Turns out I had confused myself about the situation. Just because a page
has been paged in by another process doesn't mean that the *current*
process's page directory has the present bit set! D'oh. :^)
This exposed a serious race condition in page_in_from_inode().
Reordered the logic and added a paging lock to VMObject.
Now, only one process can page in from a VMObject at a time.
There are definitely ways to optimize this, for instance by making
the locking be per-page instead. It's not something that I'm going
to worry about right now though.
Previously, calling Region::commit() would make sure to allocate any missing
physical pages, but they would contain uninitialized data. This was very
obvious when allocating GraphicsBitmaps as they would have garbage pixels
rather than being filled with black.
The MM quickmap mechanism didn't work when operating on a non-active page
directory (which happens when we're in the middle of exec, for example.)
This patch changes quickmap to reside in the shared kernel page directory.
Also added some missing clobber lists to inline asm that I stumbled on.
Make PageDirectory retainable and have each Region co-own the PageDirectory
they're mapped into. When unmapped, Region has no associated PageDirectory.
This allows Region to automatically unmap itself when destroyed.
The system can finally idle without burning CPU. :^)
There are some issues with scheduling making the mouse cursor sloppy
and unresponsive that need to be dealt with.
The old approach only worked because of an overpermissive accident.
There's now a concept of supervisor physical pages that can be allocated.
They all sit in the low 4 MB of physical memory and are identity mapped,
shared between all processes, and only ring 0 can access them.
This container is really just there to keep a retain on the individual
PhysicalPages for each page table. A HashMap does the job with far greater
space efficiency.
Process page directories can now actually be freed. This could definitely
be implemented in a nicer, less wasteful way, but this works for now.
The spawn stress test can now run for a lot longer but eventually dies
due to kmalloc running out of memory.
mmap() will now map uncommitted pages that get allocated and zeroed upon the
first access. I also made /proc/PID/vm show number of "committed" bytes in
each region. This is so cool! :^)
...by adding a new class called Ext2Inode that inherits CoreInode.
The idea is that a vnode will wrap a CoreInode rather than InodeIdentifier.
Each CoreInode subclass can keep whatever caches they like.
Right now, Ext2Inode caches the list of block indices since it can be very
expensive to retrieve.
- Process::exec() needs to restore the original paging scope when called
on a non-current process.
- Add missing InterruptDisabler guards around g_processes access.
- Only flush the TLB when modifying the active page tables.
This is really sweet! :^) The four instances of /bin/sh spawned at
startup now share their read-only text pages.
There are problems and limitations here, and plenty of room for
improvement. But it kinda works.
All right, we can now mmap() a file and it gets magically paged in from fs
in response to an NP page fault. This is really cool :^)
I need to refactor this to support sharing of read-only file-backed pages,
but it's cool to just have something working.
It only works for sending a signal to a process that's in userspace code.
We implement reception by synthesizing a PUSHA+PUSHF in the receiving process
(operating on values in the TSS.)
The TSS CS:EIP is then rerouted to the signal handler and a tiny return
trampoline is constructed in a dedicated region in the receiving process.
Also hacked up /bin/kill to be able to send arbitrary signals (kill -N PID)
sys$fork() now clones all writable regions with per-page COW bits.
The pages are then mapped read-only and we handle a PF by COWing the pages.
This is quite delightful. Obviously there's lots of work to do still,
and it needs better data structures, but the general concept works.
This turned out way better than the old code. ELF loading is now quite
straightforward, and we don't need the weird concept of subregions anymore.
Next step is to respect the is_writable flag.
This was the fix:
-process.m_page_directory[0] = m_kernel_page_directory[0];
-process.m_page_directory[1] = m_kernel_page_directory[1];
+process.m_page_directory->entries[0] = m_kernel_page_directory->entries[0];
+process.m_page_directory->entries[1] = m_kernel_page_directory->entries[1];
I spent a good two hours scratching my head, not being able to figure out why
user process page directories felt they had ownership of page tables in the
kernel page directory.
It was because I was copying the entire damn kernel page directory into
the process instead of only sharing the two first PDE's. Dang!
