Commit Graph

164 Commits

Author SHA1 Message Date
Bananymous 9bc02c81f8 Kernel: Map DMA and PCI MMIO as uncached 2024-07-30 11:10:08 +03:00
Bananymous 539afb329a Kernel: All processors use LAPIC timer when running with APIC
This makes scheduler preemption much cleaner as bsb does not have to
send smp messages to notify other processes about timer interrupt.

Also PIT percision is now "full" 0.8 us instead of 1 ms that I was using
before.
2024-07-23 02:28:52 +03:00
Bananymous f8261c60c0 Kernel: Rewrite the whole scheduler and re-architecture SMP handling
Change Semaphore -> ThreadBlocker
  This was not a semaphore, I just named it one because I didn't know
  what semaphore was. I have meant to change this sooner, but it was in
  no way urgent :D

Implement SMP events. Processors can now be sent SMP events through
IPIs. SMP events can be sent either to a single processor or broadcasted
to every processor.

PageTable::{map_page,map_range,unmap_page,unmap_range}() now send SMP
event to invalidate TLB caches for the changed pages.

Scheduler no longer uses a global run queue. Each processor has its own
scheduler that keeps track of the load on the processor. Once every
second schedulers do load balancing. Schedulers have no access to other
processors' schedulers, they just see approximate loads. If scheduler
decides that it has too much load, it will send a thread to another
processor through a SMP event.

Schedulers are currently run using the timer interrupt on BSB. This
should be not the case, and each processor should use its LAPIC timer
for interrupts. There is no reason to broadcast SMP event to all
processors when BSB gets timer interrupt.

Old scheduler only achieved 20% idle load on qemu. That was probably a
very inefficient implementation. This new scheduler seems to average
around 1% idle load. This is much closer to what I would expect. On my
own laptop idle load seems to be only around 0.5% on each processor.
2024-07-22 00:33:50 +03:00
Bananymous 0578d41500 Kernel: Implement WriteCombining memory
This makes framebuffer much faster on real hardware
2024-07-15 22:09:09 +03:00
Bananymous 2be4fe8404 Kernel: Make PageTable::s_fast_page_lock non-recursive
This lock is only used in wrapper of PageTable. There is no possiblity
of taking the lock outside of these wrappers.
2024-05-24 14:12:35 +03:00
Bananymous 061012a268 Kernel: Fix signal trampoline
32 bit did not even support it and 64 bit did not align stack
2024-04-18 13:32:40 +03:00
Bananymous e7ef7a9e55 Kernel: Implement barebones AML parser
This implements only parsing for AML in qemu. InvokeMethods are not
parsed since number of arguments to Methods is not yet known.

Parsing AML uses multiple kilobytes of stack space, so I increased
boot stack size by a lot :D

I am not sure where my own AML is going, but this is good start if
I decide to implement full ACPI on my own.

This code is very much just ugly macro expansion.

Qemu has 2 DefPackage elements that I am not able to parse. Package
data ends while there should be still multiple elements.
2024-04-07 17:03:30 +03:00
Bananymous f8c01418b1 Kernel: Fix multiprocessing on x86_64
I did not even start APs after initializing them... :D
2024-04-03 14:21:55 +03:00
Bananymous d2df55b1ac Kernel: Allow booting with multiple processors on i686
Also remove unnecessary ds clearing in x86_64.
2024-04-03 02:30:38 +03:00
Bananymous 0dd74e3c9d Kernel: Implement syscalls for i686 and cleanup x86_64
This actually allows i686 to boot properly!
2024-04-03 02:23:23 +03:00
Bananymous fe386fa819 Kernel: Implement thread start trampoline for userspace
This is needed on i686 to set segment registers.
2024-04-03 00:42:39 +03:00
Bananymous 2106a9e373 Kernel: Rework scheduler/processor stacks. 2024-04-02 12:34:42 +03:00
Bananymous 5050047cef Kernel: Rewrite whole scheduler
Current context saving was very hacky and dependant on compiler
behaviour that was not consistent. Now we always use iret for
context saving. This makes everything more clean.
2024-03-29 18:02:12 +02:00
Bananymous d7bf34ecd0 Kernel: Write isr handler for x86_32 and cleanup x86_64 2024-03-26 20:16:20 +02:00
Bananymous 1943c3e7a1 Kernel: Unify IDT and GDT code between x86_64 and x86_32
The code is pretty much the same, so there are just couple macros
differiating initialization.
2024-03-26 16:42:02 +02:00
Bananymous 2dd0bfdece Kernel: Make i386 thread tramplines crash 2024-03-26 00:10:42 +02:00
Bananymous fe17958b9f Kernel: Rename rsp->sp and rip->ip
This makes more sense if we support i386
2024-03-22 15:41:15 +02:00
Bananymous 45cea14165 Kernel: Move sys_fork trampolines to kernel/arch/ directory 2024-03-22 12:48:54 +02:00
Bananymous 42469b83fe Kernel: kernel panic is now sent to all processors 2024-03-15 13:45:01 +02:00
Bananymous e65bc040af Kernel: Now all active processors are used in scheduling
When a timer reschedule happens, ipi is broadcasted too all
processors for them to perform a reschedule!
2024-03-09 23:53:50 +02:00
Bananymous 89ca4c8a8b Kernel: Implement IPI broadcasting 2024-03-09 23:53:38 +02:00
Bananymous 55d2a64f54 Kernel: Map interrupt handlers for all processors
This doesn't mean that processors will actually handle the irqs
2024-03-09 23:50:57 +02:00
Bananymous 29fd682672 Kernel: Store current processor pointer in IA32_GS_BASE
This allows easier access to processors fields
2024-03-07 16:05:29 +02:00
Bananymous 54f64e7618 Kernel: Move current page table to Processor
APs can now map kernel page table and print current time!
2024-03-06 02:19:59 +02:00
Bananymous 76b0f80169 Kernel: Move IDT to Processor 2024-03-06 00:45:54 +02:00
Bananymous 58aca68726 Kernel: Move GDT to Processor 2024-03-06 00:35:45 +02:00
Bananymous 9c36d7c338 BAN/Kernel: Rework assertion/panic system
BAN/Assert.h does not need any includes meaning it can be included
anywhere without problems.
2024-03-04 11:41:54 +02:00
Bananymous 8141b9977d Kernel: Per processor information is now stored in class Processor
This allows us to allocate processor stacks, and other per processor
structures dynamically in runtime. Giving processor stack to
ap_trampoline feels super hacky, but it works for now.
2024-03-03 22:30:06 +02:00
Bananymous c035d3c82c Kernel: Start all processors on kernel boot
Processors don't do anything, except print hello message and halt.
2024-03-03 02:19:43 +02:00
Bananymous ec0cb5fd54 Kernel: Remove CriticalScopes from memory handing code 2024-02-29 19:16:41 +02:00
Bananymous d94f6388b7 Kernel: Fix all broken locks from new mutexes 2024-02-28 22:45:34 +02:00
Bananymous 9314528b9b Kernel: Improve syscall handling
Syscalls are now called from a list of function pointers
2024-02-12 21:51:11 +02:00
Bananymous 139bb5c2a5 Kernel: Fix linker script
I have no idea why rodata was executable :D
2024-01-30 17:32:50 +02:00
Bananymous dfe5a2d665 All: Cleanup all files
Add newline to end of files and remove whitespace from end of lines
2024-01-24 15:53:38 +02:00
Bananymous b0ff2392a1 Kernel: Add some helpful debug prints
Log RIP when stack pointer is out of bounds.

Log all syscalls that return ENOTSUP
2024-01-03 02:08:01 +02:00
Bananymous 001e95f973 Kernel: Optimize sse saving/loading
Only save and load sse when new thread is trying to execute sse
instruction. There is no need to do that every time we enter kernel.
2024-01-03 02:06:49 +02:00
Bananymous 4307968182 All: Start work again on sse support 2023-12-28 19:14:42 +02:00
Bananymous 284a012509 Kernel: Add framebuffer information to kernel image 2023-12-09 17:33:58 +02:00
Bananymous 1c78671078 Kernel: Rework all signal code
Signal handling code was way too complex. Now everything is
simplified and there is no need for ThreadBlockers.

Only complication that this patch includes is that blocking syscalls
have to manually be made interruptable by signal. There might be some
clever solution to combat this is make this happen automatically.
2023-12-06 13:02:17 +02:00
Bananymous 84040e64b8 Kernel: Don't use multiboot2 explicitly. Parse it to common structure
This allows support of multiple different bootloaders
2023-11-17 18:54:59 +02:00
Bananymous ca34ece8ef BAN: Update ASSERT api
its now much harder to mix < with <= and > with >=
2023-11-02 00:01:12 +02:00
Bananymous 71ecdd7245 Kernel: Canonicalize vaddr before using it 2023-10-30 19:20:17 +02:00
Bananymous 42772611ce Kernel: Add fast page to page table
Add "fast page" to KERNEL_OFFSET. This is always present in page
tables and only requires changing the page table entry to map. This
requires no interrupts since it should only be for very operations
like memcpy.

I used to map all temporary mappings to vaddr 0, but this is much
better. C++ standard always says that nullptr access is undefined
and this gets rid of it.

Fixed some bugs I found along the way
2023-10-30 19:02:09 +02:00
Bananymous 0c590821ed Kernel: only map kernel from g_kernel_start onwards 2023-10-30 16:22:02 +02:00
Bananymous 6ef153b6ca Kernel: Enable Write Protect. This seems to be good for security 2023-10-30 12:23:22 +02:00
Bananymous 130a69bac6 Kernel: Fix stack OOB detection
I now check both interrupt and normal stack to detect OOB. Processes
are killed if they encouner stack over/under flow.
2023-10-30 12:17:08 +02:00
Bananymous 0b5fcb3f88 Kernel/LibC: Add crt* files to LibC and remove crt0 from kernel
There was no reason for libc get crt0 from kernel.
2023-10-30 11:06:13 +02:00
Bananymous c17cdb39d5 Kernel: Map multiboot2 memory in PageTable initialization
It cannot be assumed that multiboot data lies between kernel_end
and 2 GiB mark, so I properly allocate virtual address space for it.
2023-10-17 01:15:08 +03:00
Bananymous 69a39b7077 Kernel: Start using multiboot2 instead of multiboot
This allows better compatibility with (U)EFI and gives RSDP location
instead of me having to scan ram to find it.
2023-10-17 01:06:24 +03:00
Bananymous 128202a55a Kernel: Cleanup GDT code 2023-10-13 16:18:22 +03:00