to qualify next, and were thinking about a Cortex-A8 board (BeagleBone Black for example), but we haven't decided yet.
We understand Github is very "hot" now, we certainly considered it. We ended up choosing sourceforge, at least for now.
If we can help you, please don't hesitate to ask
Currently, our idea is to use a Cortex-A e.g. 96 boards - Hikey960 and a Cortex-M e.g. STM32F7 to set up the basic system for robotics/automotive.
However, as everyone knows security and safety are very important, so we're looking at what we can do in these areas.
For software part, my draft idea is to adapt seL4 on Cortex-A. It's like a hypervisor and we can run multiple ROS2 instances(based on Linux or maybe unikernel) to provide FuSa. But not sure if this is a correct way.
For Cotex-M, current plan is to build on top of Nuttx. But I'm not sure if it can fit the requirements of FuSa. So that's the reason I'm considering pharos.
For the test boards, I suggest to use Armv8 not v7,as v8 has 64bit support which is the main-tream and I think anyway you will move the aarch64 (Arm 64bit) in the future. Raspberry Pi3 is cheap enough for aarch64 development and it has many many developers. BBB is good but a little bit "old". :)
Our focus is to build up the Arm ecosystem on robotics and automotive areas. So our solution will be open source and anyone/any company can build their projects on top of our work (e.g. build a robot with a few Cortex-A and Cortex-M subsystems). Furthermore, this solution (most components) should be certificated by the industrial standards (I think this matches your idea also).
Thank you again for your inputs. They are appreciated.
For us it is a little difficult to speak about seL4 and Nuttx since we do not have a neutral point of view and we are very biased towards Pharos. But I think it all comes down to the certification standard that you want to follow and if the technology suffice your needs.
For example, for avionics (DO178B) there are criticality levels defined (DAL-A being the most critical and DAL-E the least critical). The higher you go tn criticality, the most difficult it is to qualify it. You need a lot of artifacts to demonstrate the certification, tests (a lot), requirements, design, criticality analysis, verification, quality assurance, etc. But that's another game.
If you want to classify your product according to one of these levels of criticality then ALL of the software that you run (including the operating system, libraries and your own code) must be qualified to same level or higher. Pharos is certainly not qualified to any level of criticality.
It was our intent to provide Pharos free from external libraries (because those would need to be qualified as well). Of course you could add any external library directly in your application.
The qualification issue is a very big one to discuss so briefly here. There are many standards and ways to do. Even the HW you choose must be planned carefully (e.g. has EDAC? has dual-step core?). If you have already a specific standard in mind maybe we could take a look and help you.
As to the boards themselves, as I told you, we are focused now on a M4, the http://www.ti.com/tool/EK-TM4C129EXL. I remember taking a look at the STM32 boards and found them interesting. We just ended choosing the TI board since we already worked on it on the previous Pharos release (http://www.ti.com/tool/LAUNCHXL2-570LC43). Maybe in the future we can take a look on the STM boards. Is it possible for you to take a look to the http://www.ti.com/tool/EK-TM4C129EXL?
Regarding the port for the Cortex-A family, we need to study a little bit. Just looking at the aarch64 and multi-core architecture seems a big step right now for us. Maybe a simpler port would be a more conscientious step, then go to armv8.
Maybe a suggestion to you, if you have interested in Pharos is to build right now a prototype of yourapplication on top Pharos and thinking about using EK-TM4C129EXL. We are about to release the port to this board. We could then try the concept and better decide what could be the steps forward. It could also allow you to demonstrate your application.
Please let me know, kind regards
Thanks.
Board is not a big issue. We can use IT's board or port stm32 by ourselves. (most probably move to Armv8m this year)
We're not prototyping a product. Our intention is to build the robotics/automotive ecosystem for Arm. What we understand is that certifications are important to these areas. So thinking about if we can facilitate developers' certifications by helping/prompting Pharos.
I'm still in investigation stage. Need some time to check the details of Pharos.
I will ask for your help if any questions. :)