Fix broken NPF BPFJIT tests.
when compiling BPF byte code, bpf_jit_generate_code called by npf_bpf_compile
expects the size in array count not the raw byte size of the bpf instruction struct.
also, use npf's kernel prepare function to initialize the byte code args.
sparc: Stop exposing default KERNBASE to modules
Otherwise, on kernels with non-standard KERNBASE, modules reference
illegal addresses when using curcpu(), curlwp, and so on.
Unfortunately, this cannot be pulled-up to release branches, as
this changes module ABI.
Ride on 11.99.10 version bump.
struct proc: Move p_pgrp out of p_startcopy/endcopy region.
It cannot be safely copied with the current memcpy logic that logic
uses happens without holding proc_lock, so we have to copy it
separately anyway. Leave a comment emphasizing this aspect of the
locking rules so we don't have this bug again.
Looks like p_mqueue_cnt may need similar treatment.
As a result of this kernel ABI change, welcome to 11.99.10!
PR kern/60962: fork and spawn read p_pgrp without proc_lock
menu(3x): Constification and voidification of pointer arguments.
libmenu has several functions that take a non-const char * argument
and then (effectively) strdup(3) it. I've changed it so that these
functions take a const char * argument. Modern GCC objects to string
literals being passed to these functions otherwise.
In other cases, a userptr allows some data to be associated with a
data structure, but its type is char *. I've changed it to void *.
Both of these changes are consistent with similar changes in ncurses,
although are inconsistent with SVR4 libmenu versions. libmenu has
unfortunately never been standardized.
"that is fine" - blymn@
change failure and success path in testing bpf in npf
There are two tests here, standlone BPF byte code and JIT-compiled code
we don't need if/else block, we need an assertion to trigger
on any if there is unsuccessful compilation.
this avoids tests failures being unnoticed because we only see failures when
running in verbose mode
atto(1): Sprinkle __dead on functions that never return.
Pacifies -Werror=missing-noreturn in clang:
/home/source/ab/HEAD-llvm/src/usr.bin/atto/main.c:24:1: error: function 'usage' could be declared with attribute 'noreturn' [-Werror,-Wmissing-noreturn]
{
^
/home/source/ab/HEAD-llvm/src/usr.bin/atto/main.c:134:1: error: function 'fatal' could be declared with attribute 'noreturn' [-Werror,-Wmissing-noreturn]
{
^
2 errors generated.
aarch64: no need to DSB after every DC instruction in ln_dcache_*_all
The ln_dcache_{wb,wbinv,inv}_all helper do not need to DSB after each
DC instruction. Instead DSB after completing the cache level.
FUNCTION_PROFILER(): On 68010, use the same two step profile counter
reference strategy as ColdFire does. This avoids overflowing R_68K_PC16
relocations for pretty much every reasonably sized program build for
profiling.
m68k/conf/Makefile.kern.m68k.inc now sets -fno-omit-frame-pointer
automagically when it's needed, so don't bother doing it in
individual kernel configs.
