10.4 : Warum Inline-Variablen?

[Der schöne Günther]

Ein schönes praktisches Beispiel wie man sich selbst ins Knie schießen kann hat Microsoft kürzlich erzählt. Ein uralter Bug in FAT32 lag im Endeffekt auch an einer unglücklich platzierten Deklaration einer Variable:

(Runterscrollen zu "Use-After-Free in FAT32")
https://msrc-blog.microsoft.com/2020...ry-on-windows/

Zitat:

The code was something along the lines of this:

Code:

for(int i = 0; i < size; i++)
{
      int tmp;
      DoStuff(&tmp, i);
}

This code was running in a loop. A variable was declared inside of the loop. On the first iteration of the loop, the function DoStuff would initialize the variable “tmp” that it is passed the address to. On every additional iteration of the loop, the variable “tmp” was used as an in/out parameter. In other words, the variable would be read from and then updated.

The issue here is that the variable comes in scope at the beginning of each iteration of the loop and goes out of scope after each iteration of the loop. With InitAll enabled, this variable is zero initialized for each iteration of the loop. This is effectively a use-after-free. This code is depending on the value of “tmp” being preserved each iteration of the loop even though it goes out of scope at the end of each iteration.