1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-24 20:57:35 +00:00
Commit graph

579 commits

Author SHA1 Message Date
Andreas Kling
13a2e91fc5 Kernel: No need to use safe_memcpy() when handling an inode fault
We're copying the inode contents from a stack buffer into a page that
we just quick-mapped, so there's no reason for this memcpy() to fail.
2021-07-23 14:19:47 +02:00
Andreas Kling
082ed6f417 Kernel: Simplify VMObject locking & page fault handlers
This patch greatly simplifies VMObject locking by doing two things:

1. Giving VMObject an IntrusiveList of all its mapping Region objects.
2. Removing VMObject::m_paging_lock in favor of VMObject::m_lock

Before (1), VMObject::for_each_region() was forced to acquire the
global MM lock (since it worked by walking MemoryManager's list of
all regions and checking for regions that pointed to itself.)

With each VMObject having its own list of Regions, VMObject's own
m_lock is all we need.

Before (2), page fault handlers used a separate mutex for preventing
overlapping work. This design required multiple temporary unlocks
and was generally extremely hard to reason about.

Instead, page fault handlers now use VMObject's own m_lock as well.
2021-07-23 03:24:44 +02:00
Andreas Kling
64babcaa83 Kernel: Remove unused MAP_SHARED_ZERO_PAGE_LAZILY code path 2021-07-23 03:24:44 +02:00
Andreas Kling
e44a41d0bf Kernel: Convert Region to east-const style 2021-07-22 23:34:33 +02:00
Gunnar Beutner
f2be1f9326 Kernel: Fix the variable declaration for some linker script symbols
Despite what the declaration would have us believe these are not "u8*".
If they were we wouldn't have to use the & operator to get the address
of them and then cast them to "u8*"/FlatPtr afterwards.
2021-07-22 22:27:11 +02:00
Andreas Kling
0642f8f2c6 Kernel: Make committed physical page allocation return NonnullRefPtr
Since we're taking from the committed set of pages, there should never
be a reason for this call to fail.

Also add a Badge to disallow taking committed pages from anywhere but
the Region class.
2021-07-22 14:20:05 +02:00
Andreas Kling
5217875f6a Kernel: Consolidate API for creating AnonymousVMObject with given pages
We don't need to have a dedicated API for creating a VMObject with a
single page, the multi-page API option works in all cases.

Also make the API take a Span<NonnullRefPtr<PhysicalPage>> instead of
a NonnullRefPtrVector<PhysicalPage>.
2021-07-22 09:17:02 +02:00
Andreas Kling
9e15708aa0 Kernel: Convert VMObject & subclasses to east-const style 2021-07-22 09:17:02 +02:00
Gunnar Beutner
b4272d731f Kernel: Make sure crash dumps are properly aligned on x86_64 2021-07-22 08:57:01 +02:00
Gunnar Beutner
36e36507d5 Everywhere: Prefer using {:#x} over 0x{:x}
We have a dedicated format specifier which adds the "0x" prefix, so
let's use that instead of adding it manually.
2021-07-22 08:57:01 +02:00
Gunnar Beutner
31f30e732a Everywhere: Prefix hexadecimal numbers with 0x
Depending on the values it might be difficult to figure out whether a
value is decimal or hexadecimal. So let's make this more obvious. Also
this allows copying and pasting those numbers into GNOME calculator and
probably also other apps which auto-detect the base.
2021-07-22 08:57:01 +02:00
Gunnar Beutner
be795d5812 Prekernel: Use physical addresses for some of the BootInfo parameters
The kernel would just turn those virtual addresses into physical
addresses later on, so let's just use physical addresses right from the
start.
2021-07-20 15:12:19 +02:00
Gunnar Beutner
dd42093b93 Kernel: Move boot info declarations to a header file
Instead of manually redeclaring those variables in various files this
now adds a header file for them.
2021-07-20 15:12:19 +02:00
Brian Gianforcaro
85e95105c6 Kernel: Mark read only RegisterState function parameters as const 2021-07-20 03:21:14 +02:00
Brian Gianforcaro
27e1120dff Kernel: Move syscall precondition validates to MM
Move these to MM to simplify the flow of the syscall handler.

While here, also make sure we hold the process space lock for
the duration of the validation to avoid potential issues where
another thread attempts to modify the process space during the
validation. This will allow us to move the validation out of the
big process lock scope in a future change.

Additionally utilize the new no_lock variants of functions to avoid
unnecessary recursive process space spinlock acquisitions.
2021-07-20 03:21:14 +02:00
Brian Gianforcaro
308396bca1 Kernel: No lock validate_user_stack variant, switch to Space as argument
The entire process is not needed, just require the user to pass in the
Space. Also provide no_lock variant to use when you already have the
VM/Space lock acquired, to avoid unnecessary recursive spinlock
acquisitions.
2021-07-20 03:21:14 +02:00
Gunnar Beutner
a364f5c7b7 Kernel: Make sure super pages are in the first 16MiB of physical memory
This was broken by recent changes.
2021-07-19 11:29:09 +02:00
Gunnar Beutner
5ff1416076 Kernel: Rename bootloader to prekernel
There are a few occurrences of the old name that slipped through.
2021-07-18 22:08:03 +02:00
Gunnar Beutner
f56ee10a13 Kernel: Hand out 64-bit addresses to userspace 2021-07-18 17:31:13 +02:00
Gunnar Beutner
7e94b090fe Kernel: Introduce basic pre-kernel environment
This implements a simple bootloader that is capable of loading ELF64
kernel images. It does this by using QEMU/GRUB to load the kernel image
from disk and pass it to our bootloader as a Multiboot module.

The bootloader then parses the ELF image and sets it up appropriately.
The kernel's entry point is a C++ function with architecture-native
code.

Co-authored-by: Liav A <liavalb@gmail.com>
2021-07-18 17:31:13 +02:00
Gunnar Beutner
357ddd393e Kernel: Make allocate_randomized() work for 64-bit addresses
The odds of finding a suitable address in 1000 attempts were not in our
favor given the size of the 64-bit address space.
2021-07-18 17:31:13 +02:00
Gunnar Beutner
f0c4941beb Kernel: Fix 64-bit address truncation in MemoryManager::ensure_pte() 2021-07-18 17:31:13 +02:00
Gunnar Beutner
b708b23b13 Kernel: Add support for kernel addresses other than 3-4GB 2021-07-18 17:31:13 +02:00
Andreas Kling
9457d83986 Kernel: Rename Locker => MutexLocker 2021-07-18 01:53:04 +02:00
Andreas Kling
cee9528168 Kernel: Rename Lock to Mutex
Let's be explicit about what kind of lock this is meant to be.
2021-07-17 21:10:32 +02:00
Brian Gianforcaro
c0987453e6 Kernel: Remove double RedBlackTree lookup in VM/Space region removal
We should never request a regions removal that we don't currently
own. We currently assert this everywhere else by all callers.

Instead lets just push the assert down into the RedBlackTree removal
and assume that we will always successfully remove the region.
2021-07-17 16:22:59 +02:00
Brian Gianforcaro
d879709ec7 Kernel: Convert the PhysicalPage bool parameter to an enum 2021-07-17 18:38:28 +04:30
Brian Gianforcaro
eb282ad410 Kernel: Declare VM/RangeAllocator trivial destructor as default
This is a clang tidy recommendation.
2021-07-17 13:02:09 +02:00
Brian Gianforcaro
24bd664980 Kernel: Remove stale include from VM/RangeAllocator.cpp
This was left over after the latest big refactor of the VM subsystem.
2021-07-17 13:02:09 +02:00
Brian Gianforcaro
dbc77148c9 Kernel: Convert RangeAllocator VERIFY to proper error handling
If a user allocates above 0x0 and below the allowable usermode
virtual address space, we need to return error instead of asserting.

Fixes: #8484
2021-07-17 13:00:21 +02:00
Gunnar Beutner
cbdb488578 Kernel: Move end_of_kernel_image after the .ksyms section
Without this we won't be able to detect whether .ksyms overlaps the end
of the page table we set up for the kernel image.
2021-07-16 18:50:59 +02:00
Andreas Kling
15ad4a8fd6 Kernel: Convert RangeAllocator to using a RedBlackTree internally
This data structure is a much better fit for what is essentially a
sorted list of non-overlapping ranges.

Not using Vector means we no longer have to worry about Vector buffers
getting huge. Only nice & small allocations from now on.
2021-07-15 02:03:57 +02:00
Andreas Kling
b0d9b88c49 Kernel: Hoist VERIFY from a loop in RangeAllocator::allocate_specific() 2021-07-15 01:48:09 +02:00
Andreas Kling
7ff14fecba Kernel: Remove unnecessary locking in RangeAllocator::contains()
The total range managed by a RangeAllocator doesn't change, so there's
no need to take a spinlock while comparing against it.
2021-07-15 01:48:09 +02:00
Andreas Kling
d4c73daacb Kernel: Convert RangeAllocator to east-const style 2021-07-15 01:48:09 +02:00
Idan Horowitz
be475cd6a8 Kernel: Handle OOM when adding memory regions to Spaces :^) 2021-07-15 00:49:41 +02:00
Andreas Kling
dc26c02379 Kernel: Convert MemoryManager to east-const style 2021-07-14 13:31:21 +02:00
Andreas Kling
c42807e3dc Kernel: Remove debug spam when PhysicalRegion::take_free_page() fails
We can have multiple PhysicalRegions (often the case when there is a
huge amount of RAM) so we really shouldn't print a debug message any
time someone tries to allocate from one. They will move on to another
region anyway.
2021-07-14 01:37:31 +02:00
Andreas Kling
5c24d18923 Kernel: Fix logic error in PhysicalRegion::contains()
This was incorrectly returning true for the address one byte past the
end of the region.
2021-07-14 01:37:19 +02:00
Andreas Kling
6cc1247395 Kernel: Cut allocation size for physical buddy bitmaps in half
We were allocating twice as much memory as we needed for these bitmaps
due to a silly typo. Found by tomuta trying to boot with 24 GiB of RAM.
2021-07-13 23:47:49 +02:00
Andreas Kling
d8ff46594a Kernel: Re-add accidentally removed friendship
PageDirectory and MemoryManager need to remain friends, for now..
2021-07-13 23:21:22 +02:00
Andreas Kling
de4ba1f39b Kernel: Remove some friendships and make some classes non-copy/moveable 2021-07-13 23:19:00 +02:00
Andreas Kling
424afdd72b Kernel: Remove some unnecessary includes in VM/Physical* 2021-07-13 23:11:06 +02:00
Andreas Kling
0a21d421d9 Kernel: Print a summary of physical zones during boot
Let's not print out every single zone, since that gets very noisy on
machines with a lot of RAM. :^)
2021-07-13 23:08:45 +02:00
Andreas Kling
bf5e4326ac Kernel: Fix bogus address calculation in initialize_physical_pages()
We were incorrectly using sizeof(PhysicalPageEntry) for some address
calculations instead of sizeof(PageTableEntry).

It still worked correctly because they happen to be the same size.
2021-07-13 23:08:45 +02:00
Andreas Kling
e323942623 Kernel: Only loop through usable zones when allocating >1 physical page
We still have to loop here, since a zone can be "usable" while not
being able to satisfy a multi-page allocation request.
2021-07-13 23:08:45 +02:00
Andreas Kling
379bcd26e4 Kernel: Avoid O(n) loop over zones when allocating from PhysicalRegion
We now keep all the PhysicalZones on one of two intrusive lists within
the PhysicalRegion.

The "usable" list contains all zones that can be allocated from,
and the "full" list contains all zones with no free pages.
2021-07-13 23:08:45 +02:00
Andreas Kling
9ae067aa7f Kernel: Make PhysicalRegion eternally allocated 2021-07-13 22:40:25 +02:00
Andreas Kling
959ceb4424 Kernel: Remove PhysicalRegion::finalize_capacity()
There's no reason to delay calculating the capacity (total page count)
of each PhysicalRegion. Just do it in the constructor.
2021-07-13 22:40:25 +02:00
Andreas Kling
5171249540 Kernel: Simplify the way PhysicalRegions are constructed
Instead of creating a PhysicalRegion and then expanding it over and
over as we traverse the memory map on boot, we now compute the final
size of the contiguous physical range up front, and *then* create a
PhysicalRegion object.
2021-07-13 22:40:25 +02:00