RetroBSD

2.11BSD operating system for microcontrollers
It is currently Thu Jun 27, 2019 2:53 am

All times are UTC




Post new topic Reply to topic  [ 29 posts ]  Go to page Previous  1, 2
Author Message
PostPosted: Wed Jan 20, 2016 4:56 pm 
Committer
User avatar

Joined: Thu Oct 11, 2012 8:45 am
Posts: 1801
Location: Room 217, Floor 8, Arm 8, Wheel S7, Mars Base Alpha 3
Ah right, that's your usage of the word "native". It confused me.

Ok - so by your definition (and it's actually mine too if I got you right) is:
  • MCU: A CPU core with peripherals and flash and ram on a single die which can run with no external (except passives etc) components
  • SoC: A CPU core with peripherals but no flash or ram which requires external flash and ram to execute. An internal ROM is often included as a bootloader.
  • CPU: A CPU core with no peripherals (except maybe simple ones like a timer or watchdog) which requires external peripherals, RAM and Flash / ROM to operate. There is never any internal ROM but a pre-defined execute path from external Flash / ROM

_________________
Why not visit my shop? http://majenko.co.uk/catalog
Universal IDE: http://uecide.org
"I was trying to find out if it was possible to only eat one Jaffa Cake. I had to abandon the experiment because I ran out of Jaffa Cakes".


Top
 Profile  
 
PostPosted: Wed Jan 20, 2016 5:09 pm 
Contributor
User avatar

Joined: Thu Nov 08, 2012 7:04 am
Posts: 2401
Location: Rapa Nui
SoC is rather a marketing buzz. I do not care what is on the chip actually, the question is whether the "cpu" has got the rom/ram on the die, or, whether it is off the die..
Allwinner A20 is called a SoC as it has got the whole desktop PC on the chip, except a dram. Therefore I would call it CPU and not MCU in this regard.

So back to my Q: what is the market share for MIPS based MCUs (as the MCUs are the core biz of Atmel and Microchip)?

PS: my guess is 0.023%.

_________________
Pukao Hats Cleaning Services Ltd.


Top
 Profile  
 
PostPosted: Thu Jan 21, 2016 3:24 am 
User avatar

Joined: Fri Mar 06, 2015 9:13 pm
Posts: 37
Location: Barcelona, SP
https://olimex.wordpress.com/2016/01/20/you-guys-will-buy-your-avrs-from-microchip-from-now-on/ 8-) 8-)

http://www.microchip.com/investor/Pressrelease/Atmel%20Acquisition%20Investor%20Presentation.011916.pdf 8-) 8-)

_________________
Dennis MacAlistair Ritchie (Sep.09.1941 - Oct.12.2011)


Top
 Profile  
 
PostPosted: Thu Jan 21, 2016 10:12 am 
Contributor
User avatar

Joined: Thu Nov 08, 2012 7:04 am
Posts: 2401
Location: Rapa Nui
That is why I am asking experts herein on the actual pic32mx/mz market share or product line revenues - see the Olimexe's comment - I think the same - the pic32mx/mz has not been a success at Microchip..
They spent huge effort into it, came late, with issues, they managed expectation poorly. The main marketing point - the huge ram (as of 2008) and even a bigger ram with mz (as of today) seems to be the only differentiator on the market. I doubt MCHP is happy about the pic32mx/mz endeavor.. So, we will see within next 6months whether they will still invest in this product line, or..
Hopefully yes, of course :)
PS: as I was a part of big "merges" several times in past :P - my lessons learned - next 6months there will be a fight between product lines (led by their managers) on both sides and it will be held about $$ only. No big discussions with techies :)

_________________
Pukao Hats Cleaning Services Ltd.


Top
 Profile  
 
PostPosted: Thu Jan 21, 2016 10:38 pm 
User avatar

Joined: Fri Mar 06, 2015 9:13 pm
Posts: 37
Location: Barcelona, SP
They make the effort to get out EF series nice series can look at the erratas most of them gone, and really for me it's nice CPU+MCU+FPU and really we can call it SoC I never understand what it's the problem for the manufacturers to include big RAM inside the main chip, I don't know so much physical IC's with the capacity to run powerful OS from few capacitors and a few resistors, the key will be when they releases MZDA, and MZ series drop it's price for very affordable price range 3-5 $ in small quantities.

Really IoT it's growing market and I think LiteBSD will be perfect companion for that on a coin sized device you can run LiteBSD with a bash script, make something interesting and then sleeping.

The framework ecosystem for the MZ proposed by the manufacturer, and all the series 8bits 16bits is turning a huge and really complicated ecosystem based on something called "assistant" :? :? in the other hand POSIX based platform is perfect ramp to take off in working system, and with the compiler inside the system your tool becomes a terminal unit to connect and deploy.

A real deploy on IoT is complicated, limited comunications, firmware updates, to a lot of nodes deployed etc, thinking on BSD system running for me it's a dream that comes true, If we can solve the power consumption things, etc, really it becomes fantastic !!

At all nice "fusion" and huge portfolio coming soon ..... hope the documentation and availability of the Atmel products will increase.... now then we will have 3 big actors on MCU scene, I think;

- TI
- STM
- MCHP

_________________
Dennis MacAlistair Ritchie (Sep.09.1941 - Oct.12.2011)


Top
 Profile  
 
PostPosted: Thu Jan 21, 2016 10:46 pm 
Committer
User avatar

Joined: Thu Oct 11, 2012 8:45 am
Posts: 1801
Location: Room 217, Floor 8, Arm 8, Wheel S7, Mars Base Alpha 3
The one thing that will make LiteBSD really rock for the IoT is auto-self updating. That is downloading and installing not only a new version of the binaries for packages, but the whole kernel too. That way nodes can keep themselves up to date with new packages and kernel without you having to manually connect to each node and update it. Of course, you should be able to schedule it to a good time (cron) :)

The equivalent to "apt-get update && apt-get upgrade -y" on Debian.

So the hard part will be how to upgrade the kernel. I guess it would require a package that downloads a kernel image and writes it to flash then reboots into the new version of the kernel. I am wondering how easy it would be to "lock" a process into execution so the swapper never swaps it out (effectively turn off the kernel) so it can erase the flash then write a new kernel into it before rebooting. Or would it be better to make use of the split-flash facility and split the flash in half - have half as the running kernel and the other half as the newly installed kernel - then switch the split boot flag thingy (whatever that is - something the bootloader could do?) to boot off the other half of the flash - gives you a fallback to boot off the old kernel if needed too. Is that possible do you think?

_________________
Why not visit my shop? http://majenko.co.uk/catalog
Universal IDE: http://uecide.org
"I was trying to find out if it was possible to only eat one Jaffa Cake. I had to abandon the experiment because I ran out of Jaffa Cakes".


Top
 Profile  
 
PostPosted: Fri Jan 22, 2016 7:43 am 
Committer
User avatar

Joined: Wed Oct 10, 2012 11:01 pm
Posts: 1081
Location: Sunnyvale, CA
majenko wrote:
Serge, do MIPS actually produce any chips of their own these days (for public consumption)? I know that back in the day they used to manufacture CPUs (DEC used them in the DECstations that I used to use when younger) but do they actually make them these days, or is it all about licensing to other manufacturers?
No: last 16 years, since it was spun off by SGI, MIPS does not produce chips by itself. It only sells licenses for their cores. It's a pure IP company. Actually the world's first company of such kind, as MIPS is believed to invent the IP licensing business model.


Top
 Profile  
 
PostPosted: Fri Jan 22, 2016 7:50 am 
Committer
User avatar

Joined: Wed Oct 10, 2012 11:01 pm
Posts: 1081
Location: Sunnyvale, CA
majenko wrote:
The one thing that will make LiteBSD really rock for the IoT is auto-self updating.

Definitely we can do this. The update can be performed directly from the bootloader, on reboot. It's not so hard to read the kernel image directly from SD card and program it into the main Flash memory. This can be combined with split-flash technique for reliability.


Top
 Profile  
 
PostPosted: Fri Jan 22, 2016 9:22 am 
Contributor

Joined: Mon Nov 12, 2012 1:34 pm
Posts: 1092
Hi Serge,

A simple way to begin might be a 'user level' application program to:

1. Update kernel from SD card- Writing control program running in bootflash or RAM.

2. Update either boot space from SD card- Could be done from kernel flash or RAM.

It is probably a VERY good idea to include the ability to be able to store whatever in the 'unused' spaces in kernel and boot flash memory without modifying the main kernel or used portion of the bootloader.

Once running, I can try out the latest boot loader or kernel by just plugging in the right SD and telling the programmer to go.

And I can backup to the previous kernel by putting in an older SD and telling it to go.

Or in my case, I can easily put my version of the kernel into the PIC32 to try it out. Debug it, etc. Or my serial version of the bootloader.

Wiz


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 29 posts ]  Go to page Previous  1, 2

All times are UTC


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
cron




Powered by phpBB® Forum Software © phpBB Group

BSD Daemon used with permission