Page faults occurring on RO pages with the CopyOnWrite flag set will be
handled by the page handler as follows:
- allocate new frame
- establish temporary mapping for new frame
- copy original page to new frame
- update entry for the page where the fault occurred:
- set physical frame address to the allocated frame
- clear CoW flag and set Present, RW flags
- return from the fault handler to resume execution at the instruction
that caused the fault
Any other page faults will still cause a kernel panic
Function EarlyReserveRegion reserves contiguous virtual address space
regions beginning at the end of the available kernel space and moving
towards lower virtual addresses. The only state that is tracked by this
function is the last allocated virtual page address which is adjusted
after each reservation request.
Starting at the end of the kernel address space ensures that we will not
step on the virtual addresses used by the kernel code and data sections.
This allows us to remove the allocFn argument from the vmm functions
which causes the compiler's escape analysis to sometimes incorectly flag
it as escaping to the heap.
The linked.ld script is extended to include the _kernel_start and
_kernel_end symbols which are passed by the rt0 code to Kmain. The
allocator converts these addresses to a start/end frame index by
rounding down the kernel start address to the nearest page and rounding
up the kernel end address to the nearest page.
When allocating frames, the allocator will treat the region defined by
these 2 indices as reserved and skip over it.
AllocFrame now rounds up the region start address to the nearest page
multiple and rounds down the region end address to the nearest page
multiple. It also ignores memory regions with size smaller than a page.
Instead of using frame indices and converting them to a pmm.Frame, the
allocator now just keeps track of the last allocated pmm.Frame.
As the allocator is now unexported, a package-exported Init() method is
now provided whose purpose is to initialize the physical allocator
sub-system.
The Map/Unmap methods of PageDirectoryTable operate similar to the
global Map/Unmap functions. While the global functions work with the
currently active PDT, the PageDirectoryTable methods can also
work with inactive page tables by temporarily modifying the recursive
mapping of the active PDT to point to the inactive PDT frame before
delegating the mapping/unmapping to the global Map/Unmap functions.
The API provides the Map() and MapTemporary() functions that establish
virtual -> physical address mappings using the currently active page
directory table.
Mapped pages can be unmapped using the Unmap() function. When unmapping
virtual addresses, the page tables leading to them will not be
automatically released even if they are empty. This will be addressed by
a future commit.
The page table walker provides a mechanism for accessing the individual
page table entries that correspond to a particular virtual memory
address. This implementation will serve as the basis for implementing
page mapping/unmapping and virtual to physical address translation.
This is equivalent to pmm.Frame (also a uintptr) but having different
types for physical and virtual frames serves as an additional layer of
protection for functions/methods that receive physical and/or virtual
page arguments.
To keep the implementation portable, the Frame type had to be changed
from uint64 to uintptr. Using uintptr ensures that the frame will always
match the pointer size of the platform.
When converting strings to []byte so that they can be used with the tty
io.Writer interface Go calls a runtime method called
"stringtoslicebyte". If the input length exceeds a particular size then
this method will allocate a new []byte and copy the data into it. This
obviously causes our kernel to crash.
To fix this, all early_fmt functions have been changed to iterate any
string arguments and output them to the active TTY one byte at a time.
The switch to 64-bit mode allows us to use 48-bit addressing and to
relocate the kernel to virtual address 0xffff800000000000 + 1M. The
actual kernel is loaded by the bootloader at physical address 1M.
The rt0 code has been split in two parts. The 32-bit part provides the
entrypoint that the bootloader jumps to after loading the kernel. Its
purpose is to make sure that:
- the kernel was booted by a multiboot-compliant bootloader
- the multiboot info structures are copied to a reserved memory block
where they can be accessed after enabling paging
- the CPU meets the minimum requirements for the kernel (CPUID, SSE,
support for long-mode)
Since paging is not enabled when the 32-bit code runs, it needs to
translate all memory addresses it accesses to physical memory addresses
by subtracting PAGE_OFFSET. The 32-bit rt0 code will set up a page table
that identity-maps region: 0 to 8M and region: PAGE_OFFSET to
PAGE_OFFSET+8M. This ensures that when paging gets enabled, we will still
be able to access the kernel using both physical and virtual memory
addresses. After enabling paging, the 32-bit rt0 will jump to a small
64-bit trampoline function that updates the stack pointer to use the
proper virtual address and jumps to the virtual address of the 64-bit
entry point.
The 64-bit entrypoint sets up the minimal g0 structure required by the
go function prologue for stack checks and sets up the FS register to
point to it. The principle is the same as with 32-bit code (a segment
register has the address of a pointer to the active g) with the
difference that in 64-bit mode, the FS register is used instead of GS
and that in order to set its value we need to write to a MSR.
The kfmt/early package provides a minimal printf implementation that
does not use memory allocations (everything is allocated on the stack).
This implementation can be used to emit debug messages before the memory
manager is initialized.
We still keep the required main func in stub.go to prevent the compiler
from optimizing the code out. We also force the compiler not to inline
the call to kernel.Kmain so we can find the symbol in the generated .o
file.
Since the actual size of each memory entry is not known in advance
(bootloaders may append additional information to it) but needs to be
queried off the memory map tag header we cannot reserve space for it as
no memory allocation is yet available.
Instead, a visitor pattern was implemented to allow the memory
manager initialization block to easily mark the appropriate pages as reserved
