Apparently some keyboards can have led control on another interface from
the one where input is received. Now led mask is global to the usb
devices instead of per keyboard instance.
If we cannot determine current scancode set because either the keyboard
does not respond or responds with bogus value, assume scancode set 1 and
mark the scancode set as uncertain. If we receive the byte 0xF0 while
having the uncertain flag set, swap to scancode set 2. 0xF0 is unused in
scancode set 1 but indicates key release in scancode set 2, so it will
be sent after every key press.
If we cannot find PS/2 devices in the acpi namespace check 8042 bit in
FADT. Also if user explicitly specified PS/2 scancode set, assume that
we do have a PS/2 controller even when other detection mechanisms fail
This makes insertion to block list O(log n) instead of O(n) when
blocking with wake time. This is more heavy and currently slower than
the previous implementation, but that is just because we don't really
have too many threads running at any given time. This will be more
scalable in the future and even allows implementing priorities if I ever
wish so
Now modifying Scheduler.h or Thread.h doesnt trigger practically a full
kernel rebuild. This required moving Mutex and RWLock of of line but
that should be fine :^)
Before this I was writing audio data to kernel in ~20 ms chuncks. This
feels like way too many syscalls, so now we allow sending upto 1 second
per syscall :D
We now use the correct PIN interrupt from the interrupt line field
instead of writing to it, this had no effect :D
PIN interrupts are now mapped as shared and level triggered, this
properly allows PIN interrupt sharing with multiple devices
There is no need to pass the current mode as an argument as ACPI has
access to that info either way.
Only route INTx objects when using ACPI, they won't be used otherwise
When shrinking a mmap based allocation, we can just unmap the trailing
pages that will no longer be needed. When extending an allocation we try
to MMAP_FIXED_NOREPLACE the need extra data to avoid memcpy between old
and new allocation. I'm not sure if the extension logic is really worth
it but it is just 15 lines of code to avoid possibly very large memcpys.
There was a lot of stuff wrong with this driver and it was only getting
speeds of ~2.5 MB/s. Now 4K blocks get ~17 MB/s which is not good but
definitely better. I'm pretty sure the issue is qemu's emulation as 85%
of time is spent between sending the command and it to complete.
Large blocks are now supported too, qemu with block size 1M gets around
600 MB/s read speeds.
The old SpinLockAsMutex was pretty confusing first of all. It also was
causing the issue thats been around for maybe a year now which is the
only consistently happening kernel panic. I've been pretty confused
about this and finally figured out what was causing this.
The main issue was that we were accidentally enabling interrupts when
blocking a thread that passed SpinLockAsMutex from normally interrupt
enabled context. This led to receiving IPI for thread unblock while
we were actively blocking the thread. I'm very suprized this had't
caused any more serious issues than occasional kernel panics :^)
Update dtv[0] to the max loadad module when loading dynamic modules
Free uthread if stack allocation or mprotect fails on pthread_create
Don't unmap the currently used stack when detached thread exits
This supports most(?) options except for the separate fields. Should
work for simple use cases
Also merge is currently just a wrapper around sort so its not optimized
:D
Don't wake up threads when getting the next timer deadline.
If we are idling this can: remove thread from block list, get the
deadline for the next blocked thread, return to idle thread. This will
end up sleeping until the next wake up condition which in the worst
case would be rebalance, or indefinitely when not running SMP.
Make sure we always set the next timer deadline. I was not doing that if
the timer interrupt did not lead to a yield
Scheduler will now set a timer interrupt deadline for when it should get
interrupted next. This fixes a big issue I've had for a long time where
sleeping on an idle core would have to wait for the next periodic
interrupt to wake up. This could cause 1 ms sleeps to be close to 10 ms.
This currently only supports lapic timers and will still fallback to
periodic interrupts when running with PIC. I should add deadline support
to PIT/HPET but why would you run with APIC disabled ;)
Remove random template and just make qsort be a wrapper around qsort_r
Recurse only into the smaller partition. This removes the possibility of
O(n) recursion in the worst case
Use insertion sort for ranges shorter than 16 elements.
Split qsort_swap into qsort_swap_fixed and qsort_swap_general
- fixed one handles 1, 2, 4 and 8 byte elements which can be done directly
with mov instructions
- general copies 64 bytes at a time followed by a single copy for the
rest. the 64 byte chunk loop gets nicely optimized into sse
Bananymous