This creates more object files in the static library which reduces the size
of statically linked files as only the relevant object files are included.
In my experience, it reduced the size of the system initrd from 1.9 MiB to
1.6 MiB which is valuable.
Unfortunately this area is not standardized by POSIX. Linux uses an ioctl
which is not that bad, but I'd like to have a designated function. I'm not
sure if this facility is powerful enough and whether it should be improved.
Also note that I use a struct winsize as on Linux, but I use size_ts instead
for the heck of it. Perhaps I should use another name for the struct.
The console renderer now renders to a text buffer, which can be implemented
on any device, whether it's the VGA text buffer or a bitmap graphics device
with font rendering. This replaces the older code that could only render to
a VGA framebuffer and where the input parsing was tightly coupled with the
device rendering phase.
This enables useful features such as directories, CRC32 checksums, and other
useful features. The initrdfs in the kernel is now hooked up against the new
API, although the kernel's current limited FS support is a problem for now.
To work around that, directories are not supported at runtime, although the
internal API understands them wonderfully. This will be fixed when the
kernel gets a real VFS.
Programs were crashing randomly at startup on the kthread branch. After some
investigation, it turned out that the programs weren't correctly loaded by
the program loader in rare cases. Although, all investigation showed that
the program loader was correct and so was the interrupt routines (well,
almost, but nothing that could really trigger this). Yada yada, a few months
later I discovered that memcpy(3) was being corrupted by an interrupt handler
(which was correct). Turns out memcpy used stack space it hadn't allocated.
This is a Linux optimization that I had forgotten to disable with
-mno-red-zone in libmaxsi and thus interrupts just overwrote the stack of
optimized functions. Eek!
This provides control over the caching of memory, which makes write-combined
IO possible. Graphics drivers can use this to transfer data at a much higher
rate to the video memory.
The implementation is a bit hacky but it'll do for now. It provides enough
support for the experimental VBE driver to work on the real computers I
tested it on, even if the BIOS uses screwed up default MTRRs.
The virtual memory layer now automatically uses the PAT feature if available
but in a backwards compatible manner and otherwise just tries to approximate
PAT features if they are asked for.
