| Age | Commit message (Collapse) | Author |
|---|
|
|
|
syscalls.h shouldn't define a random struct etc
|
|
Those had a lot of repeating code, but I'm not sure if this is the
right change. Well, apart from making pipe_joinqueue more consistent.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
doesn't need to be a shell builtin now
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
~3x speedup in tests
|
|
|
|
|
|
|
|
|
|
|
|
/bin/tests with test_sleep disabled now runs almost 2x as fast
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
similar to /dev/{stdin,stdout,stderr} on Linux, except handled by the
libc instead of the kernel
because that's the simplest way
|
|
Everything other than %s and %x outputs a single char at once. The
speedup is easily visible when e.g. hexdumping.
|
|
$ iostress 32 512 0 > /vtty # before
512 calls, 0 bytes. avg 121133
$ iostress 32 512 0 > /vtty # after
512 calls, 0 bytes. avg 103540
103540/121133 = ~85%
I think the tiny bit of added complexity is worth it here.
|
|
|
|
|
|
|
