Commit Graph

23 Commits

Author SHA1 Message Date
Justin C. Miller
ac19d3f532 Allow page table copying and unmapping
Lots of rearranging in page_manager as well, moving constants out as
well as helper structs.
2019-03-03 01:52:21 -08:00
Justin C. Miller
194527e0fe Fix address-marking bugs
* Non-blocksize-aligned regions could fail to be found. Have the
  bootloader load them aligned.
* Consolidating used frame blocks in the bootstrap means these would
  have been impossible to free as address space
* mark_permanent wasn't actually removing blocks from the free list
2019-03-03 01:42:32 -08:00
Justin C. Miller
28cf5562ac Use the address_manager to place allocations 2019-02-28 00:37:00 -08:00
Justin C. Miller
8cdc39fdee Switch page_manager to use frame_allocator.
Removed the frame allocation logic from page_manager and replaced it
with using an instance of frame_allocator instead. This had several
major ripple effects:

- memory_initalize() had to change to support this new world
  - Where to map used blocks is now passed as a flag, since blocks don't
    track their virtual address anymore
  - Instead of the complicated "find N contiguous pages that can be
    mapped in with one page table", we now just have the bootloader give
    us some (currently 64) pages to use both for tables and scratch
    space.
  - frame_allocator initialization was split into two steps to allow
    mapping used blocks before std::move()ing them over
2019-02-28 00:37:00 -08:00
Justin C. Miller
d8399e3c07 Fix for page faults under KVM
Under KVM we were hitting what look like out-of-order and/or issues
during initialization when writing to the page tables and then
immediately writing to the mapped memory.  Adding a memory barrier and
an io_wait() in memory_bootstrap.cpp fixed it.
2018-09-21 20:34:26 -07:00
Justin C. Miller
cef0a71bce Use uintptr_t instead of addr_t
They're never actually going to change independently, and it's also
brining in kutil headers more places than they should be.
2018-09-20 09:37:30 -07:00
Justin C. Miller
593cda3ee8 Convert page_block to use kutil::linked_list
- Created a new linked_list-based slab allocator
- Simplified memory bootstrap code by using the slab allocator and
  linked_lists
2018-09-11 20:46:48 -07:00
Justin C. Miller
dc40c2f6ad Changes from the reorg branch
Add CR4 options: global pages, FXSAVE, PCIDs
Better page manager page-in flags
Remove obsolete rflags-saving in create_process
2018-09-05 22:26:23 -07:00
Justin C. Miller
799fbbdd10 _Actually_ move the kernel to the last TiB.
More work on process page tables, including only mapping the last 2 pml4
entries (the highest 1TiB of the address space, ie, kernel space) into a
new table.

Includes the work of actually moving the kernel there, which I had
apparently done in name only previously. Oops.
2018-09-01 14:54:12 -07:00
Justin C. Miller
0ddcf668cb Allow for 2MiB large pages 2018-05-21 09:07:52 -07:00
Justin C. Miller
9128bfc5f1 Switch to clang and improve cpprt 2018-05-09 01:17:18 -07:00
Justin C. Miller
0f54630725 Move memory_manager and assert into kutil. 2018-05-08 01:11:03 -07:00
Justin C. Miller
b3e49590a7 Add logging framework 2018-04-28 19:18:53 -07:00
Justin C. Miller
25b9625635 paging finally works 2018-04-25 19:53:22 -07:00
Justin C. Miller
bed882f41c Enable paging WIP 2018-04-25 10:48:14 -07:00
Justin C. Miller
1113164505 Join page_block insert methods into one 2018-04-22 23:27:15 -07:00
Justin C. Miller
1de73de2e3 Move page table allocation to top 256GiB.
I forgot to account for tracking page table physical addresses, so
this is a bit of an overhaul. Major changes:
- Refactor bootstrap code into more functions and:
  - Only allocate 32 pages of scratch space
  - Remap remaining space into top 256GiB, the "page table space"
- Use the page table space to directly offset-map page table pages
  from their physical addresses, to avoid tracking overhead.
- Refactor page_block list functions into static functions to better
  handle null/empty lists
2018-04-22 21:52:59 -07:00
Justin C. Miller
95d52b87f4 Initialize page_manager.
Page manager now:
- Caches mapped pages and page_block structs
- Can unmap memory ranges
- Unmaps extra kernel memory during it's init
2018-04-22 02:48:45 -07:00
Justin C. Miller
07fd3abe2c Move page size and higher half offset to constants 2018-04-21 20:58:58 -07:00
Justin C. Miller
5dedd2e0e0 Finish memory bootstrap sequence.
Now we're setting up all our own page tables, and handing off to
page_manager's init function. (Which is still NYI.)
2018-04-21 19:32:39 -07:00
Justin C. Miller
57abb03deb Rearrange memory manager into two classes.
page_manager and memory_manager are now separate, and are also pre
allocated in the kernel so they don't have to allocate themselves.
2018-04-21 17:34:33 -07:00
Justin C. Miller
4a38a74b16 Refactor memory code.
Break out some more bootstrap code into functions. Add the start of
some Doxygen doc comments to help organize my thoughts.
2018-04-21 16:49:39 -07:00
Justin C. Miller
da404f520d Rearrange memory bootstrapping code.
Move EFI-related code and initial memory manager bootstrap code
to memory_bootstrap.cpp, move memory page structs to their own
memory_page.h/cpp files.
2018-04-21 02:52:52 -07:00