248557 Commits

Author SHA1 Message Date
dim
248f9affc9 Pull in r323998 from upstream clang trunk (by Richard Smith):
PR36157: When injecting an implicit function declaration in C89, find
  the right DeclContext rather than injecting it wherever we happen to
  be.

  This avoids creating functions whose DeclContext is a struct or
  similar.

This fixes assertion failures when parsing certain not-completely-valid
struct declarations.

Reported by:	ae
PR:		225862
MFC after:	3 months
X-MFC-With:	r327952
2018-02-13 17:05:50 +00:00
hselasky
4287266c4f Import the mthca kernel side infiniband driver from Linux 4.9 and fix
compilation under FreeBSD. The mthca driver was temporarily removed as
part of the Linux 4.9 RoCE/infinband upgrade.

Top commit in Linux source tree:
69973b830859bc6529a7a0468ba0d80ee5117826

Sponsored by:	Mellanox Technologies
2018-02-13 17:04:34 +00:00
bdrewery
8b002dd3b0 Allow overriding VTABSIZE at compile-time.
Reviewed by:	jilles
Differential Revision:	https://reviews.freebsd.org/D14339
2018-02-13 16:48:57 +00:00
lwhsu
4e36a1e0b3 Fix non-64-bit platform build by printing bus_addr_t values using %#jx
Reviewed by:	slm
Differential Revision:	https://reviews.freebsd.org/D14344
2018-02-13 16:26:06 +00:00
royger
d123f846b5 list myself in the MAINTAINERS file for Xen bits 2018-02-13 16:25:43 +00:00
brd
85830188f4 Fix resolv to run when it should and not when it should not..
Approved by:	manu
Reported by:	manu
Pointy hat to:	brd
2018-02-13 16:07:39 +00:00
kib
068e7ccc44 linuxkpi: Do not leak pages on put.
When the owner of the wire reference releases the last reference, it
might be that the page was already attempted to be freed (but free
cannot be performed at that time due to wire).  Check that the page
was removed from the object as the indicator of the free attempt and
finish the free operation if so.

Reported and tested by:	Slava Shwartsman
Reviewed by:	hselasky
Sponsored by:	Mellanox Technologies
MFC after:	1 week
2018-02-13 15:44:35 +00:00
kib
0c20f07bdd Do not leak rv->psind in some specific situations.
Suppose that we have an object with a mapped superpage, and that all
pages in the superpages are held (by some driver).  Additionally,
suppose that the object is terminated, e.g. because the only process
mapping it is exiting.  Then the reservation is broken, but the pages
cannot be freed until later, when they are unheld.  In this situation,
the reservation code cannot clean psind, since no pages are freed, and
the page is freed and then reused with invalid psind.

Clean psind on vm_reserv_break() to avoid the situation.

Reported and tested by:	Slava Shwartsman
Reviewed by:	markj
Sponsored by:	Mellanox Technologies
MFC after:	1 week
Differential revision:	https://reviews.freebsd.org/D14335
2018-02-13 15:36:28 +00:00
kib
59d970a7ef Fix build with gas.
Do not use C constant suffixes.  Bit values are small enough to not
require typing, despite they are used for 64bit MSR writes.  The added
cast in hw_ibrs_recalculate() is redundand but I prefer to add it for
clarity.

Reported by:	bde
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2018-02-13 15:30:31 +00:00
remko
2dd567b6c9 Update the secteam/security officer list to match reality. 2018-02-13 15:13:43 +00:00
hselasky
680b68c4f6 Fix for incorrect PnP information used by devmatch(8).
MFC after:	1 week
Sponsored by:	Mellanox Technologies
2018-02-13 11:43:57 +00:00
n_hibma
6d5125e8c2 DSCP values passed to setdscp need to be lowercase.
See definition of f_ipdscp values. They are compared against using bcmp
which is case sensitive.

MFC after:	1 week
2018-02-13 10:11:39 +00:00
hselasky
65a55b1b71 Remove leftover empty directory. 2018-02-13 08:17:19 +00:00
hselasky
0880adcb5c Add new USB quirk.
PR:		225844
MFC after:	1 week
Sponsored by:	Mellanox Technologies
2018-02-13 08:13:20 +00:00
hselasky
d332fd9c7d Add missing semicolon to not break devd during system startup. 2018-02-13 08:10:17 +00:00
araujo
e63cefb08a - I'm not using nis(8) and ypldap(8) anymore, so I'm removing myself
from MAINTAINERS list.
2018-02-13 06:36:27 +00:00
jhibbits
8edcc09486 Unify metadata load files for arm, mips, powerpc, sparc64
Summary:
All metadata.c files are very similar, with only trivial changes.  Unify them
into a single common file, with minor special-casing where needed.

Reviewed By: imp
Differential Revision: https://reviews.freebsd.org/D13978
2018-02-13 03:44:50 +00:00
dteske
51d4dde7fd Use tabs in io.d, fix alignment issues, remove extraneous newlines 2018-02-12 23:53:38 +00:00
jeff
ba27b5187b Make v_wire_count a per-cpu counter(9) counter. This eliminates a
significant source of cache line contention from vm_page_alloc().  Use
accessors and vm_page_unwire_noq() so that the mechanism can be easily
changed in the future.

Reviewed by:	markj
Discussed with:	kib, glebius
Tested by:	pho (earlier version)
Sponsored by:	Netflix, Dell/EMC Isilon
Differential Revision:	https://reviews.freebsd.org/D14273
2018-02-12 22:53:00 +00:00
cem
b604274973 devmatch.8: Link to MODULE_PNP_INFO(9)
Sponsored by:	Dell EMC Isilon
2018-02-12 22:43:47 +00:00
cem
80c14b338f Take a first stab at documenting use of MODULE_PNP_INFO
Sponsored by:	Dell EMC Isilon
2018-02-12 22:42:06 +00:00
landonf
9d02c8f760 bwn(4): Fix outstanding bug in PHY-G tssi2dbm table generation caught by
-Wconstant-conversion, and remove now unnecessary warning suppression
flags.
2018-02-12 22:21:11 +00:00
vangyzen
63337e8587 Update the MTU in affected routes when IPv6 RA changes the MTU
ip6_calcmtu() only looks at the interface MTU if neither the TCP hostcache
nor the route provides an MTU.  Update the routes so they do not provide
stale MTUs.

This fixes UNH IPv6 conformance test cases v6LC_4_1_08 and v6LC_4_1_09,
which use a RA to reduce the link MTU from 1500 to 1280.

Reported and tested by:	Farrell Woods <Farrell_Woods@Dell.com>
Reviewed by:	dab, melifaro
Discussed with:	ae
MFC after:	1 week
Sponsored by:	Dell EMC
Differential Revision:	https://reviews.freebsd.org/D14257
2018-02-12 19:49:20 +00:00
landonf
652dabad0f siba(4): Ignore disabled per-core address match entries.
Previously, the address regions described by disabled admatch entries would
be treated as being mapped to the given core; while incorrect, this was
essentially harmless given that the entries describe unused address space
on the few affected devices.

We now perform parsing of per-core admatch registers and interrupt flags in
siba_erom, correctly skip any disabled admatch entries, and use the
siba_erom API in siba_add_children() to perform enumeration of attached
cores.
2018-02-12 19:36:26 +00:00
asomers
e87bca0221 Fix a comment. No functional change.
MFC after:	3 weeks
Sponsored by:	Spectra Logic Corp
2018-02-12 17:42:28 +00:00
ian
e1340a314d Add a new sysctl, debug.clock_do_io, to allow manully triggering a one-shot
read or write of all registered realtime clocks.  In the read case, the
values read are simply discarded.  For writes, there's no alternative but
to actually write the current system time to the device.
2018-02-12 17:41:11 +00:00
ian
e5909c170b Add a set of convenience routines for RTC drivers to use for debug output,
and a debug.clock_show_io sysctl to control debugging output.
2018-02-12 17:33:14 +00:00
jtl
c63f80e908 Mark the pages used for the initial page-table entries as wired. This
makes them consistent with the way other page-table pages are allocated.
It also provides the rest of the VM system a good clue that these pages
are used.

Reviewed by:	alc, kib, markj
Sponsored by:	Netflix
Differential Revision:	https://reviews.freebsd.org/D14269
2018-02-12 17:27:50 +00:00
ian
19e2991d59 Replace the existing print_ct() private debugging function with a set of
three public functions to format and print the three major data structures
used by realtime clock drivers (clocktime, bcd_clocktime, and timespec).
2018-02-12 16:25:56 +00:00
imp
0d0658d73e Add lua to build script 2018-02-12 15:32:05 +00:00
imp
044b6c0684 Add the lua scripts from the lua-bootloader SoC
These are the .lua files from from Pedro Souza's 2014 Summer of Code
project. Rui Paulo, Pedro Arthur and Wojciech A. Koszek also
contributed.

Obtained from: https://wiki.freebsd.org/SummerOfCode2014/LuaLoader
Sponsored by: Google Summer of Code

Improve the SoC lua menu code to bring it in line with forth
menu functionality

Submitted by: Zakary Nafziger
Sponsored by: FreeBSD Foundation

Use loader.setenv and loader.unsetenv instead of loader.perform

Convert from include("/boot/foo.lua") to foo = require("foo");
to bring in line with latest lua module conventions.

Enforce a uniform style for the new .lua files:
	o hard tab indenation for 8 spaces
	o don't have if foo then bar; else bas; end on one line

MFC After: 1 month
Relnotes: yes
Differential Review: https://reviews.freebsd.org/D14295
2018-02-12 15:32:00 +00:00
imp
974e9d9ccc Add Lua as a scripting langauge to /boot/loader
liblua glues the lua run time into the boot loader. It implements all
the runtime routines that lua expects. In addition, it has a few
standard 'C' headers that nueter various aspects of the LUA build that
are too specific to lua to be in libsa. Many refinements from the
original code to improve implementation and the number of included lua
libraries. Use int64_t for lua_Number. Have "/boot/lua" be the default
module path. Numerous cleanups from the original GSoC project,
including hacking libsa to allow lua to be built with only one change
outside luaconf.h.

Add the final bit of lua glue to bring in liblua and plug into the
multiple interpreter framework, previously committed.

Add LOADER_LUA option, currently off by default.

Presently, this is an experimental option. One must opt-in to using
this by defining WITH_LOADER_LUA and WITHOUT_FORTH. It's been
lightly tested, so keep a backup copy of your old loader handy.
The menu code, coming in the next commit, hasn't been exhaustively
tested. A LUA boot loader is 60k larger than a FORTH one, which is
80k larger than a no-interpreter one. Subtle changes in size
may tip things past some subtle limit (the binary is ~430k now
when built with LUA). A future version may offer coexistance.

Bump FreeBSD version to 1200058 to mark the milestone.

Pedro Souza's 2014 Summer of Code project. Rui Paulo, Pedro Arthur,
Zakary Nafziger and Wojciech A. Koszek also contributed. Warner Losh
reworked it extensively into its current form.

Obtained from: https://wiki.freebsd.org/SummerOfCode2014/LuaLoader
Sponsored by: Google Summer of Code
Relnotes: Yes
MFC After: 1 month
Differential Review: https://reviews.freebsd.org/D14295
2018-02-12 15:31:53 +00:00
imp
8256f7ec08 Move __va_list and related defines to sys/sys/_types.h
__va_list and related defines are identical in all the
ARCH/include/_types.h files. Move them to sys/sys/_types.h

Sponsored by: Netflix
2018-02-12 14:48:20 +00:00
imp
c78af2dc5f We don't support gcc < 4.2.1, so varargs.h now is just #error
always. Unifdef for versions prior to 4.2.1 and remove now-unused
header files.

Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D14323
2018-02-12 14:48:14 +00:00
imp
f91c747c4d Use standard pattern for stdargs.h
We don't support older compilers. Most of the code in these files is
for pre-3.0 gcc, which is at least 15 years obsolete. Move to using
phk's sys/_stdargs.h for all these platforms.

Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D14323
2018-02-12 14:48:05 +00:00
tychon
e942535b5d Provide further mitigation against CVE-2017-5715 by flushing the
return stack buffer (RSB) upon returning from the guest.

This was inspired by this linux commit:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/x86/kvm?id=117cc7a908c83697b0b737d15ae1eb5943afe35b

Reviewed by:	grehan
Sponsored by:	Dell EMC Isilon
Differential Revision:	https://reviews.freebsd.org/D14272
2018-02-12 14:45:27 +00:00
imp
b3f3f8e9c5 Move devmatch to sbin from usr/sbin.
Since we want to use devmatch in context before a split /,/usr system
has mounted /usr, move devmatch to /sbin.

Sponsored by: Netflix
2018-02-12 14:44:21 +00:00
imp
ebc6d15cf1 Fix typo 2018-02-12 06:52:49 +00:00
imp
07ac0a0cfb Turn devmatch on by default.
Turn devmatch on by default. However, use 'start' instead of
'onestart' in the devmatch.conf file so the setting of
'devmatch_enable' is honored. Give an example of what to put in
devd.conf if you want to disable just the run-time part of devmatch.

Relnotes: yes
2018-02-12 06:51:20 +00:00
imp
c0ed339548 Add usb.conf to ObsoleteFiles.
Add a note to UPDATING.
Fix a missing tab.

Relnotes: Yes
2018-02-12 06:42:38 +00:00
imp
43df9c2b26 Install devmatch.conf, don't install usb.conf 2018-02-12 04:54:51 +00:00
imp
8d22013907 Install devmatch int /etc/rc.d and echo modules being installed. 2018-02-12 04:52:25 +00:00
imp
60be4b058e Switch to using devmatch to autoload drivers. Remove usb.conf
as obsolete because devmatch gets its information from the same
place as the genration scripts.
2018-02-12 04:45:26 +00:00
imp
3fce0a572b Add devmatch rc.d integration
Create simple script to load modules on demand based on the device
identifying information.

Sponsored by: Netflix
2018-02-12 04:45:17 +00:00
cem
c4503b16c3 device_add_child.9: reference logical opposite, device_delete_child(9) 2018-02-11 22:09:07 +00:00
scottl
084e8354e2 Print out the shared memory queues during initialization
Sponsored by:	Netflix
2018-02-11 20:15:47 +00:00
brooks
73867e4fb2 Use syscall_helper_register(9) rather than syscall_register().
The usage is simpler, documented, and more common.

Reviewed by:	cem
Sponsored by:	DARPA, AFRL
Differential Revision:	https://reviews.freebsd.org/D14227
2018-02-11 18:37:08 +00:00
imp
908bb2c3cc Consistent macro indentation is the hobgoblin of small minds
Line up the macro definitions and names here like the machine/stdarg.h
files that this replaced. This is easier to read and also makes it
easier to match up with other includes. Also two space indent va_end
to match the rest of the surrounding if block.
2018-02-11 17:45:38 +00:00
ian
4bdfc364c4 Regenerate devd/usb.conf after the recent addition of several new device IDs. 2018-02-11 16:35:56 +00:00
ian
fffc10fb1d Add a device ID to uftdi for TIAO USB Multi Protocol Adapter (TUMPA).
PR:		225810
2018-02-11 16:35:23 +00:00