Initializing a Buffer or a strlist_t with zero-valued bytes only works
by conincidence, but because it would be the correct way. In the code
path "missing `in' in for", that zero-filled Buffer is freed using
Buf_Destroy, which could have invoked undefined behavior.
* route: ensure IPv4LL routes come last in priority
* DHCP: fix many issues with extending the last lease
* privsep: don't read control group from config in privsep
* privsep: only the master process responds to signals
* privsep: use a socketpair for stderr/stdin rather than dupping /dev/null
* privsep: right limit stdin/stderr/stdout
* privsep: dumping a lease is now run in a sandbox
* options: check if kernel supports INET or INET6 before enabling default
* options: let clientid override a prior duid
* options: allow -1 to represent infinity for requested lease time
* dhcpcd: fix a crash initing a new interface after route overflow
slightly updated in status. x86-64 and sparc and mips64-eb
all appear to work well enough to complete or get well into
an atf run (still going.)
note first ICE segv on sh3
right now. new address-of-packed-member and format-overflow
warnings have new GCC_NO_ADDR_OF_PACKED_MEMBER amd
GCC_NO_FORMAT_OVERFLOW variables to remove these warnings.
apply to a bunch of the tree. mostly, these are real bugs that
should be fixed, but in many cases, only by removing the 'packed'
attribute from some structure that doesn't really need it. (i
looked at many different ones, and while perhaps 60-80% were
already properly aligned, it wasn't clear to me that the uses
were always coming from sane data vs network alignment, so it
doesn't seem safe to remove packed without careful research for
each affect struct.) clang already warned (and was not erroring)
for many of these cases, but gcc picked up dozens more.
- pmap(9) needs uvm/uvm_extern.h.
- x86/pmap.h is not usable on its own; it is only usable if included
via uvm/uvm_extern.h (-> uvm/uvm_pmap.h -> machine/pmap.h).
- Make nvmm.h and nvmm_internal.h standalone.
rillig