Timeline for Two FTDI devices
Current License: CC BY-SA 4.0
19 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 26, 2018 at 22:16 | comment | added | Chris Stratton | @BenVoigt - if you had every actually tried the USB task you'd realize how irrelevant what you are arguing really is. But because you've never done so, you just continue spouting falsely imagined difficulties. Seriously, give it an actual try - document what you need to do by studying the PC implementation, then get an example that does similar types of operations for the embedded platform, and modify it to do what you documented as required. You'll be surprised how straightforward it actually is. | |
| Jun 26, 2018 at 22:15 | comment | added | Ben Voigt | @ChrisStratton I may not have done "porting" but I have written peripheral control code for both MCUs and desktop OSes and it's nothing alike. Number crunching stuff can be used between embedded and desktop just by passing the code to the cross-compiler. Hardware access is extremely specific to the environment. Sure, there will be reusable pieces for processing the buffers after you get them from the low-level layer -- as long as the low-level layer on both platforms uses the same paradigm. If one is event-driven and the other uses blocking calls, good luck. | |
| Jun 26, 2018 at 22:10 | comment | added | Chris Stratton | @BenVoigt - userspace is only relevant in demonstrating that interrupts and kernel paradigms are not required for such an API. If you had ever actually tried such a porting exercise - as I have - you'd know it's not nearly as complicated as you imagine from your perspective lacking any actual knowledge of the subject. Frankly for something like this where you configure it and then just keep servicing it, often all you have to do is use something like wireshark to capture a few packets between your PC and it, and then do the same thing in your own code. | |
| Jun 26, 2018 at 22:08 | comment | added | Ben Voigt | @ChrisStratton: ... and userspace is an abstraction level that doesn't exist on an MCU either. | |
| Jun 26, 2018 at 22:06 | comment | added | Chris Stratton | @BenVoigt - you're mixing up host and device considerations. USB is timed by the host. Also, Android USB API is a userspace one. As is something like libusb on linux. | |
| Jun 26, 2018 at 22:06 | comment | added | Ben Voigt | @ChrisStratton: Yeah but the Linux (and Android) ones have to use split interrupt handlers and such. | |
| Jun 26, 2018 at 22:02 | comment | added | Chris Stratton | @BenVoigt - actually it's not very different at all, since they're all just wrapping the same underlying types of USB operations. Further, there are already MCU versions. | |
| Jun 26, 2018 at 21:55 | comment | added | Ben Voigt | @ChrisStratton: Ok, but it's still a significant porting effort to get one of those running on an actual MCU, because the abstraction level is totally different. | |
| Jun 26, 2018 at 19:50 | comment | added | Chris Stratton | @BenVoigt - no, other MCUs are not going to have a terrible time. The protocol is very well understood, as evidenced by the fact that there's been a Linux driver since basically forever, there are implementations atop Android's userspace USB host API, etc. | |
| S Jun 26, 2018 at 19:46 | history | suggested | SolveEtCoagula07 | CC BY-SA 4.0 |
Grammar, capitalization.
|
| Jun 26, 2018 at 18:02 | comment | added | Ben Voigt | FTDI made a microcontroller with a USB host with special support for other FTDI devices (plus HID and mass storage). However it (Vinculum) never had good tooling and is pretty well abandoned. Other MCUs are going to have a terrible time talking to an FTDI due to the fact that it uses a non-standard protocol. | |
| Jun 26, 2018 at 17:34 | comment | added | Chris Stratton | Generally speaking a determined person could bypass the USB-serial chips; quite likely there are test points downstream of them, and even if not it's typically only 2 signals to pick up. Otherwise, a flash-MCU based USB host, while taking longer than a pi to get going, will be far more robust, and potentially low power in operation. | |
| Jun 26, 2018 at 17:25 | comment | added | Wouter van Ooijen | "i only have access to the the interfaces after the FTDI chips" - I presume "after" means the USB side | |
| Jun 26, 2018 at 16:30 | comment | added | Cristobol Polychronopolis | I'm not clear on why you don't just connect the serial ports together without the FTDI in between? | |
| Jun 26, 2018 at 16:24 | review | Suggested edits | |||
| S Jun 26, 2018 at 19:46 | |||||
| Jun 26, 2018 at 16:19 | review | Close votes | |||
| Jul 19, 2018 at 3:04 | |||||
| Jun 26, 2018 at 16:00 | answer | added | Wouter van Ooijen | timeline score: 3 | |
| Jun 26, 2018 at 15:59 | review | First posts | |||
| Jun 26, 2018 at 16:01 | |||||
| Jun 26, 2018 at 15:54 | history | asked | Javier Verde | CC BY-SA 4.0 |