apue | Advanced Programming in UNIX Environment

From this sigaction manual page:

This is a pointer to a ucontext_t structure, cast to void *. The structure pointed to by this field contains signal context information that was saved on the user-space stack by the kernel; ... Commonly, the handler function doesn't make any use of the third argument.

So this argument is a kernel-specific pointer to a special signal context. It's not possible to reuse it for user-data.

There’s nothing particularly special about this function. It’s specified by POSIX to read a single, newline-delimited line from stream, into the buffer whose address is given by lineptr and which must be large enough to hold n bytes.

The reason lineptr and n are pointers is that they are used both as input to the function and potentially output from it. If lineptr is NULL on entry, or n indicates that its size is too small to hold the line read from stream, then getline (re)allocates a buffer and updates lineptr and n with the corresponding information.

getline is easier to use than fgets because the latter will stop reading when it reaches the end of the buffer. So if fgets tries to read a line longer than the provided buffer, it will return a partial read and the caller will have to read again and connect the multiple parts. In such circumstances, getline would reallocate the provided buffer, if any.

As explained in the GNU C Library documentation,

Standard C has functions to do this, but they aren’t very safe: null characters and even (for gets) long lines can confuse them. So the GNU C Library provides the nonstandard getline function that makes it easy to read lines reliably.

(getline originated in the GNU C Library and was added to POSIX in the 2008 edition.)

I think that's because according to the manual of pthread_join, it stores the return value of the thread, not a memory address: as you can see on the thr_fn1 and 2 functions, the return value is (void *)1 and (void *)2.

Dereferencing the pointer is pointless, as we are interested in his value, not the address he points to.

The second case uses this technique because ptr is actually a memory address.