- 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.
this commit adds the facilities to extract eh_frame[_header]
sections and pass it to each process.
A bump is required to `std` to extract the passed parameters.
it works with e.g. shellchat, but not vault.
still scratching my head on what's gone wrong -- it's
crashing in the pre-boot, which means something is not
mapped right for some case that wasn't caught in the
earlier testing.
Atomic operations were previously broken. RISC-V breaks atomic
operations into multiple instructions, and there is some state kept
between those instructions. This state is not visible, and is internal
to the processor.
We were ignoring this state, which resulted in inconsistencies
throughout the system. The following sequence was possible:
1. A thread loads a value, and obtains the `reservation set`
2. A context switch happens on the following instruction
3. A different thread loads the value, modifies it, and continues
4. A third thread (or a different process) loads a value and obtains the
`reservation set`
5. The original thread runs, sees the `reservation set` is held, and
updates the value, clobbering what the second thread did in #3
A workaround for this is to perform a dummy `sc` as part of the context
switch. This invalidates any atomic operations that may be pending.
Signed-off-by: Sean Cross <sean@xobs.io>
The previous patch didn't properly restore thread IDs when switching
contexts -- there were spots where we missed things, and spots where we
were double-borrowing.
Fix these corner cases.
Signed-off-by: Sean Cross <sean@xobs.io>
Note the current thread in the original process when its quantum
expires. This will allow us to clean up after ourselves when a context
switch occurs, and will help to prevent thread starvation.
Signed-off-by: Sean Cross <sean@xobs.io>
Currently, returning a Message (e.g. a Scalar1, Scalar2, or
MemoryMessage) always results in a context switch from the server to the
client. This is what makes these syscalls cheaper.
However, this results in process starvation. For example, if three
processes send a message to a server, then the server will only be able
to respond to one of those messages at a time. This works well with a
lightly-loaded system, but quickly falls apart when many threads are
running.
Keep track of which process and thread were originally switched to as
part of the `SwitchTo` call. If the Server is responding to a message
and owns the current quantum, don't immediately switch to the Client
thread.
Signed-off-by: Sean Cross <sean@xobs.io>
For some reason, the current thread becomes 0 during suspend or resume.
This causes an error during boot.
Patch this up while we figure out what's causing it.
Signed-off-by: Sean Cross <sean@xobs.io>
Unify scalar return calls so that we only have one return path rather
than three.
We can almost unify memory returning as well, but it's not quite there.
Signed-off-by: Sean Cross <sean@xobs.io>
There were a few corner cases when it came to thread selection. Namely,
the same thread had a higher chance of repeatedly being executed.
Clean up these corner cases, and as a result get libstd tests to pass.
Signed-off-by: Sean Cross <sean@xobs.io>
Add notes on how to properly hand memory to userspace. This does not yet
work, but is committed here as a work-in-progress.
Signed-off-by: Sean Cross <sean@xobs.io>
The default mem size got shrunk down to 4 MB earlier, but there is a
lingering issue that prevents specifying virtual addresses. As a result,
pddb and others don't work.
Temporarily revert this change.
Signed-off-by: Sean Cross <sean@xobs.io>
The process search now uses iterators properly. It's still not ideal
because it needs to investigate every process, but it's better than it
was.
Signed-off-by: Sean Cross <sean@xobs.io>
Use a multiply-based algorithm to count the number of leading zeroes.
This results in a function that's just 20 instructions long.
Signed-off-by: Sean Cross <sean@xobs.io>
THe act of finding a runnable thread is so common, yet the kernel
contains three ways of doing this.
Use a unified function to select the next available thread.
Signed-off-by: Sean Cross <sean@xobs.io>
Previously the `current_thread` was set to 0, which is invalid. Set the
`current_thread` so that it defaults to the initial thread ID.
Signed-off-by: Sean Cross <sean@xobs.io>
This memory region is much smaller than the previous value of 256 MB,
but it will limit the amount of memory used by superpages in
long-running processes.
Previously, we would march through memory as messages were passed back
and forth and a process' memory space would quickly be eaten up by 1024
pages.
Signed-off-by: Sean Cross <sean@xobs.io>
When waiting on a thread, it may be the case that the thread we're
joining has exited already. In this case, the correct approach is to
resume execution rather than switching to the parent process.
This fixes an issue where `SWITCHTO_CALLER` is None when joining a
short-lived thread.
Signed-off-by: Sean Cross <sean@xobs.io>
When attempting to return an invalid item, return `DoubleFree`. This is
because such cases are most likely due to instances where a message was
responded to twice.
Signed-off-by: Sean Cross <sean@xobs.io>
- we don't need a general memset call -- just a zero page call
- use native rust semantics to embody the call. some alignment checking is
done on the arguments on the way in and out that seem worth it
- the unrolled loop doesn't pay off because we're at the limit of
write bandwidth on the CPU at this point. we would need write-back
caches to gain more performance through unrolling.
@xobs found a scheduler bug that would trigger a runnable process
search every time a scalar message returned. This fixes the problem,
causing the quantum to go back to the caller as intended.
Improves performance on the syscall by about 40%.
Fix a segfault that occurred when using ReplyAndReceiveNext on hosted
mode. This now properly passes memory messages back to the client.
Signed-off-by: Sean Cross <sean@xobs.io>
Implement `ReplyAndReceiveNext` as a syscall. This will reply to a
message, pull the next message out of the queue, and then re-queue the
server to run. This should make servers much more responsive.
This patch has been tested on both hosted and native modes.
Signed-off-by: Sean Cross <sean@xobs.io>
I've added a `platform` module to store all platform specific code, (not
architecture specific) as some platforms will require custom drivers, e.g.
for debug output and for the random number generation code. The platform
code is in charge of mapping any peripherals needed to work, also to
initialize any other parts of the kernel that require drivers, for example
the debug shell.
Signed-off-by: Jean-Pierre De Jesus DIAZ <me@jeandudey.tech>
The rationale is that xous-rs should have only architecture
specific code, e.g.: riscv, arm, x86, etc. To make other
crates that depend on it not specify the features.
This could be useful when using xous-rs with `libstd`
for example.
Signed-off-by: Jean-Pierre De Jesus DIAZ <me@jeandudey.tech>
This allows for a more reliable detection of target_os configuration
in build scripts. The target is inconsistent between some architectures,
notably armv7a is `armv7a-none-eabi` skipping the `unknown` (the vendor,
I guess) part. This happen with some of the unrelated Cortex-M targets
present, and others may also be.
Signed-off-by: Jean-Pierre De Jesus DIAZ <me@jeandudey.tech>
- strip out gdb-stub. Per xobs, the feature never worked, and it
depends upon a branch of a git repo which is not allowed by crates.io
Since the feature hasn't been developed, and crates.io will not
allow even an optional feature that points to a git repo, the feature
was stripped out.
- rename "kernel" to "xous-kernel", so that we can live in a crates.io
package namespace that is specific to Xous.
still have to figure out the riscv package issue.
this will help speed up performance profiling as we can
resolve the physical address of the performance buffer allocated
in RAM and use a USB command to read it out, instead of streaming
it over serial.
Instead of passing a constructed mutable ref from the pagetable entry,
pass a raw pointer and use `read_volatile()` / `write_volatile(x)` on
this pointer.
This should remove a little bit of UB from the kernel.
Signed-off-by: Sean Cross <sean@xobs.io>
Previously, the kernel would dutifully try to map memory that was in the
kernel space. It did this by investigating the kernel memory offset,
which immediately causes a double-fault.
Check to make sure a userspace process is not trying to map kernel
memory before checking the valid page.
Signed-off-by: Sean Cross <sean@xobs.io>
Implement the syscall to adjust a process' limit. Currently this
only supports adjusting the heap maximum and the heap size.
In order to adjust a limit, you must first present the current limit.
This can be obtained by requesting a bogus value. The current limit is
returned as part of the result.
Signed-off-by: Sean Cross <sean@xobs.io>
Return the heap size and offset when decreasing the heap. Additionally,
return the heap size when increasing the heap by 0 bytes.
Signed-off-by: Sean Cross <sean@xobs.io>
this contains a dependency to `failure` two levels down that has a critical
vulnerability. Unfortunately no new versions of debug-here have been
published in a while. It's no longer used, so just stripping it out
altogether.
Hosted mode spawns processes very differently, and this difference was
not reflected after changes were made to native mode.
Signed-off-by: Sean Cross <sean@xobs.io>
Unmap the pages from the child process after we're done setting them up.
This ensures they're not still aliased by our process, despite being
marked as "owned" by the other process.
Signed-off-by: Sean Cross <sean@xobs.io>
Implement a functioning CreateProcess syscall on riscv. This call
creates a server in the child process, and connects the server to the
parent. The parent can communicate with the child process using this
server.
The Server ID is passed as four `u32` values to the entrypoint function.
This can be reassembled into something that can then be passed to the
`ReceiveMessage` syscall.
Signed-off-by: Sean Cross <sean@xobs.io>
We now us `ProcessStartup` as the return type. This builds in tset
mode, hosted mode, and RISC-V mode.
Process Startup is still not implemented for RISC-V mode.
Signed-off-by: Sean Cross <sean@xobs.io>
Turns out we didn't end up needing this capability, so, commenting it
out because it is a shy too powerful and is like packing a live
mortar shell in a storage unit: eventually, someone is going to
get hurt, even though it seems mostly harmless at the moment.
Code just commented out in case we need this feature in the future,
but, the main reason it was asked for (the PDDB) was accomplished
without it.
This moves the kernel into the main monorepo, allowing it to share
packages with the rest of the system. This change will allow us to
reduce the number of rebuilds that take place, as well as reducing disk
usage.
Note that the loader is still an indepenent project, and the integration
of that into the monorepo is pending.
Signed-off-by: Sean Cross <sean@xobs.io>
Each Cargo.toml now includes a comment which remnids maintainers
of the dependency policy for Xous (that is, nail down every dependency
to the minor number).
Correctly handle transitions to and from exception handling.
Previously, the thread ID was incorrectly set such that the exception
handler became a valid thread. This has been adjusted so that the
exception handler is no longer considered a thread, and all other
threads will cease scheduling during this process.
Signed-off-by: Sean Cross <sean@xobs.io>
Memory is on-demand allocated, meaning it's entirely possible to
allocate and free memory without ever touching physical pages.
Previously, when we attempted to free memory, it would skip over pages
that were never backed by real memory.
With this patch, we continue to free the virtual page even if there is
no associated physical page.
Signed-off-by: Sean Cross <sean@xobs.io>
Memory is on-demand allocated, meaning it's entirely possible to
allocate and free memory without ever touching physical pages.
Previously, when we attempted to free memory, it would skip over pages
that were never backed by real memory.
With this patch, we continue to free the virtual page even if there is
no associated physical page.
Signed-off-by: Sean Cross <sean@xobs.io>
If a process is currently running, it needs to be placed into the
`Debug` state rather than the `Ready` state.
Additionally, the process that's being switched to must enter the
`Running` state.
Signed-off-by: Sean Cross <sean@xobs.io>
If a process fails for some reason (e.g. by performing a fault) and
there is no exception handler, either terminate the process or (if
gdbserver is enabled) have it suspend the process.
Signed-off-by: Sean Cross <sean@xobs.io>
Previously, we were creating a stack beginning at the initial page,
and then growing it upwards. The correct behavior is to begin it at
the top and grow it downwards.
In order to do this, subtract the size of the stack from the initial
pointer.
Signed-off-by: Sean Cross <sean@xobs.io>
This fixes the exception handler so that it can support user-defined
exceptions. Additionally, kernel mode errors are called out now.
Signed-off-by: Sean Cross <sean@xobs.io>
When the program counter is NULL, we can still get exceptions. This
happens when we jump to an invalid address, for example.
Signed-off-by: Sean Cross <sean@xobs.io>
This compresses the entirety of the memory range finder into a single
statement. As a result, duplicate code is removed and we can elide more
range checks.
Signed-off-by: Sean Cross <sean@xobs.io>
By vendoring this package, we can avoid causing a rebuild unnecessarily.
This will be removed when Rust 1.55 is released and we can bring the
`kernel` package back into the main workspace.
Signed-off-by: Sean Cross <sean@xobs.io>
This fixes the MemoryRange refactor where `.addr` and `.size` are now
private fields.
Previously it had only been tested on baremetal mode.
Signed-off-by: Sean Cross <sean@xobs.io>
Mark the private fields in a `MemoryRange` struct as being private. This
allows us to perform operations on the memory range armed with the
knowledge that the entire range is valid and unaliased.
Signed-off-by: Sean Cross <sean@xobs.io>
Since `MemoryRange::new()` is now unsafe, add an `unsafe {}` note when
calling this function on Hosted mode.
This fixes the build.
Signed-off-by: Sean Cross <sean@xobs.io>
This function allows for creating arbitrary memory ranges. Mark this
function as `unsafe` in order to warn users that this can be used to
create invalid memory regions.
Signed-off-by: Sean Cross <sean@xobs.io>
This allows us to support different printers in the kernel. For example,
this can be used to swap out the println!() function to print to GDB
when a debugger is attached.
Signed-off-by: Sean Cross <sean@xobs.io>
Ensure pages of memory can be written to when we write to them. This
lets us patch program code to, for example, insert breakpoints.
Signed-off-by: Sean Cross <sean@xobs.io>
When running in CI, cause processes to panic rather than waiting for a
debugger to attach. This will ensure test failures are caught sooner
rather than later.
Signed-off-by: Sean Cross <sean@xobs.io>
Switch to the message client when returning memory in Hosted mode. This
properly marks a process as ready-to-run, which was otherwise not
correctly done.
Signed-off-by: Sean Cross <sean@xobs.io>
This changes the `p` command to only print processes, and renames the
existing behavior to `P` which prints processes and threads.
Signed-off-by: Sean Cross <sean@xobs.io>
This enables processes to be "suspended" by going into a Debug state.
Messages will back up and processes will continue to be scheduled around
the suspended process, while the system will continue to operate.
Signed-off-by: Sean Cross <sean@xobs.io>
When a process is terminated, we disconnect it from all other processes.
We also need to free the slot in the server table.
Signed-off-by: Sean Cross <sean@xobs.io>
THe MMU flushing logic got disabled due to an experiment. Re-enable it
to ensure pages are consistent when shifting them across boundaries.
Signed-off-by: Sean Cross <sean@xobs.io>
This avoids using uninitialized arrays to manage the pagetables, and
instead uses raw pointers. This should avoid using potentially
uninitialized data.
Signed-off-by: Sean Cross <sean@xobs.io>
When the kernel is processing an interrupt, use the exception stack.
This ensures that the kernel interrupt handler can switch processes
without having its stack go away.
This is used for debug information.
Signed-off-by: Sean Cross <sean@xobs.io>
These calls can be used to print out thread information for debugging.
This augments the existing `print_thread()` call, which has been renamed
to `print_current_thread()`.
Signed-off-by: Sean Cross <sean@xobs.io>
Due to several errors, multi-page lending did not work properly.
To begin with, if a page was not allocated then it would not be able to
be sent. That is, if you allocated 8192 bytes but only ever accessed the
first 4096 bytes, then an error would be encountered when lending the
unaccessed upper page.
Furthermore, due to a striding error in accessing pointers, half-pages
were treated as full pages.
Signed-off-by: Sean Cross <sean@xobs.io>
This brings intra-process lending in-line with intra-process sending,
where memory is not remapped the pointer is simply reused.
We may change this at a later point in order to further enforce the
concept of lent-memory-is-inaccessible, so do not rely in this behavior.
Signed-off-by: Sean Cross <sean@xobs.io>
This implements a syscall to shut down a server. Servers may not be shut
down if they have outstanding messages that they have seen but not
responded to.
Signed-off-by: Sean Cross <sean@xobs.io>
This rewrites the entire project so that most runnable projects now use
`xous-ipc` instead of `xous::String` where possible.
This also removes `rkyv` from the list of dependencies in `xous-rs`.
Signed-off-by: Sean Cross <sean@xobs.io>
When moving memory between processes, update the owner of that page to
match the new PID. Otherwise, memory can never be released.
Signed-off-by: Sean Cross <sean@xobs.io>
The `current_thread` variable is used to determine which thread to
switch back to after a context switch occurs. Keep track of the thread
more closely.
Signed-off-by: Sean Cross <sean@xobs.io>
This will enable us to have more information about the result that is
returned when mutable memory is modified by a server.
Signed-off-by: Sean Cross <sean@xobs.io>
The `can_resume` variable doesn't change, and can be made a constant.
This slightly improves reasoning about what the function is doing.
Signed-off-by: Sean Cross <sean@xobs.io>
This reworks the server queue to add a "generation counter". This
counter ensures messages are removed from the queue in the order they
were added, even if they are jumbled up within the queue itself.
Signed-off-by: Sean Cross <sean@xobs.io>
Instead of queueing the client thread number inside the server, queue
the server's thread number inside the server.
This prevents many spooky messages that were occurring when server
queues got full.
Signed-off-by: Sean Cross <sean@xobs.io>