RDTSC readouts might not actually be monotonically increasing due to
desynchronization between CPU cores, so checked_sub was causing a kernel
panic.

This reverts commit 4626c74b.

This is much more diff- and by extension rebase-friendly.

I tested compiling the kernel with a more recent rustc, and this
triggered UB. Don't know for sure why there was an invalid entry passed
to the kernel, but it did.

mark the canaries mutable so they go into data and bss and not in rodata

A range of usable memory could be small enough not to have any MAX_ORDER sized frames.

Pretty straightforward changes. This commit tries to avoid making any non-trivial fixes.

bjorn3 authored 5 months ago and Jeremy Soller committed 5 months ago

This allows removing the -Csoft-float argument when building.

This avoids TLB stalls after context switching, that are very visible
from flamegraphs.

I broke this in !321

This would allow logd to write all received log messages to debug:
without causing an infinite loop when it at the same time reads all
messages in sys:log.

Diagonal as in

Thread A        Thread B
Send to B       Send to A
Wait for B      Wait for A.

I haven't reproduced this bug, but this should fix it, since the wait
steps now checks the local percpu block for pending TLB shootdowns onto
itself.

Rather than disabling it right before entering userspace. This allows
showing debug messages while the graphics subsystem hasn't initialized
yet.

This doesn't give any issues both for UEFI boot (with framebuffer
present) and U-Boot boot (without framebuffer present).