| Age | Commit message (Collapse) | Author |
|---|
|
+ Gives a very slight improvement to RPC speeds, but mostly cleans up
code a little bit.
|
|
+ Fairly considerable speedup, as we don't have to look up the rpc pte
every time separately, instead cacheing them. Adds an architecture
specific limitation to total rpc stack size, though.
|
|
+ Try to write arguments to memory as fast as possible to free up some
registers that the compiler can then play with.
Qemu runs ~1 700 000 rpc's per second, visionfive2 @1GHz (I think) ~800 000.
|
|
|
|
|
|
+ Unlikely to ever run into billions of threads, and this keeps in line
with stuff like Linux. Also allows C to relatively painlessly
implement hashmaps of pids, if neccessary.
|
|
+ Now 10M alloc/frees succeed, which totals more memory than the virtual
machine has, so no too obvious leaks are occuring. For future
debugging speed, changed 10M to 1M.
+ Also quick fix to rpcs, stacks are now assigned. Not entirely sure why
they worked before this, but good that I found it.
|
|
|
|
|
|
|
|
|
|
+ The secret is using gravestones. I'll have to write up full
documentation for the feature but essentially riscv lets us encode
whatever we want into page table entries, as long as they're not
active. We use this to encode highest user address that is not a zero,
in that all entries in the top level are either active or gravestones.
When an active entry is removed, it is either a gravestone (if there
are other active entries above it) or it starts a cascade of removing
entries that have been previously removed
Slight runtime overhead to page mapping, pretty major advantage in rpc
calls. Feature will need to be tested more thorougly, and the init
program is sort of a best scenario with just one top level userspace
page table entry active at a time, leading to incredibly fast context
switches.
Current implementation limits a process' max virtual memory to 248 GiB
(in Sv39), but I don't think the missing 8 GiB is that big of a deal.
|
|
+ Eliminated need for save/load_regs, now the register save area is
behind a pointer which _slightly_ increases overall syscall delay, but
speeds up ipc requests by a fair bit.
|
|
|
|
|
|
|
|
+ Not strictly done yet, but we can swap between two processes :D
|
|
+ Not actually working (at least fully), need to continue debugging when
I have time.
|
|
|
|
|
|
|
|
|
|
+ Slightly unsure if it should be a syscall, or if assign would be
enough. Also, which thread should run in a fork/spawn? For now, old
thread, though this might increase latency. Or add swap flag to these
calls?
|
|
|
|
|
|
+ Skip cow for now, maybe implement later but try to keep things simple.
|
|
|
|
|
|
+ Does not mean documentation is done, but it's a nice little
achievement nonetheless.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
+ Essentially, separate root process and remote procedure call handling.
I really should write down some documentation so I don't forget, but
essentially: All threads share a common process virtual memory, and
when a thread wants to make an rpc, it switches over to its own rpc
vmem space that is linked to the remote process.
|
|
|
|
+ Next step, start documenting contents of each file.
|
|
+ Closer to what it's for rather than what it is.
|
|
|
|
+ next up would probably be to start working on process/thread
relationships, largely as documented. Still not entirely sure about
how I want to handle fork, but I suppose I might figure it out at some
point.
|
|
|
|
|
|
|
|
|
|
+ At least initial in the sense that this seems like the smartest way to
handle things, that is.
|
|
|
|
|
|
|
