kernel: support memory mapped stacks #70810
Commits on Apr 1, 2024
-
x86: gen_gdt: remove extra unnecessary parens
Pylint complains so we fix. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
x86: do not advertise demand paging support for x86_64
x86_64 does not currently support demand paging so don't advertise it. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
x86: synchronize usage of CONFIG_X86_STACK_PROTECTION
Most places use CONFIG_X86_STACK_PROTECTION, but there are some places using CONFIG_HW_STACK_PROTECTION. So synchronize all to use CONFIG_X86_STACK_PROTECTION instead. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
x86: correct size for stack bound check for privileged stack
Previous commit changed the privileged stack size to be using kconfig CONFIG_PRIVILEGED_STACK_SIZE instead of simply CONFIG_MMU_PAGE_SIZE. However, the stack bound check function was still using the MMU page size, so fix that. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
kernel: add options to cleanup after aborting current thread
This adds the mechanism to do cleanup after k_thread_abort() is called with the current thread. This is mainly used for cleaning up things when the thread cannot be running, e.g., cleanup the thread stack. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
kernel: mm: introduce k_mem_phys_map()/_unmap()
This is similar to k_mem_map()/_unmap(). But instead of using anonymous memory, the provided physical region is mapped into virtual address instead. In addition to simple mapping physical ro virtual addresses, the mapping also adds two guard pages before and after the virtual region to catch buffer under-/over-flow. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
kernel: add the ability to memory map thread stacks
This introduces support for memory mapped thread stacks, where each thread stack is mapped into virtual memory address space with two guard pages to catch under-/over-flowing the stack. This is just on the kernel side. Additional architecture code is required to fully support this feature. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
tests: mem_protect/userspace: extend tests for mapped stack
This extends the test for memory mapped stack, as the address of memory mapped stack object would be different than the actual stack object. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
tests: threads/thread_stack: extend tests for mapped stack
This extends the test for memory mapped stack, as the address of memory mapped stack object would be different than the actual stack object. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
cmsis_rtos_v1: support memory mapped stack at abort hook
This extends the thread abort hook to support memory mapped stack. The calculation to find out to which instance of thread pools the outgoing thread belongs requires physical address. So find the physical address via the memory mapped stack for that. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
tests: mem_protect/stackprot: test mapped stack
This extends the mem_protect/stackprot tests to support testing with memory mapped stack. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
boards: qemu_x86/atom/virt: forcibly disable memory mapped stack
The current memory mapped stack code requires actual physical addresses for stacks, and cannot deal with stacks already using virtual addresses. So disable mapped stack via defconfig. Note that this is done before enabling memory mapped stacks on x86 so test won't fail when the support in the architecture code is introduced. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
-
x86: add support for memory mapped stack for threads
This adds the necessary bits to enable memory mapping thread stacks on both x86 and x86_64. Note that currently these do not support multi level mappings (e.g. demand paging and running in virtual address space: qemu_x86/atom/virt board) as the mapped stacks require actual physical addresses. Signed-off-by: Daniel Leung <daniel.leung@intel.com>
