[kernel] Use IST for kernel stacks for NMI, #DF, #PF

We started actually running up against the page boundary for kernel
stacks and thus double-faulting on page faults from kernel space. So I
finally added IST stacks. Note that we currently just
increment/decrement the IST entry by a page when we enter the handler to
avoid clobbering on re-entry, but this means:

* these handlers need to be able to operate with only a page of stack
* kernel stacks always have to be >1 pages
* the amount of nesting possible is tied to the kernel stack size.

These seem fine for now, but we should maybe find a way to use something
besides g_kernel_stacks to set up the IST stacks if/when this becomes an
issue.
This commit is contained in:
Justin C. Miller
2021-02-02 18:27:37 -08:00
parent 8575939b20
commit 68a2250886
8 changed files with 166 additions and 87 deletions

View File

@@ -9,6 +9,7 @@
#include "device_manager.h"
#include "frame_allocator.h"
#include "gdt.h"
#include "io.h"
#include "log.h"
#include "msr.h"
@@ -117,6 +118,29 @@ memory_initialize_post_ctors(args::header &kargs)
reinterpret_cast<uintptr_t>(kargs.page_tables),
kargs.table_count);
using memory::frame_size;
using memory::kernel_stack_pages;
constexpr size_t stack_size = kernel_stack_pages * frame_size;
for (int ist = 1; ist <= 3; ++ist) {
uintptr_t bottom = g_kernel_stacks.get_section();
log::debug(logs::boot, "Installing IST%d stack at %llx", ist, bottom);
// Pre-realize and xerothese stacks, they're no good
// if they page fault
kutil::memset(reinterpret_cast<void*>(bottom), 0, stack_size);
// Skip two entries to be the null frame
tss_set_ist(ist, bottom + stack_size - 2 * sizeof(uintptr_t));
}
#define ISR(i, s, name) if (s) { idt_set_ist(i, s); }
#define EISR(i, s, name) if (s) { idt_set_ist(i, s); }
#define IRQ(i, q, name)
#include "interrupt_isrs.inc"
#undef IRQ
#undef EISR
#undef ISR
}
static void