FS#48 - SQUASHFS error: xz decompression failed, data probably corrupt about openwrt HOT 5 CLOSED

stefanct:

I can now confirm that it happens even if the MD5 in LuCI matches.

from openwrt.

jow-:

Can you reproduce this error on unmodified hardware?

from openwrt.

stefanct:

No, since I don't own an unmodified model. Since this one is going to be used "in production" soon, I certainly won't replace the flash chip again to test this out. If you have other things I could test then please tell.

However, from what I can tell the model itself really should not matter since the flash chips are mostly compatible and differ only in details that are not used by the MTD layer AFAICT (I know flash chips in detail but not the MTD layer). Also, the SCK is only 25 MHz AFAICT so it really shouldn't be a signal deterioration either. If at all then the chip itself or the board is damaged. In that case however there must still be a change in the code between 15.05.1 and the current-ish HEAD because with the former flashing always succeeded so far.

from openwrt.

jow-:

No I don't have any other suggestions for tests. As long as this bug isn't confirmed on unmodified HW I'm reluctant to spend any effort into this. You could try to bisect the error.

from openwrt.

john:

flashed a unmodified 3040 and it worked 3 times in a row.

from openwrt.