| Age | Commit message (Collapse) | Author |
|---|
|
This allows addressing all cores on more modern processors. For now,
we still have a hardcoded limit of 64 due to s_processors being a
static array.
|
|
Some CPUID functions (e.g. 0xb) require input values in ecx.
|
|
Initializing the variable this way fixes a kernel panic in Clang where
the object was zero-initialized, so the `m_in_scheduler` contained the
wrong value. GCC got it right, but we're better off making this change,
as leaving uninitialized fields in constant-initialized objects can
cause other weird situations like this. Also, initializing only a single
field to a non-zero value isn't worth the cost of no longer fitting in
`.bss`.
Another two variables suffer from the same problem, even though their
values are supposed to be zero. Removing these causes the
`_GLOBAL_sub_I_` function to no longer be generated and the (not
handled) `.init_array` section to be omitted.
|
|
...to is_owned_by_current_processor(). As Tom pointed out, this is
much more accurate. :^)
|
|
Rename these API's to make it more clear what they are checking.
|
|
This avoids a race between getting the processor-specific SchedulerData
and accessing it. (Switching to a different CPU in that window means
that we're operating on the wrong SchedulerData.)
Co-authored-by: Tom <tomut@yahoo.com>
|
|
Co-authored-by: Tom <tomut@yahoo.com>
|
|
There is no value in exposing particular sizes for these files.
|
|
|
|
|
|
This function returns the same identical FPU state for all CPU's, so
there's no point requiring a Processor instance.
|
|
|
|
We can use ThreadRegisters::set_flags() to avoid the #ifdef's here.
|
|
|
|
And let id() be the non-static version that gives you the ID of a
Processor object.
|
|
This closes the race window between Processor::current() and a context
switch happening before in_irq().
|
|
Don't assume that a platform machine will provide at least the 32 bit
version of SMBIOS tables. If there's no SMBIOS tables, don't expose
directory entries in the /sys/bios/ directory.
|
|
This matches MutexLocker, and doesn't sound like it's a lock itself.
|
|
|
|
This has several benefits:
1) We no longer just blindly derefence a null pointer in various places
2) We will get nicer runtime error messages if the current process does
turn out to be null in the call location
3) GCC no longer complains about possible nullptr dereferences when
compiling without KUBSAN
|
|
Add some arch-specific getters and setters that allow us to merge blocks
that were previously specific to either ARCH(I386) or ARCH(X86_64).
|
|
Previously we allowed using immediate values here which is ambiguous
as to which specific mov instruction the compiler should choose.
|
|
Leave interrupts enabled so that we can still process IRQs. Critical
sections should only prevent preemption by another thread.
Co-authored-by: Tom <tomut@yahoo.com>
|
|
By making these functions static we close a window where we could get
preempted after calling Processor::current() and move to another
processor.
Co-authored-by: Tom <tomut@yahoo.com>
|
|
Processing SMP messages outside of non-SMP mode is a waste of time,
and now that we don't rely on the side effects of calling the message
processing function, let's stop calling it entirely. :^)
|
|
We were previously relying on a side effect of the critical section in
smp_process_pending_messages(): when exiting that section, it would
process any pending deferred calls.
Instead of relying on that, make the deferred invocations explicit by
calling deferred_call_execute_pending() in exit_trap().
This ensures that deferred calls get processed before entering the
scheduler at the end of exit_trap(). Since thread unblocking happens
via deferred calls, the threads don't have to wait until the next
scheduling opportunity when they could be ready *now*. :^)
This was the main reason Tom's SMP branch ran slowly in non-SMP mode.
|
|
|
|
Enter a critical section in Processor::exit_trap so that processing
SMP messages doesn't enable interrupts upon leaving. We need to delay
this until the end where we call into the Scheduler if exiting the
trap results in being outside of a critical section and irq handler.
Co-authored-by: Tom <tomut@yahoo.com>
|
|
We can't enter/leave SMP mode once the kernel is up and running.
|
|
- Use the receiver's per-CPU entry in the message, instead of the
sender's. (Using the sender's entry wasn't safe for broadcast
messages since the same entry ended up on multiple message queues.)
- Retry the CAS until it *succeeds* instead of *fails*. This closes a
race window, and also ensures a correct return value. The return value
is used by the caller to decide whether to broadcast an IPI.
This was the main reason smp=on was so slow. We had CPUs busy-waiting
until someone else triggered an IPI and moved things along.
- Add a CPU pause hint to the spin loop. :^)
|
|
Due to a boolean mistake in smp_return_to_pool(), we didn't retry
pushing the message onto the freelist after a failed attempt.
This caused the message pool to eventually become completely empty
after enough contentious access attempts.
This patch also adds a pause hint to the CPU in the failed attempt
code path.
|
|
|
|
On x86, the "pause" instruction is a "spin loop hint".
|
|
We commonly talk about "a process's address space" so let's nudge the
code towards matching how we talk about it. :^)
|
|
|
|
This directory isn't just about virtual memory, it's about all kinds
of memory management.
|
|
This patch also replaces the HashMap previously used to store coredump
properties with a plain AK::Array.
|
|
This isn't needed for Process / Thread as they only reference it
by pointer and it's already part of Kernel/Forward.h. So just include
it where the implementation needs to call it.
|
|
|
|
This was broken by the KASLR changes.
|
|
|
|
To add a new per-CPU data structure, add an ID for it to the
ProcessorSpecificDataID enum.
Then call ProcessorSpecific<T>::initialize() when you are ready to
construct the per-CPU data structure on the current CPU. It can then
be accessed via ProcessorSpecific<T>::get().
This patch replaces the existing hard-coded mechanisms for Scheduler
and MemoryManager per-CPU data structure.
|
|
This enables further work on implementing KASLR by adding relocation
support to the pre-kernel and updating the kernel to be less dependent
on specific virtual memory layouts.
|
|
This was mistakenly added in 306d898ee56c0d277d865dd4e3afba3d95eab9aa.
|
|
This makes it a lot easier to figure out what unmapped function is being
accessed, and a lot easier to reason about _why_ it is being accessed.
|
|
|
|
Let's allow passing 4 function arguments to a syscall. The 4th argument
goes into ESI or RSI.
|
|
This ensures that we can properly take the address of these symbols in
other code.
|
|
|
|
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.
|
