[syzbot] [mtd?] divide error in block2mtd_setup2

[Jörn Engel]

On Wed, Aug 12, 2026 at 09:45:44AM +0800, Pei Xiao wrote:
>
> Great test! You're truly professional, and I've learned a lot. I'll send
> the v2 patch. Thank you for your guidance.

The test was actually a failure of sorts. :)

Actual idea was some kind of fuzzing.  Up to 32bit, computers can test
every possibility in a few seconds.  The number space here is 96bit,
which is a bit much.  But we can still test 4B random combinations or
exhaustively test some subset that is equivalent to about 4B
combinations.

While this is not a mathematical proof, a little bit of brute force is
generally quite cheap and catches the vast majority of bugs.  In this
case, the compiler refused to even generate a brute force test.  So I
guess that gave us something similar to a mathematical proof.

Jörn

--
The rabbit runs faster than the fox, because the rabbit is running for
his life while the fox is only running for his dinner.
-- Aesop