My bootloader must be in the mdt1?
That's wrong ... probably the "definition" of MTD1 boundaries was the same, as for a 7490 (I'll not check this in a 5490, at least not yet, but someone else could look into a generated "support file" to verify this.):
Code:
mtd0 0x400000,0x3400000
mtd1 0x0,0x400000
mtd2 0x0,0x40000
mtd3 0x40000,0xA0000
mtd4 0xA0000,0x100000
mtd5 0x0,0x200000
But as it was written now multiple times (even in this thread), you should
know, what you're doing.
And so it doesn't matter, which name is used by the "urlader" code partition (from the view point of this boot-loader, while it's running and in control of the device) - it's usually rather "mtd2". But you really can't read/export the content of this partition with any function, that's provided by a running EVA instance.
And if you look at the definition for "mtd1" and "mtd2" from above, you'll notice, that they share the first 256 KB of the flash memory (start at 0x0, size of mtd1 is 0x400000 and size of mtd2 is 0x40000 - but the second value is usually rather the last address (+1) and not a size value ... but combined with a start address of 0x0, the values are identical) and so it's not really surprising, if a write access to "mtd1" kills the bootloader code from the first 256 KB, if the FTP server accepts a write attempt and erases the SPI flash in a size of 0x400000 (4 MB - that's more than the real size of the SPI memory) as first step.
Sure, it's a big bug from the FTP server, that it accepts and executes such a write attempt ... but it's also a big mistake of the user, to try such one for the wrong target.
Older devices (here a 7390 with NOR flash, that's mapped to a start address of 0x9F000000 on the memory bus) use definitions for "mtd2" and "mtd1", that do not share a common section:
Code:
mtd0 0x9F000000,0x9F000000
mtd1 0x9F020000,0x9FF00000
mtd2 0x9F000000,0x9F020000
mtd3 0x9FF00000,0x9FF80000
mtd4 0x9FF80000,0xA0000000
If AVM did not correctly enable or disable one command or target name from another model in the EVA code, such failures may occur - but usually they'll never get noticed, because no-one uses the wrong commands/targets for such write attempts - as long as (s)he knows, how to do it in the right manner.
And if you don't try to collect the correct info for a box, BEFORE you start any further experiments, you will kill the next device soon. Meanwhile your question, as cited at the very beginning of this post, is practically the next "misunderstanding" and not really a good sign for a systematical collection of (correct) infos.
Sure, sometimes it's useful, if anyone has the "cohones", to dare to try even unusual things. But I'd bet, correct info, that was collected in preparation for such a "brave action", will help much more, than any premature conclusions, based on old and outdated articles.
So you should try to collect such info and - if you're unsure, whether they're still valid - you may ensure here.
But such questions (at least with this - rather unusual - intention) should have a stable base of good knowledge, contain a compact description of the circumstances/conclusions/facts and shouldn't occur after the first read source already - and you should be able to show, where you found an info, if it's wrong (the still open question, how you came to the idea to write to mtd1 on a 5490 model).
Otherwise also the "helpers" will become responsible for the premature death of many innocent FRITZ!Box devices - what a horrible notion.
Addition: And to clarify another point, too ... the view of the running system on the flash partitions via "/proc/mtd" (and therefore even via "/dev/mtd[block]X") is a different one and completely unrelated to the view of EVA. Do not mix up the numbers from the views of different systems (EVA vs. FRITZ!OS).
