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.
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.
Also move the spin wait to check for pending IPI sends to the start of
ipi sending function. There is no need to wait for IPI send succeeding
right after it. It is enough to make sure there are not multiple IPIs
being sent at the same time.
This was broken when I added SMP support. This patch makes sse kind of
dumb as it is saved and restored on every interrupt, but now it at least
works properly... I'll have to look into how sse can get optimized
nicely with SMP. Simple way would be pinning each thread to a specific
processor and doing pretty much what I had before, but sse thread saved
in processor rather than static global.
/dev/keyboard and /dev/mouse can be read for events from any attached
keyboard or mouse respectively. This makes device hot-plugging support
pretty much automatic for TTY, GUI, and whatever takes input.
This is kinda hacky, as I had disable the PS/2 initialization so that
usb keyboard gets /dev/keyboard0. I should add device hot plugging
support for TTY and GUI...
PS/2 code is now kind of messed up, but it works. Keyboards and mice are
now an abstract class that is automatically exposed to userspace. This
will make adding USB input much nicer.
realpath is implemented as a syscall. This is not really required but it
was the easiest way to get it working as there is already path
canonicalization at kernel level.
Now files are installed using the install() command instead of manually
copying files to their destinations. This allows automatic recompilation
of headers that did not work previously
These are pretty much dummy functions in the kernel side. Only case that
is handled is SOL_SOCKET with SO_ERROR. This is hard coded to return no
error. Network stack is currently synchronous, so all errors are already
reported through synchronous network functions.
Inode closing is something that needs a complete rework. Currently all
sockets are closed when close() is called, which leads to connection
closing if you fork()/exec() with socket being marked as CLOEXEC.
Inodes should probably only be closed once they are not referenced
anywhere.
When a unix domain socket is closed and it has a connection to another
socket, it will make the other socket readable and recv will return 0.
This allows detection of socket closing
These can allocate memory that can be shared between processes using
a global key. There is currenly no safety checks meaning anyone can
map any shared memory object just by trying to map every possible key.
When printing early return message, current time was read twice. This
could lead to early return check failing, but when printing and reading
the time again subtraction overflow would happen.
Old StorageDevice::read_sectors() read each sector separately if the
underlying disk had a disk cache. This patch allows multiple sectors to
be read even if the disk cache exists and contains some of the sectors.
Only sectors that could not be found from the disk cache are actually
read from the disk. This optimization is not done for writing, which
still will write each sector separately, if disk cache has no memory to
store new sectors. It would feel kind of unnecessary optimization as you
have greater problems if disk cache cannot allocate a single page.
Bananymous