- invoking the swapper context is different in an IRQ because
we don't want to modify the runnability states of the original
process
- didn't want to make a custom version of swap resume but
with the IRQ issue it made sense. In the process of doing this,
we can also optimize out a memory space transition, so it's
probably worth it even if we're not entirely going down the
right path to fix the bug of interest.
probably should have done this in smaller commit chunks but it
was hard to say if this was going in the right direction until now.
at the moment, system can boot and evict quite a bit of memory before
hitting a panic. The current state is that we panic on returning
from an interrupt handler that hit a swapped page.
This is a pretty decent point to check point everything. hopefully
this next arc will be a bit cleaner.
some thought needs to go into when and how we call oom_doom
but let's test more core functionality of the swapper before we
worry about that. I can think of several places where we could
use that call, the main tricky bit is that oom_doom diverges
and so wherever we patch it has to be able to resume. If we put
it into e.g. somewhere in the main memory alloc loop, the trouble
is we would have to track all of the multiple places that we
could have arrived there...ideally we can find a single spot
to patch this, with a clean entry and exit path.
instead of sending scalar messages on every page alloc/free, which
is a lot of traffic, introduce an API where the swapper can request
a snapshot of the alloc status by mapping the kernel's alloc
table as a read-only structure into the swapper's memory space,
and then invoking the swapper in its interrupt context to read that out.
The structure is unmapped on return from the context.
additionally, a running tally of free space is now maintained
and this can be queried with a fast syscall.
Thus, the normal workflow is the swapper would have a timer that
periodically queries the kernel's free space level, and when it
falls below a critical amount, it will invoke the expensive snapshot.
this can also be extended so that the kernel can call into
the swapper's interrupt context with a "help, OOM imminent"
message down the road.
I think that structurally we might never exceed a nesting depth
of 2, but there are some specific cases where an alloc can call
an alloc in the kernel (e.g., having to create a new superpage
in the midst of a routine alloc), therefore, we have to handle
some nesting.
There's also chance that maybe nesting could be due to race
conditions or self-circular behavior in the swapper, but ideally
we avoid that entirely because we can't handle non-structural,
unbounded nesting of re-entry to the swapper routine.
We want to use the finish_callback_and_resume() function because
it's well-tested and sets up everything just right, but it expects
that we're coming from another process. Unfortunately, in the
middle of all of this, we need to map a page back into the
destination process' memory space, so we end up starting from
the swapper's memory space, changing into the destination process'
memory space, then mapping the page, then changing back into
the swapper's memory space, then using the finish_callback_and_resume()
call.
I think the impact might not be *that* bad, mainly it's a TLB
miss and the PTW isn't super bad, and this comes after a disastrously
expensive decrypt-from-RAM operation -- so probably not worth it
to angonize over optimizing this bit at this point, and stick
with well-tested primitives.
There's two ways we can enter the swapper:
1. via an interrupt due to a page fault
2. via a syscall that generated a memory event (e.g. unmap)
This patch handles case (2) better and clarifies variable names
around that to make the code read better.
It also has logic to handle divergent looping through a large
unmap request.
The unmap syscall is a little bit tricky, because it can de-allocate
several pages at once, which can trigger an advisory de-allocate
message to the swapper in the middle of the loop. The entire structure
has to be re-written to handle this case, so we patch this right at
the top level syscall point and handle it entirely with custom
logic in the swapper.
@xobs made the `state` element of a process publicly read-only visible
through a function call. I don't think this is a security issue but
please lmk if that's a problem.
for now, we just count how mayn pages we missed an advisory on,
those all effectively are "wired". If we're fast enough on registering,
it turns out we actually miss nothing.
DUe to https://github.com/rust-lang/rust/issues/114447 direct access to
static mut variables causes a warning which will be turned into an
error in the 2024 edition.
The workaround is to use `addr_of!()` and `addr_of_mut!()` macros and
then dereference that.
Signed-off-by: Sean Cross <sean@xobs.io>
With Rust 1.76.0, we no longer support building with `#![no_std]` or
`panic = "abort"`, both of which are required within the kernel. As a
workaround, go back to building the kernel under
`riscv32imac-unknown-none-elf`.
Signed-off-by: Sean Cross <sean@xobs.io>
Support passing process arguments in register 2 as part of arguments to
a process.
This requires updating the `InitialProcess` struct that is received from
the loader, and so requires modeifications to be made in tandem.
Signed-off-by: Sean Cross <sean@xobs.io>
these are just sync'd to the asm.S and checked in.
Eventually, we should migrate that .S file into using
in-line assembly macros, but now that I look at it, I remember
why it hasn't been done yet -- there are some fancy conditionals
used to enable compatibility with different architectures that
I don't know off the top of my head how to translate into
in-line assembly syntax.
a lot of re-import warnings have cropped up. While I appreciate
their studiousness on this, the problem is that removing some
of these exports in the API crate could mean that future users
who want access to some API features that aren't used in this
configuration are now unaccessible.
I think it's not a huge problem at the moment, just spam-cleaning
the warnings because there are too many to deal with carefully
at the moment.
When enumerating processes, the `find_process()` process didn't consult
whether a process was actually free. This resulted in a crash,
particularly when attaching a debugger to a crashed process.
Signed-off-by: Sean Cross <sean@xobs.io>
This command ensures memory is all owned by the correct process and is
unique in its ownership.
It takes a long time to run, so ensure the watchdog timer is disabled.
Signed-off-by: Sean Cross <sean@xobs.io>
Add a function to ensure mappings are unique across the whole system.
This takes a very long time and can crash if the watchdog timer isn't
paused.
Signed-off-by: Sean Cross <sean@xobs.io>
The phys_to_virt function can be used to determine what physical address
page is mapped to a virtual address page.
This throws an error when a page is mapped twice, which shouldn't
happen.
Signed-off-by: Sean Cross <sean@xobs.io>
The kernel stack offset should be aligned to 16 bytes. Previously it was
aligned to 8 bytes. It's unclear if this was the cause of any issues,
but it is best practice to keep things aligned to 16 bytes.
As part of this, key the atomic instructions off of the CPP macro
presence of atomic instructions.
Signed-off-by: Sean Cross <sean@xobs.io>
Fix memory zeroing when requesting a physical pointer. Previously, only
the first 25% of a page was zeroed due to a bug when casting between
*mut u8 and *mut usize.
This now uses `write_volatile` to ensure the memory clear isn't
optimized out, and uses `usize` pointers everywhere.
Signed-off-by: Sean Cross <sean@xobs.io>
The register set should be cleared already, however it is possible
(though remote) that register values are not cleared when a thread is
created.
Ordinarily threads are zeroed when destroyed, but this makes sure
they're zeroed upon creation as well.
Signed-off-by: Sean Cross <sean@xobs.io>
While it would be nice if we stopped the PLLs as we do on hardware,
Renode does not yet have the ability to issue the `wfi` instruction from
a peripheral write.
As such, directly call `wfi()` on Renode similar to how Cramium
operates.
Signed-off-by: Sean Cross <sean@xobs.io>
using the DUART means we don't get console input on the kernel
(because the DUART is incapable of this) but it's a decent compromise
I think because the DUART is fast and efficient, and mostly we want
it to capture kps?
Eventually we'll need an input mode but for now, this will get us
moving.
Mark `MemoryRange::as_slice()` and `MemoryRange::as_slice_mut()` as
unsafe.
This is due to the fact that not all representable types can be cast to
an arbitrary value. For example, if Process A sends random values to
Process B, then Process B casts the memory to a slice of `bool`, then
any values that are not 0 or 1 are undefined.
Because these functions can leak UB into the program, they are marked as
unsafe.
This patch adds `unsafe {}` blocks to all instances where this function
is called. Additionally, all instances have been checked to verify that
they do not cast to any invalid types -- most use `u8` or `u32`.
Additionally, the long-deprecated `MemoryRange::from_parts()` call was
removed, since that causes UB without being marked as unsafe.
Signed-off-by: Sean Cross <sean@xobs.io>
With Rust 1.71.0, mut detection has changed slightly, and it is more
aggressive about warning when things do not need to be mutable.
Remove these from the `services` file.
Signed-off-by: Sean Cross <sean@xobs.io>
SUpport debugging when interrupt handlers go awry, such as when they
blow out their own stack.
Do this by creating a dummy sentinal of 255 for interrupt handlers.
Signed-off-by: Sean Cross <sean@xobs.io>
When an interrupt handler panics, restore the system and allow it to
continue. This will help to debug IRQ handlers that panic for one reason
or another, such as infinite recursion.
Signed-off-by: Sean Cross <sean@xobs.io>
This changes the process.inner.pid to the correct PID. Before, it was
simply 1. This lead to problems, for example when sending messages
where an assert checks to make sure that the current PID and the
process PID are the same. Now, you can send messages from a new process.
- make the "base" field no longer public. It is accessible, but only
through an `unsafe` method, because, it is unsafe.
- add accessors to offest/mask fields. Because, it's not at all
unsafe to access these, and it can be useful for bit-ops that are
not part of the generic set of UTRA.
- add Arc to AtomicCSR, and require `std` for AtomicCSR types.
This should not break anything, I checked inside @foundation-devices
keyOS and they don't seem to be using AtomicCSR anywhere.
The Arc makes it a bit safer to clone the AtomicCSR base pointer
and share it between threads, but it will have an impact on performance.
The USB stack (which requires AtomicCSR) was regression testing and the
impact seems minimal.
Recall that the main purpose of AtomicCSR is to provide a CSR accessor
that does not require a mutable reference to the base object. This
can make certain Rust idioms a bit cleaner. We wrap the base accessor
in Arc+Atomic to try and improve the safety of throwing the borrow
checker under the bus, but it's still not perfect because hardware
mutable state can have side-effects within hardware itself that cannot
be modeled or checked by Rust's borrow checker.
Instead of always doing bytewise accesses, support 2- and 4-byte access.
Additionally, recognize when multiple 4-byte reads are required.
Note that an additional optimization may be made here in the future to
prevent having to check every address, but this is left for a future
optimization.
Signed-off-by: Sean Cross <sean@xobs.io>
Instead of remapping the page to allow the kernel to access a process'
memory, simply set the SUM bit. This makes it no longer an error for the
kernel to access userspace.
Keep setting the `W` bit when we poke, to allow for updating program
memory with breakpoints.
Signed-off-by: Sean Cross <sean@xobs.io>
Previously, we were creating a fake interrupt that was a "user software
interrupt". Update this to be a "supervisor external interrupt" to fix
hardware resume.
Signed-off-by: Sean Cross <sean@xobs.io>
Debug handling was not correct, and didn't work at all. Additionally,
interrupt masking didn't work.
Fix these problems.
Signed-off-by: Sean Cross <sean@xobs.io>
Implement the basics required to get "extended mode" working. This is
waiting on upstream to fix issues, but should allow us to attach to
processes once it works.
Signed-off-by: Sean Cross <sean@xobs.io>
When debugging an invalid process (e.g. not yet run, terminated, etc.)
return an error rather than silently continuing.
Signed-off-by: Sean Cross <sean@xobs.io>
Disable IRQs during destruction, but only if constructed with
`GdbUart::new()`. This allows characters to buffer when the GDB bridge
is destroyed while avoiding the double-free error we were getting from
creating a new UART in the IRQ handler.
Signed-off-by: Sean Cross <sean@xobs.io>
This adds a basic gdb server. Currently, it is possible to list
processes and inspect memory and threads. It is not yet possible to
actually interact with code (e.g. no breakpoints and no single
stepping).
Signed-off-by: Sean Cross <sean@xobs.io>
Add an initial reimplementation of the gdb-stub using the gdbstub
backing crate.
This supports process enumeration and basic thread investigation.
Processes will pause when the debugger is inspecting them. Memory and
registers can be investigated. Threads can be listed and selected.
It does not yet support breakpoints.
Signed-off-by: Sean Cross <sean@xobs.io>
Previously, `disable_irq()` would unconditionally mask an interrupt in
the `sim` register. However, when interrupts are disabled this register
is zero, and we really should be updating `SIM_BACKING`.
Consult the state of the interrupt handler and update the appropriate
field as necessary.
Signed-off-by: Sean Cross <sean@xobs.io>
This function never failed. Make it infallible so that it matches the
same calling convention as `enable_irq()`
Signed-off-by: Sean Cross <sean@xobs.io>
Even when the print-debug feature isn't turned on, println! macro is still able to output some critical non-debug info to the console such as panics, process crashes and violations, etc. Debugging messages are printed with klog! macro that's disabled when debug-print feature is not set.