| ??? 10/03/07 17:15 Read: times |
#145302 - suggestions/answers Responding to: ???'s previous message |
So, if anyone can offer robust methods to prevent this from happening (other than "don't let your code do that") I am all ears.
there is no such thing short of using a device with less than 64k flash and having the bootloader in external ROM. That, however will not be totally 'bullrtproof' since the reset vector will be in flash. What I do, (using SILabs devices) is to have an 'emergency' JTAG connector on the board. I guess Atmeel (that I would never use) has a "pull this low, pull this high" means of kicking the uC into ISP and a means of doing so would also provide an 'emergency method'. the obvious answer would be: "do not write errant firmwarer" :) ESPECIALLY for medical devices. ABOVE ALL use a quality supervisor. Micro 1: ~0.5 MIPS, >18 IO, internal EEPROM, 5V IO, ISP with bootloader protection sounds like a NXP P89V51Rx2 Micro 2: >4.5 MIPS, >32 IO, internal EEPROM, 5V IO, ISP with bootloader protection, 64k FLASH. looking for speed I would abandon the 5V and go for a SILabs device (the pins ARE '5V tolerant") Essentially I'm looking for the Atmel micros WHY? Ereik |
| Topic | Author | Date |
| IAP/ISP bootloader destruction vulnerability | 01/01/70 00:00 | |
| suggestions/answers | 01/01/70 00:00 | |
| certainly not the P89V51Rx2 | 01/01/70 00:00 | |
| Different devices ? | 01/01/70 00:00 | |
this is EXACTLY how it works with the Atmels... | 01/01/70 00:00 | |
| bootloader corruption? | 01/01/70 00:00 |



