Fixed non-header aligned sizes
When allocating a size that isn't aligned to the Header, then we get a incorrect alignment.
To fix this, added additional alignment padding to the next free header.
Fixed adding strings to build options
Refactored tests for the scheduler and task
Revert "Refactored tests for the scheduler and task"
This reverts commit 2bf56a368bc18f2bd2d33c385e3672d07e4431d9.
Refactored tests for the scheduler and task
Task fmt
Task fmt again >:(
Ramdisk
Added NotOpened error for file read and write
Added vfs init to initialise the root node
Added the ramdisk.initrd file to grub
Update makeiso to copy the ramdisk to the modules folder
Add a ramdisk step to create a ramdisk to be leaded by grub and parsed by the kernel
Add test files for runtime tests of ramdisk
vfs.init => vfs.setRoot
Improved ramdisk step
Also spelling
Changed name for the initrd
Rename RamdiskFS => InitrdFS
Add deinit for initrd
Fixed VMM unmap
{}
The symbol name can have spaces.
Changed the parseName to use the newline parsing.
Also used str.len
Missed whitespace => newline
Add a nothing to run release safe tests
Fix TSS
Also change to .{} syntax where appropriate.
Added the SS segment
Fixed spelling
Refactoring GDT
Multitasking working for now
WIP scheduler
Refactored Bitmap a bit
WIP still
Task switching working
Handlers return the stack pointer that will be used to restore the tasks stack, normal handlers will return the same stack pointer it was called with where task switching will return the stack pointer of the next task and restore its state using the interrupt stub.
Initial scheduler done
Created a stage 2 init task
Change u32 to usize
Move Task to arch specific
WIP
WIP2
Removed esp from task, replaced with stack_pointer
Removed the debug logs
Fixed init task stack
Change pickNextTask to pointer manipulation
This allows less allocations so faster switching
Temporary enable interrupts for some runtime tests
PIT and RTC need interrupts enabled to run their runtime tests
Renamed schedule => pickNextTask, comptime bitmap for pids not task init
And some other stuff: No pub for the task anymore
Use the leak detector allocator
Fmt
Fix unit tests
And some other stuff :P
PR review
Moved Task out of arch and have the stack init in the arch file
Mocking clean up
Removed commented code
Renamed createTask to scheduleTask where the user will have to provide a task to schedule
Removed redundant pub in log runtime test
Removed global allocator for scheduler
Cleaner assembly in paging
Fmt
Added new Scheduler test mode
Added new test mode to CI
Removed one of the prints
Added doc comment, task test for i386
Removed test
WIP
Runtime tests work
Have a global set in one task and reacted to in another. Also test that local variables are preserved after a task switch.
Removed new lines
Increased line length
Move the allocation of the bool above the task creation
This will allow the user to build the OS with more fine grain, but there is a white list of supported build targets.
Closes#129
Updated the README to reflect the new build options
x86 => i386
Added the new testing to the OS files
Some spelling
INOUT => IN/OUT
Added some doc comments to log
Added the new runtime to the build + added the None test mode
Moved some stuff around
None test mode is the default to run/build the OS normally with no runtime tests.
Add the new runtime testing to the CI
Updated README and CI
Increased timeout
Print the log message
Spelling
Move runtime to test folder
Add new RT to tty
Add a log to use the unmapped memory to cause page fault in release
Ensure the integer overflow happens even in release builds
