Of course you’re free to make your own choices, but I find it weird to accept relying on many lines of Linux kernel code and then not wanting to reuse a relatively small number of lines of an existing malloc.
The reverse (running on bare metal and tweaking an existing malloc to run on it) looks way more logical to me.
> No other system lets you avoid the libc.
On most OSes [1] it’s relatively easy to write a libOS that just wraps the system calls. Your only dependency would be on the mapping to syscall numbers.
The reverse (running on bare metal and tweaking an existing malloc to run on it) looks way more logical to me.
> No other system lets you avoid the libc.
On most OSes [1] it’s relatively easy to write a libOS that just wraps the system calls. Your only dependency would be on the mapping to syscall numbers.
[1] OpenBSD is/may be (I don’t know the status of these) an exception. See https://man.openbsd.org/pinsyscalls.2, https://lwn.net/Articles/949078/