On 10/11/22 04:20, Dennis Clarke wrote:
On 10/10/22 17:53, Warner Losh wrote:
On Mon, Oct 10, 2022 at 10:56 AM Warner Losh <i...@bsdimp.com> wrote:
I know what's causing this problem. I'll resolve.
tl/dr: _pv_entry.h depends on sys/param.h being included before its use.
https://reviews.freebsd.org/D36927 fixes it by making sys/_pv_entry.h
more
self-contained and less reliant on param.h pollution. The kernel includes
that
everywhere it's used, but userland is more hit or miss because
machine/pmap.h
isn't a well defined interface, but is needed for some things sometimes.
I am not sure where this is related, however, there is a change in QEMU:
https://github.com/qemu/qemu/commit/a4a9a4432e2bf280a989ca344466d7375db7993f
Which seems to provide a way around some really long ISA cpu name data
that gets caught in sys/riscv/riscv/identcpu.c at around line 113 or
so :
Forgot to mention :
https://github.com/qemu/qemu/blob/master/hw/riscv/virt.c#L248
I wonder what the string is that tossed the KASSERT.
--
Dennis Clarke
RISC-V/SPARC/PPC/ARM/CISC
UNIX and Linux spoken
GreyBeard and suspenders optional
