Sometimes it is a kernel bug: bind() in Linux 6.0.16

Re:Sometimes it is a kernel bug: bind() in Linux 6.0.16 by astrange on Friday January 13, 2023 @04:41PM (#34374286)

If you don't hit compiler bugs, it's mostly because you're not updating your compiler.

Large companies with compiler teams want to use them, so they do update their compiler, so they will find bugs in it.

Btw, what do you think this prints with clang? (Whether the answer is a compiler bug is debatable…)

   printf("%#x\n", 1 << 32);

[ Parent | on HN ]

Re:Sometimes it is a kernel bug: bind() in Linux 6.0.16 by valleyer on Friday January 13, 2023 @05:46PM (#34374967)

Assuming int is 32 bits wide, `1 << 32` invokes UB.

[ Parent | on HN ]

Re:Sometimes it is a kernel bug: bind() in Linux 6.0.16 by astrange on Friday January 13, 2023 @05:57PM (#34375068)

Indeed it does.

(Spoiler: it prints uninitialized stack memory. A bit closer to demons flying out your nose than you'd expect!)

[ Parent | on HN ]

Re:Sometimes it is a kernel bug: bind() in Linux 6.0.16 by valleyer on Saturday January 14, 2023 @05:01PM (#34384574)

Fun. Basically, it decides that (due to UB) it can simply not materialize the argument and carry on.

On Darwin/x86_64, this actually means that it prints out bits from an "uninitialized" register, specifically %rsi in this case, since that's where the second argument should be (even for variadic functions).

Certainly a jarring failure mode! However, if updating your compiler causes you to run into this, you already have a bug in your code (one that clang -- arguably not forcefully enough -- already warns you of).

[ Parent | on HN ]

Re:Sometimes it is a kernel bug: bind() in Linux 6.0.16 by Agingcoder on Saturday January 14, 2023 @06:27AM (#34379406)

This is exactly when I find most of my compiler bugs : every time we upgrade.

[ Parent | on HN ]