Linux – High memory (user space) and highmem (kernel space)

Is Kernel space used when Kernel is executing on the behalf of the user program i.e. System Call? Or is it the address space for all the Kernel threads (for example scheduler)?

Yes and yes.

Before we go any further, we should state this about memory.

Memory get's divided into two distinct areas:

• The user space, which is a set of locations where normal user processes run (i.e everything other than the kernel). The role of the kernel is to manage applications running in this space from messing with each other, and the machine.
• The kernel space, which is the location where the code of the kernel is stored, and executes under.

Processes running under the user space have access only to a limited part of memory, whereas the kernel has access to all of the memory. Processes running in user space also don't have access to the kernel space. User space processes can only access a small part of the kernel via an interface exposed by the kernel - the system calls. If a process performs a system call, a software interrupt is sent to the kernel, which then dispatches the appropriate interrupt handler and continues its work after the handler has finished.

Kernel space code has the property to run in "kernel mode", which (in your typical desktop -x86- computer) is what you call code that executes under ring 0. Typically in x86 architecture, there are 4 rings of protection. Ring 0 (kernel mode), Ring 1 (may be used by virtual machine hypervisors or drivers), Ring 2 (may be used by drivers, I am not so sure about that though). Ring 3 is what typical applications run under. It is the least privileged ring, and applications running on it have access to a subset of the processor's instructions. Ring 0 (kernel space) is the most privileged ring, and has access to all of the machine's instructions. For example to this, a "plain" application (like a browser) can not use x86 assembly instructions lgdt to load the global descriptor table or hlt to halt a processor.

If it is the first one, than does it mean that normal user program cannot have more than 3GB of memory (if the division is 3GB + 1GB)? Also, in that case how can kernel use High Memory, because to what virtual memory address will the pages from high memory be mapped to, as 1GB of kernel space will be logically mapped?

For an answer to this, please refer to the excellent answer by wag here

1. The kernel mapping exists primarily for the kernel’s purposes, not user processes’. From the CPU’s perspective, any physical memory address which isn’t mapped as a linear address might as well not exist. But the CPU does need to be able to call into the kernel: to service interrupts, to handle exceptions... It also needs to be able to call into the kernel when a user process issues a system call (there are various ways this can happen so I won’t go into details). On most if not all architectures, this happens without the opportunity to switch page tables — see for example SYSENTER. So at minimum, entry points into the kernel have to be mapped into the current address space at all times.

2. Kernel allocations are dynamic, but the address space isn’t. On 32-bit x86, various splits are available, such as the 3/1 GiB split shown in your diagram; on 64-bit x86, the top half of the address space is reserved for the kernel (see the memory map in the kernel documentation). That split can’t move. (Note that libraries are loaded into user space. Kernel modules are loaded into kernel space, but again that only changes the allocations, not the address space split.)

3. In user mode, there is a single mapping for the kernel, shared across all processes. When a kernel-side page mapping changes, that change is reflected everywhere.

   When KPTI is enabled, the kernel has its own private mappings, which aren’t exposed when running user-space code; so with KPTI there are two mappings, and changes to the kernel-private one won’t be visible to user-space (which is the whole point of KPTI).

   The kernel memory map always maps all the kernel (in kernel mode when running KPTI), but it’s not necessarily one-to-one — on 64-bit x86 for example it includes a full map of physical memory, so all kernel physical addresses are mapped at least twice.