OpenBSD manual page server

Manual Page Search Parameters

KBIND(2) System Calls Manual KBIND(2)

kbindupdate protected memory for lazy-binding

#include <sys/unistd.h>

struct __kbind {
	void	*kb_addr;
	size_t	 kb_size;
};
#define KBIND_BLOCK_MAX 2  /* powerpc and sparc64 need 2 blocks */
#define KBIND_DATA_MAX  24 /* sparc64 needs 6, four-byte words */


int
kbind(const struct __kbind *param, size_t psize, int64_t cookie);

The kbind syscall updates the contents of one or more blocks of the process's memory with the supplied replacement data. This is used to efficiently and securely perform lazy-binding.

param points to an array of __kbind structures giving the addresses and lengths to update. The last __kbind structure is immediately followed in the same order by the source data for the blocks to copy. psize specifies the total length of the parameters: both the __kbind structures and the associated source data. There may be at most KBIND_BLOCK_MAX __kbind structures and each block may be at most KBIND_DATA_MAX bytes in size.

kbind updates memory “as if” the thread temporarily made the memory writable with mprotect(2). If the process does not have write access to the underlying memory object, kbind will fail without making the requested change.

kbind is currently intended for use by ld.so(1) only. It is therefore not provided as a function and two security checks are performed to bind it (pun intended) to its use in ld.so(1): the first time kbind is used, the kernel records both the text address of the call and the value of the cookie argument. If those values differ in a later kbind call, then the process is killed.

Upon successful completion, the value 0 is returned; otherwise the value -1 is returned and the global variable errno is set to indicate the error.

kbind() will fail if:

[]
Cannot allocate memory when updating a copy-on-write page.
[]
The parameters are inconsistent or exceed kbind limits.
[]
Part of the structures or additional data pointed to by param is outside the process's allocated address space, or the underlying memory object is not writable.

ld.so(1), mprotect(2)

The kbind syscall is specific to the OpenBSD dynamic linker and should not be used in portable applications.

The kbind syscall appeared in OpenBSD 5.8.

November 21, 2021 OpenBSD-current