utting an OS in storage, and then updating it?

[happydud]

So a few friend's and I are building an OS for a peice of hardware we're putting together, but we're a wee bit perplexed as to how/where exactly one would put the Operating system. We were going to toss is on ROM at first, but then we weren't sure if we would be able to update/patch it if needed, it being on Read Only Memory.

So.. any ideas?

Summary: Put an OS on a peice of storage so that it can't be edited by the user from the device, but it can be updated if need be.

 

[bmcculla]

What you need is a separate memory segment to contain a bootloader. The CPU runs code from this second memory and writes to the first memory.

ST has an interesting memory called PSD. Its basicly a bunch of flash and some ram in one package with some programmable logic for custom address decoding. It's JTAG programmable as well as having 2 memory segments to support reading from one while writing to the other. The only down side is its a bit pricy (but not that bad for a fairly large amount of flash) and not all that fast (~10Mhz max).

Brent

 

[Nigel Goodwin]

So a few friend's and I are building an OS for a peice of hardware we're putting together, but we're a wee bit perplexed as to how/where exactly one would put the Operating system. We were going to toss is on ROM at first, but then we weren't sure if we would be able to update/patch it if needed, it being on Read Only Memory.

So.. any ideas?

Summary: Put an OS on a peice of storage so that it can't be edited by the user from the device, but it can be updated if need be.

The classic way is to use UV eraseable EPROM in sockets, it provides everything you want - but you need an eraser and a programmer!. But no matter what sort of memory you use, you're going to require some kind of programmer. You can now get FLASH memory to take the place of EPROM as well, these don't require an eraser.