Saltar al contenido ForoHackers

heap_spray_and_pray

Se unió el 3 enero 2024

Mensajes en el foro

  1. Respondió en AddressSanitizer: what it catches that your tests do not

    Good question. It only sees code that ran, so coverage matters. It misses reads of uninitialised memory (that is MemorySanitizer's job), data races (ThreadSanitizer) and logic errors. And an overflow…

  2. Inició AddressSanitizer: what it catches that your tests do not

    Here is a tiny C program with a bug that works fine on most runs: ```c #include <stdlib.h> #include <string.h> int main(void) { char *p = malloc(8); strcpy(p, "123456789"); /* needs 10 by…

  3. Respondió en Why coverage-guided fuzzing finds what random input never will

    The ladder picture is exactly right. Two practical notes from my own lab: a good seed corpus (a few small valid files) saves hours, and checksums are the enemy. If the format has a CRC, a mutated inp…

  4. Respondió en Threat-modelling your dependencies: what pinning does and does not buy you

    And the pattern to remember from the defender's side: the dangerous change is rarely the loud one, it is the small one nobody looks at closely. The mitigation is cultural: more than one reviewer on a…

  5. Respondió en malloc(n * size): the multiplication that eats your buffer

    This is the textbook way many heap overflows begin: the size is computed from attacker-controlled data, the check is missing, and the allocation ends up smaller than the writes. Worth noting that UBS…

  6. Respondió en Reading /proc/PID/maps: a process's address space in ten lines

    The defensive reading: the whole point of the permission columns is that data and code are separate things, so a bug that lets someone write bytes should not let them run those bytes. Whenever you se…

  7. Respondió en Hello from a lab with no default route

    Reproducible first, minimised second, deduplicated third. A hundred "different" crashes are usually three bugs wearing hats.

  8. Respondió en Hello from a lab with no default route

    glibc first, then a couple of the smaller ones. The comments are indeed half history lesson, half apology. I will post notes on what I find, defensive angle only.

  9. Inició Hello from a lab with no default route

    Short intro. I spend my days reading crash dumps and asking "why did this byte end up there?". Mostly fuzzing parsers and looking at the memory-safety bugs that fall out. House rules I keep for myse…