<bslsk05>
elixir.bootlin.com: Making sure you're not a bot!
<Ermine>
_Heat: wow
<Ermine>
i looked in src/thread
Left_Turn has joined #osdev
Turn_Left has quit [Ping timeout: 248 seconds]
netbsduser`` has quit [Read error: Connection reset by peer]
Maja has quit [Ping timeout: 244 seconds]
sdomi has quit [Ping timeout: 265 seconds]
Turn_Left has joined #osdev
Maja has joined #osdev
sdomi has joined #osdev
Left_Turn has quit [Ping timeout: 252 seconds]
GeDaMo has quit [Quit: 0wt 0f v0w3ls.]
dumb0 has joined #osdev
Matt|home has joined #osdev
dumb0 is now known as kb26
ZipCPU has quit [Ping timeout: 260 seconds]
kb26 is now known as daax
FreeFull has joined #osdev
daax has quit [Quit: leaving]
daax has joined #osdev
manmshuk has joined #osdev
manmshuk has left #osdev [#osdev]
daax has quit [Quit: leaving]
daax has joined #osdev
jcea has joined #osdev
FreeFull has quit []
netbsduser has joined #osdev
Lucretia has quit [Remote host closed the connection]
<Ameisen>
ugh. I think that I just found a long-lasting but random segfault cause. Took a long time. recompiled chunks stored jump patches (for address lookups) in a `std::vector<patch>`. However, when generating the machine code, we directly inject the address of that entry into the code. Thus, if the patch vector grows past capacity, we are now potentially pointing to arbitrary memory (usually a pointer in the chunk table) and clobbering data when a jump