cy [Tue, 14 Mar 2017 02:10:37 +0000 (02:10 +0000)]
Disconnect ipftest and ipresend from the build until it can be verified
that they still work. These utilities have become out of sync with the
code in the kernel and need work to bring them back into shape.
Most people test on real systems or VMs on real networks.
cy [Tue, 14 Mar 2017 01:00:09 +0000 (01:00 +0000)]
Disconnect ipftest and ipresend from the build until it can be verified
that they still work. These utilities have become out of sync with the
code in the kernel and need work to bring them back into shape.
Most people test on real systems or VMs on real networks.
sbruno [Mon, 13 Mar 2017 22:53:06 +0000 (22:53 +0000)]
IFLIB updates
- unconditionally enable BUS_DMA on non-x86 architectures
- speed up rxd zeroing via customized function
- support out of order updates to rxd's
- add prefetching to hardware descriptor rings
- only prefetch on 10G or faster hardware
- add seperate tx queue intr function
- preliminary rework of NETMAP interfaces, WIP
Submitted by: Matt Macy <mmacy@nextbsd.org>
Sponsored by: Limelight Networks
ngie [Mon, 13 Mar 2017 18:01:01 +0000 (18:01 +0000)]
bsnmpd: fix segfault when trans_insert_port(..) is called with multiple
out of order addresses
Move `port->transport` initialization before the TAILQ_FOREACH(..) loop
to ensure that the value is properly initialized before it's inserted
into the TAILQ.
ngie [Mon, 13 Mar 2017 17:35:55 +0000 (17:35 +0000)]
lib/libcam/cam_cdbparse.3: fix manpage warnings
- Add comma before and after 'e.g.'; remove surrounding parentheses
that were unnecessary after this change [1].
- Add .Mt when referencing ken and Peter Dufault's email addresses
[2].
- Sprinkle around .An use where proper [2].
adrian [Mon, 13 Mar 2017 16:44:06 +0000 (16:44 +0000)]
[net80211] begin fleshing out per-VAP WME configurations.
As part of (eventual) p2p/tdls/multi-STA support, a lot of global configuration
parameters (WME, ERP (slot, preamble), channel width, HT protmode, etc are the
biggest offenders) should be per-VAP.
For multi-BSS VAP operation they may be linked, but for p2p/TDLS
operation that can be off-channel they can definitely be different
(think: 2GHz STA, 5GHz p2p.)
The global configuration view of these is mostly a "configure the current
non-smart-firmware NIC" view. This should be split up into per-VAP state,
then a global non-smart-firmware-NIC management layer to handle channel
changes and such in software.
This is step one in a loooong road for this. It should be a big non-functional
change for everyone.
* create a per-VAP WME update call.
* call it if its set, and call the global callback if it isn't
This still uses global WME state - it's just preparation for a future change
which will make the WME state per-VAP and fall through to a shared WME
configuration for hostap operation.
Note: this requires a full kernel recompile; the internal net80211 ABI has changed.
ngie [Mon, 13 Mar 2017 01:05:25 +0000 (01:05 +0000)]
Fix CFLAGS for including netipsec headers #includes
The netipsec headers are referenced via netipsec/..., not ./... .
Thus, assuming that the netipsec/... is nested under ${SRCTOP}/sys/netipsec
is wrong.
This tripped up some individuals building ^/head on systems pre-r314812.
imp [Sun, 12 Mar 2017 18:59:16 +0000 (18:59 +0000)]
Convert gnu to using SRCTOP
Prefer SRCTOP over CURDIR/../../contrib, etc. However, retain the "up
one level" instances of ../ because they are really relative to this
part of the tree and not a means to find the root of the tree. As
such, it's better to leave them since that further the goal of being
able to move directories if watned to in the future.
Differential Revision: https://reviews.freebsd.org/D9932
Sponsored by: Netflix
Silence On: arch@ (twice)
imp [Sun, 12 Mar 2017 18:59:00 +0000 (18:59 +0000)]
Convert include over to SRCTOP
Use SRCTOP in place of .CURDIR/.. as appropriate. The hand-crafted
relative paths for the "links" option remain, though, since those are
relative to /usr/include/sys/<blah> not to the source tree.
Differential Revision: https://reviews.freebsd.org/D9932
Sponsored by: Netflix
Silence On: arch@ (twice)
imp [Sun, 12 Mar 2017 18:00:02 +0000 (18:00 +0000)]
Export the actual LID state via sysctl. This allows more complex
policies about when to suspend based on other events happening AND the
lid being closed through the use of clever devd or other scripts.
pfg [Sun, 12 Mar 2017 16:03:34 +0000 (16:03 +0000)]
libc: provide some bounds-checking through reallocarray(3).
reallocarray(3) is a non portable extension that originated in OpenBSD.
Given that it is already in FreeBSD's libc it is useful for the cases
where reallocation involves a multiplication.
kib [Sun, 12 Mar 2017 13:58:51 +0000 (13:58 +0000)]
Avoid reusing p_ksi while it is on queue.
When sending SIGCHLD informing reaper that a zombie was reparented to
it, we might race with the situation where the previous parent still
not finished delivering SIGCHLD and having its p_ksi structure on the
signal queue. While on queue, the ksi should not be used for another
send.
Fix this by copying p_ksi into newly allocated ksi, which is directly
put onto reaper sigqueue. The later ensures that siginfo for reaper
SIGCHLD is always present, similar to guarantees for siginfo of child.
Reported by: bdrewery
Discussed with: jilles
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
kib [Sun, 12 Mar 2017 13:53:13 +0000 (13:53 +0000)]
Follow-up to r313690.
Fix two missed places where vm_object offset to index calculation
should use unsigned shift, to allow handling of full range of unsigned
offsets used to create device mappings.
Reported and tested by: royger (previous version)
Reviewed by: alc (previous version)
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
kib [Sun, 12 Mar 2017 13:51:13 +0000 (13:51 +0000)]
Accept linkers representation for ELF segments with zero on-disk length.
For such segments, GNU bfd linker writes knowingly incorrect value
into the the file offset field of the program header entry, with the
motivation that file should not be mapped for creation of this segment
at all.
Relax checks for the ELF structure validity when on-disk segment
length is zero, and explicitely set mapping length to zero for such
segments to avoid validating rounding arithmetic.
PR: 217610
Reported by: Robert Clausecker <fuz@fuz.su>
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
kib [Sun, 12 Mar 2017 13:48:24 +0000 (13:48 +0000)]
Ktracing kevent(2) calls with unusual arguments might leads to an
overly large allocation requests.
When ktrace-ing io, sys_kevent() allocates memory to copy the
requested changes and reported events. Allocations are sized by the
incoming syscall lengths arguments, which are user-controlled, and
might cause overflow in calculations or too large allocations.
Since io trace chunks are limited by ktr_geniosize, there is no sense
it even trying to satisfy unbounded allocations. Export ktr_geniosize
and clamp the buffers sizes in advance.
PR: 217435
Reported by: Tim Newsham <tim.newsham@nccgroup.trust>
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
tijl [Sun, 12 Mar 2017 10:56:19 +0000 (10:56 +0000)]
- Remove separate handling of /bin and /usr/bin in manpath. They are no
longer a special case.
- Prefer PREFIX/share/man over PREFIX/man.
- Add /usr/local/share/man to man_default_path.
- Update manpath man page.
kp [Sun, 12 Mar 2017 05:42:57 +0000 (05:42 +0000)]
pf: Fix incorrect rw_sleep() in pf_unload()
When we unload we don't hold the pf_rules_lock, so we cannot call rw_sleep()
with it, because it would release a lock we do not hold. There's no need for the
lock either, so we can just tsleep().
While here also make the same change in pf_purge_thread(), because it explicitly
takes the lock before rw_sleep() and then immediately releases it afterwards.
ngie [Sun, 12 Mar 2017 05:20:49 +0000 (05:20 +0000)]
lib/libcam/cam.3: fix manpage warnings
- spelling: "mis-named" should be "misnamed".
- delete spaces interspersed in literal representation of
`struct cam_device` as hard-tabs separate the types and fields.
- Add commas after `e.g.`.
araujo [Sun, 12 Mar 2017 04:15:56 +0000 (04:15 +0000)]
Add the capability to refresh the gpart(8) label without need a reboot.
gpart(8) has functionality to change the label of an GPT partition.
This functionality works like it should, however, after a label change
the /dev/gpt/ entries remain unchanged. glabel(8) status output remains
unchanged. The change only takes effect after a reboot.
ngie [Sun, 12 Mar 2017 03:58:54 +0000 (03:58 +0000)]
Add 3 more testcases demonstrating how install -l sr works
The additional testcases use absolute paths for sources and targets,
as the other testcase which tested `-l sr` used flat relative paths in
the same directory.
Please note that these testcases do not test `-l a` -- that's already
addressed in the battery of tests.
ngie [Sun, 12 Mar 2017 03:29:49 +0000 (03:29 +0000)]
Clarify src vs dest path mismatch in :symbolic_link_{absolute,relative}_body
Unfortunately kyua does not omit the path mismatch on failure, so it must be coded
into the error message.
Cache the values, run the test(1) call, then print out the values in an atf_fail
call to emit the required diagnostics to debug why things are failing.
pfg [Sun, 12 Mar 2017 03:26:25 +0000 (03:26 +0000)]
libc: small cleanups.
Unsign setlen: it is local and will never be negative. Having one more bit
for growth is beneficial and it avoids a cast when it's going to be used
for allocation.
pfg [Sun, 12 Mar 2017 03:22:18 +0000 (03:22 +0000)]
libc: small cleanups.
Rename nitems to numitems: it shares the anme with an existing macro in
sys/params.h. Also initialize the value later which avoids asigning the
value if we exit early.
ian [Sat, 11 Mar 2017 22:34:02 +0000 (22:34 +0000)]
Extend the pl011 small-fifos fix to other SoCs that indicate rev 5
hardware but lack the larger fifos rev 5 hardware should have.
The linux world (where our FDT data comes from) solved this by adding
a new property to pl011 nodes, "arm,primecell-periphid". When this
property is present, its values override the values in the hardware
periphid registers. For pl011 rev 5 hardware with small fifos, they
override the id so that it appears to be rev 4 hardware.
The driver now uses the new property when present. It also continues
to check the device compat string, to handle older fdt data that may
still be in use on existing systems (on RPi systems it is common to
update system software without updating fdt data which is part of the
boot firmware).
mav [Sat, 11 Mar 2017 18:46:46 +0000 (18:46 +0000)]
Increase device openings to tagged maximum.
Some SIMs report much less untagged device openings then tagged ones.
Target mode devices are not handled by regular probing routines, and so
there is nothing to increase queue size for them to the SIM's maximum.
To fix that resize the queue explicitly on ctl periph registration.
This radically improves performance of mpt(4) in target mode.
Also fetch and report device queue statistics in `ctladm dumpstructs`,
since regular way of `camcontrol tags` is not usable in target mode.
avg [Sat, 11 Mar 2017 16:43:38 +0000 (16:43 +0000)]
uma: fix pages <-> items conversions at several places
Those places were not taking into account uk_ppera.
At present one allocation is always used by one slab, so uk_ppera must
be used to convert between pages and slabs.
uk_ipers is used to convert between slabs and items.