The state of Twitter programmers

The state of Twitter programmers
>inb4 Twitter programmers

>2^64 bytes of memory on x86
>NULL isn't a valid address for memory access

Attached: D7c5ukNXkAAZuKA.jpg (1200x783, 189K)

Isn't the real address of 0x0 memory mapped by the BIOS?

Not to the program. What the program sees is "virtual memory". Basically paged addresses laid out for it by the operating system. Nearly all OS's map the page at 0x0 to something usable, since programs expect to at least be able to read 0x0.

0x0 has to be readable, because NULL is (often) defined as a void pointer to 0x0. Meaning, NULL->NULL (null should point to null, but this requires 0x0 to be readable, and present in the virtual memory map).

No, the first megabyte of memory is RAM on x86. But most OSes map this segment as protected for the obvious reason as catching NULL pointer exceptions (as general protection faults).

Also, this >0x0 has to be readable
Actually no, dereferencing a NULL pointer should also lead to a segmentation fault.
int a = *((int*) NULL); // this should lead to a segmentation fault

Null pointer exceptions are a high-level concept tho. Reading 0x0 and therefore NULL is legal. C programs get a SEGFAULT when they touch an invalid page which usually only happens when another type of check fails.

>programs expect to at least be able to read 0x0
wut

>Null pointer exceptions are a high-level concept tho
They're not, they are triggered by hardware.

>Reading 0x0 and therefore NULL is legal
Usually no. See >C programs get a SEGFAULT when they touch an invalid page
Yes, correct. That's why the the lower end of memory is mapped but protected.

>2^64 bytes of memory on x86
current implementations only provide 48bits, but the standard does allow for a 64bit implementation

Was meant for