mirror of
https://git.FreeBSD.org/src.git
synced 2024-12-04 09:09:56 +00:00
d650c3efb6
The host_* syscalls are all raw Linux system calls, not the POSIX wrappers that glibc / musl create. So we have to ranage change the return value of host_llseek correctly to use the negative value hack that all Linux system calls use. This fixes a false positive error detection when we do something like lseek(fd, 0xf1234567, ...); This returns 0xf1234567, which is a negative value which used to trigger the error path. Instead, we check using the is_linux_error() and store the return value in a long. Translate that errno to a host errno and set the global errno to that and return -1. lseek can't otherwise return a negative number, since it's the offset after seeking into the file, which by definition is positive. This kept the 'read the UEFI memory map out of physical memory' from working on aarch64 (whose boot loader falls back to reading it since there are restrictive kernel options that can also prevent it), since the physical address the memory map was at on my platform was like 0xfa008018. Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D44286 |
||
|---|---|---|
| .. | ||
| include | ||
| kboot | ||
| libkboot | ||
| Makefile | ||
| Makefile.inc | ||
| README | ||
So to make a Linux initrd:
(1) mkdir .../initrd
(2) mkdir -p .../initrd/boot/defaults
(3) cd src/stand; make install DESTDIR=.../initrd
(4) Copy kernel to .../initrd/boot/kernel
(5) cd .../initrd
(6) cp boot/loader.kboot init
(7) find . | sort | cpio -o -H newc | gzip > /tmp/initrd.cpio
(8) download or build your linux kernel
(9) qemu-system-x86_64 -kernel ~/vmlinuz-5.19.0-051900-generic \
-initrd /tmp/initrd.cpio \
-m 256m -nographic \
-monitor telnet::4444,server,nowait -serial stdio \
-append "console=ttyS0"
(though you may need more than 256M of ram to actually boot FreeBSD and do
anything interesting with it and the serial console to stdio bit hasn't
been the most stable recipe lately).
Notes:
For #6 you might need to strip loader.kboot if you copy it directly and don't
use make install.
For #7 the sort is important, and you may need LC_ALL=C for its invocation
For #7 gzip is but one of many methods, but it's the simplest to do.
For #9, this means we can automate it using methods from
src/tools/boot/rootgen.sh when the time comes.
#9 also likely generalizes to other architectures
For #8, see https://kernel.ubuntu.com/~kernel-ppa/mainline/ to download
a kernel suitable for testing... For arm, I've been using the
non 64k page kernels and 5.19 seems to not suck.
aarch64:
qemu-system-aarch64 -m 1024 -cpu cortex-a57 -M virt \
-kernel ~/linuxboot/arm64/kernel/boot/vmlinuz-5.19.0-051900-generic \
-initrd ~/linuxboot/arm64/initrd.img -m 256m -nographic \
-monitor telnet::4444,server,nowait -serial stdio \
-append "console=ttyAMA0"
General
Add -g -G to have gdb stop and wait for the debugger. This is useful for
debugging the trampoline (hbreak will set a hardware break that's durable across
code changes). If you set the breakpoint for the trampoline and it never hits,
then there's likely no RAM there and you got the PA to load to wrong. When
debugging the trampiline and up to that, use gdb /boot/loader. When debugging
the kernel, use kernel.full to get all the debugging. hbreak panic() is useful
on the latter since you'll see the original panic, not the panic you get from
there not being an early console.