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.
Only segment 0 is supported, but devices can now be accessed through
mmio.
Adding more segments would require adding argument to every PCI API so
it is left for later.
This fixes a bug where userspace provided address is not fully mapped
and the kernel tries to read/write it while using PageTable fast page.
In the future userspace input should be copied on syscall entry, so
userspace could not modify the input during syscall. Currently there
is change that userspace input passes kernel syscall validation and
after that userspace could modify the input before the value is
actually used.
Remove race condition if two acks are to be sent one after another.
Always unblock semaphore once TCP thread has done something. This
allows better chance of TCP sending to succeed.
There are multiple places in the networking code that would require
thread-safe entering to blocking mode. I should add some API for this
so that a lot of race conditions could be removed.
I had written the ICR register check backwards which lead to interrupt
handling only when it was not needed, and no handling when it was
needed. This somehow still worked, just much slower often requiring tcp
resends from the server.
Remove tree-like structure from AML. This allows more spec compliant
parsing of named objects inside not yet declared devices.
This also allows AML to be run thread safely. All object adds/removes
are now guarded by a mutex.
This patch implements basic support for power button using ACPI
fixed events. I still need to implement general purpose events
and embedded controller for full power button support.
Now global lock uses the actual global lock. Currenly if no lock
can be acquired, we just panic the kernel so that I remember to
implement it properly once AML is running concurrently.
I had forgotten that Processors used to be a different definition
in AML.
I also implemented reads/writes for FieldElement/IndexFieldElement
that fit in 64 bits. Reads and writes to buffer are still a TODO.
Method invocation is starting to come together. This implemenetation
can interpret some of the qemu's functions to enter ACPI mode.
PCI config space access is currently the one thing is between
entering ACPI mode.
Now AML parsing is actually done while respecting namespaces and
scopes. I implemented the minimal functionality to parse qemu's AML.
Next step is to implement AML interpreting and then we can drop lai
as a dependency.
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.
This allows signals to be called even if the process does no syscalls
The old scheduler did signal handling but I feel like it should be
enough to handle them only after syscalls and IRQs. ISRs already
handle signals that caused the ISR and there is no other route to
kernel space.
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.
Bananymous