]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
6 years agoAvoid use of the fdt_get_property_*() API, which is intrinsically
Nathan Whitehorn [Sat, 30 Dec 2017 20:28:29 +0000 (20:28 +0000)]
Avoid use of the fdt_get_property_*() API, which is intrinsically
incompatible with FDT versions < 16. This also simplifies the code a bit.

MFC after: 1 month

6 years agoGarbage-collect loader.ps3. It is currently disconnected from the build and
Nathan Whitehorn [Sat, 30 Dec 2017 20:27:13 +0000 (20:27 +0000)]
Garbage-collect loader.ps3. It is currently disconnected from the build and
is superseded by either direct loading of the kernel by petitboot (soon to
become the installer default) or loader.kboot.

6 years agoRemove logic for early console with loader.ps3 now that loader.ps3 is dead.
Nathan Whitehorn [Sat, 30 Dec 2017 20:25:33 +0000 (20:25 +0000)]
Remove logic for early console with loader.ps3 now that loader.ps3 is dead.

6 years agoChange the way SMP startup works to match the new multi-AP features in
Nathan Whitehorn [Sat, 30 Dec 2017 20:24:33 +0000 (20:24 +0000)]
Change the way SMP startup works to match the new multi-AP features in
locore64.S introduced in r327358.

MFC after: 3 weeks

6 years agoCheck more aggressively for whether the desired properties actually exist.
Nathan Whitehorn [Sat, 30 Dec 2017 20:23:14 +0000 (20:23 +0000)]
Check more aggressively for whether the desired properties actually exist.
If they don't, the code would look up some random part of the device tree
and seize the console inappropriately.

MFC after: 2 weeks

6 years agoAdd macro for vxlan list mutex lock and unlock
Bryan Venteicher [Sat, 30 Dec 2017 19:49:40 +0000 (19:49 +0000)]
Add macro for vxlan list mutex lock and unlock

This will simplify some later VNET support.

Submitted by: hrs
MFC after: 2 weeks

6 years agoAdvertise IFCAP_LINKSTAT after r326480 added link status support
Bryan Venteicher [Sat, 30 Dec 2017 19:35:12 +0000 (19:35 +0000)]
Advertise IFCAP_LINKSTAT after r326480 added link status support

MFC after: 2 weeks

6 years agoFix r327383:
Cy Schubert [Sat, 30 Dec 2017 19:27:22 +0000 (19:27 +0000)]
Fix r327383:

.../sys/dev/ep/elink.c:31:1: error: '/*' within block comment
[-Werror,-Wcomment]
/* $NetBSD: elink.c,v 1.6 1995/01/07 21:37:54 mycroft Exp $
^

6 years agoelink.[ch]: Move historic VCS tags after the license.
Pedro F. Giffuni [Sat, 30 Dec 2017 15:13:33 +0000 (15:13 +0000)]
elink.[ch]: Move historic VCS tags after the license.

This matches better our common practices and style.

6 years agoImprove usbconfig(8) manual page by adding descriptions for subcommands.
Edward Tomasz Napierala [Sat, 30 Dec 2017 12:56:50 +0000 (12:56 +0000)]
Improve usbconfig(8) manual page by adding descriptions for subcommands.

Reviewed by: hselasky
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D13445

6 years agoMove i386/isa/elink.[hc] to dev/ep.
Konstantin Belousov [Sat, 30 Dec 2017 11:42:49 +0000 (11:42 +0000)]
Move i386/isa/elink.[hc] to dev/ep.

The ep(4) driver is the only consumer of the two functions from
elink.c.  I removed the standalone module as well, and most likely,
the module metadata is not needed anywhere, but this is for later
cleanup.

Discussed with: imp, jhb
Sponsored by: The FreeBSD Foundation

6 years agoMove i386/isa/npx.c to i386i386/npx.c.
Konstantin Belousov [Sat, 30 Dec 2017 11:33:04 +0000 (11:33 +0000)]
Move i386/isa/npx.c to i386i386/npx.c.

The i386 FPU (AKA npx) code does not depend on ISA devices at all,
after the support for IRQ13 FPU exceptions was removed.  Put the file
into the expected place in the kernel source tree.

Discussed with: jhb
Sponsored by: The FreeBSD Foundation

6 years agoOn further testing on actual machines with this hardware, we should
Warner Losh [Sat, 30 Dec 2017 08:16:31 +0000 (08:16 +0000)]
On further testing on actual machines with this hardware, we should
only warn for devices that are attached. Add missing \n.

6 years agoRemove two stray references to wl driver, removed some time ago.
Warner Losh [Sat, 30 Dec 2017 07:59:32 +0000 (07:59 +0000)]
Remove two stray references to wl driver, removed some time ago.

6 years agolibsa: Pull in strnlen from libc
Kyle Evans [Sat, 30 Dec 2017 07:03:52 +0000 (07:03 +0000)]
libsa: Pull in strnlen from libc

strnlen is not used at the moment, but it will be when libfdt gets updated.
Prepare for the not-so-distant future by pulling in strnlen.

Noticed because: segfault in ld.bfd due to strnlen missing

6 years agostand/fdt: Swap libfdt include order
Kyle Evans [Sat, 30 Dec 2017 06:53:27 +0000 (06:53 +0000)]
stand/fdt: Swap libfdt include order

libfdt.h should be included before fdt.h, as hinted at by all of libfdt/;
standard include order being libfdt.h, libfdt_env.h, fdt.h.

The current include order also causes problems when libfdt gets updated, as
fdt.h requires some definitions from libfdt.h.

Differential Revision: https://reviews.freebsd.org/D13688

6 years agoAdd support for IPv6 scoped addresses to vxlan
Bryan Venteicher [Sat, 30 Dec 2017 04:03:53 +0000 (04:03 +0000)]
Add support for IPv6 scoped addresses to vxlan

MFC after: 2 weeks

6 years agoRemove a redunant check.
John Baldwin [Sat, 30 Dec 2017 03:08:49 +0000 (03:08 +0000)]
Remove a redunant check.

6 years agogeom_ccd.c: Fix the licenses properly
Pedro F. Giffuni [Sat, 30 Dec 2017 02:07:18 +0000 (02:07 +0000)]
geom_ccd.c: Fix the licenses properly

The license merging in r109471 didn't take into account that licensing
could change. Just removing the 3rd clause obviates the copyright
assignment to the NetBSD Foundation.

We do have plenty of files that have two or more licensing as in this
case, so fix this properly by splitting back the licenses as they are
upstream.

Obtained from: NetBSD

6 years agogeom_ccd.c: Update the license with changes from upstream.
Pedro F. Giffuni [Sat, 30 Dec 2017 01:37:08 +0000 (01:37 +0000)]
geom_ccd.c: Update the license with changes from upstream.

Part of this file originated in NetBSD, with the original file
carrying two versions of 4-clause BSD licenses. r109471 attempted to
simplify the situation by putting both licenses together.

Meanwhile, NetBSD dropped Clauses 3 and 4 from their own license, and
eventually NetBSD got permission from the University of Utah to drop the
3rd clause.

Keep the license "simple" by dropping the third clause since both TNF,
Utah/Berkeley and phk agree in principle that it can be dropped.

Obtained from: NetBSD (ccd.c CVS 1.128, 1.138)

6 years agonet80211: sanitize input for ieee80211_output()
Andriy Voskoboinyk [Sat, 30 Dec 2017 00:40:34 +0000 (00:40 +0000)]
net80211: sanitize input for ieee80211_output()

- Add some basic checks for i_fc* bits (ToDS, FromDS, MoreFrag, Protected);
those are used / checked across various places in Tx path.
- Mark injected 802.11 frame as encapsulated (just as it should be).
- Classify 802.11 frame in a proper way (extract ether_type from LLC header
for Data frames, use AC_BE queue for others (NoData / Management / Control).
- Subtract header length from tx_bytes statistics (so it will correspond
to the comment).

Was checked with RTL8188EU (AP) + Intel 6205 (STA).

Reviewed by: adrian
Differential Revision: https://reviews.freebsd.org/D13161

6 years agoccdconfig: Move VCS tags to be more consistent with our style.
Pedro F. Giffuni [Sat, 30 Dec 2017 00:26:42 +0000 (00:26 +0000)]
ccdconfig: Move VCS tags to be more consistent with our style.

Update a now-bogus SPDX tag while here.

6 years agonet80211: handle VHT nodes in ieee80211_node_setuptxparms()
Andriy Voskoboinyk [Sat, 30 Dec 2017 00:24:53 +0000 (00:24 +0000)]
net80211: handle VHT nodes in ieee80211_node_setuptxparms()

Select proper mode when node can do VHT.

Currently there are no drivers with VHT support in the tree,
so this should be noop.

Reviewed by: adrian
Differential Revision: https://reviews.freebsd.org/D9806

6 years agoccdconfig: Update licensing terms to match NetBSD.
Pedro F. Giffuni [Sat, 30 Dec 2017 00:22:47 +0000 (00:22 +0000)]
ccdconfig: Update licensing terms to match NetBSD.

The code originated in NetBSD which has since removed Clauses 3 and 4.

Approved by: phk (concerning his own copyright)
Obtained from: NetBSD (CVS ccdconfig.c 1.47, ccdconfig.8 1.24)

6 years agoMake kernel option KERNVIRTADDR optional, remove it from std.<platform>
Ian Lepore [Sat, 30 Dec 2017 00:20:49 +0000 (00:20 +0000)]
Make kernel option KERNVIRTADDR optional, remove it from std.<platform>
files that can use the default value.

It used to be required that the low-order bits of KERNVIRTADDR matched
the low-order bits of the physical load address for all arm platforms.
That hasn't been a requirement for armv6 platforms since FreeBSD 10.
There is no longer any relationship between load addr and KERNVIRTADDR
except that both must be aligned to a 2 MiB boundary.

This change makes the default KERNVIRTADDR value 0xc0000000, and removes the
options from all the platforms that can use the default value.  The default
is now defined in vmparam.h, and that file is now included in a few new
places that reference KERNVIRTADDR, since it may not come in via the
forced-include of opt_global.h on the compile command line.

6 years agoexamples: fix typo
Eitan Adler [Fri, 29 Dec 2017 23:58:05 +0000 (23:58 +0000)]
examples: fix typo

6 years agoRemove a reference to burncd
Eitan Adler [Fri, 29 Dec 2017 23:56:06 +0000 (23:56 +0000)]
Remove a reference to burncd

I'm unclear of the purpose of this pseudo-target
but while its here keep it up to date

6 years agoAdd AHCI/XHCI device IDs found on AMD 1950X+X399 system
Conrad Meyer [Fri, 29 Dec 2017 22:24:41 +0000 (22:24 +0000)]
Add AHCI/XHCI device IDs found on AMD 1950X+X399 system

A follow-up to r327094.

Sponsored by: Dell EMC Isilon

6 years agofind(1): Fix -newer and -samefile to conform to POSIX[0]
Conrad Meyer [Fri, 29 Dec 2017 22:08:43 +0000 (22:08 +0000)]
find(1): Fix -newer and -samefile to conform to POSIX[0]

By default, or with the -P flag, find(1) should evaluate paths "physically."
For symlinks, this means using the link itself instead of the target.

Historically (since the import of BSD 4.4-lite from CSRG), find(1) has
failed to refer to the link itself, at least for -newer and -samefile.

[0]: http://pubs.opengroup.org/onlinepubs/9699919799/utilities/find.html

PR: 222698
Reported by: Harald Schmalzbauer <bugzilla.freebsd AT omnilan.de>
Sponsored by: Dell EMC Isilon

6 years agoSort by time.
Maxim Konovalov [Fri, 29 Dec 2017 21:37:36 +0000 (21:37 +0000)]
Sort by time.

6 years agoEnhance the CHRP/pSeries platform layer:
Nathan Whitehorn [Fri, 29 Dec 2017 21:09:17 +0000 (21:09 +0000)]
Enhance the CHRP/pSeries platform layer:
- Densely number CPUs to avoid systems with CPUs with very high ID numbers
- Always have the BSP be CPU 0 to avoid remnant brokenness with non-0 BSPs
  in other parts of the kernel.
- Improve parsing of the device tree CPU listings on SMT systems.
- Allow reboot via RTAS as well as OF for pSeries systems booted by FDT
  without functioning Open Firmware.

Obtained from: projects/powernv
MFC after: 3 weeks

6 years agoDo not lock vm map in swapout_procs().
Konstantin Belousov [Fri, 29 Dec 2017 20:33:56 +0000 (20:33 +0000)]
Do not lock vm map in swapout_procs().

Neither swapout_procs() nor swapout() access the map.  Since the
process' vmspace is referenced only to obtain the pointer to the
vm_map, the reference is not needed as well.

Reviewed by: alc
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D13681

6 years agoAdd support for 64-bit PowerPC kernels to be directly loaded by kexec, which
Nathan Whitehorn [Fri, 29 Dec 2017 20:30:10 +0000 (20:30 +0000)]
Add support for 64-bit PowerPC kernels to be directly loaded by kexec, which
is used as the bootloader on a number of PPC64 platforms. This involves the
following pieces:
- Making the first instruction a valid kernel entry point, since kexec
  ignores the ELF entry value. This requires a separate section and linker
  magic to prevent the linker from filling the beginning of the section
  with stubs.
- Adding an entry point at 0x60 past the first instruction for systems
  lacking firmware CPU shutdown support (notably PS3).
- Linker script changes to support the above.

MFC after: 1 month

6 years agoMaintain alignment of in-code 64-bit quantities by design rather than luck.
Nathan Whitehorn [Fri, 29 Dec 2017 20:25:15 +0000 (20:25 +0000)]
Maintain alignment of in-code 64-bit quantities by design rather than luck.
If these are not aligned, the linker has to emit a different type of
relocation that the early boot self-relocation code cannot handle, even
in principle, resulting in them being set to zero and the kernel crashing.

MFC after: 1 week

6 years agoCorrect a mistake and reword a couple sentences to clarify that "the value"
Ian Lepore [Fri, 29 Dec 2017 20:00:19 +0000 (20:00 +0000)]
Correct a mistake and reword a couple sentences to clarify that "the value"
refers to the scale value, not the kmem_arena size that results from scaling.

Suggested by: alc@

6 years ago- Don't allow userland to switch partitions; it's next to impossible
Marius Strobl [Fri, 29 Dec 2017 19:07:50 +0000 (19:07 +0000)]
- Don't allow userland to switch partitions; it's next to impossible
  to recover from that, especially when something goes wrong.
- When userland changes EXT_CSD, update the kernel copy before using
  relevant EXT_CSD bits in mmcsd_switch_part().

6 years agoStyle.
Konstantin Belousov [Fri, 29 Dec 2017 19:05:07 +0000 (19:05 +0000)]
Style.

Reviewed by: alc
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D13678

6 years agogeli: fix the resize test on arm64
Alan Somers [Fri, 29 Dec 2017 18:42:55 +0000 (18:42 +0000)]
geli: fix the resize test on arm64

The resize test used bsdlabel(8), which is not available on all
architectures.  Change it to use gpart(8) instead, which should be available
everywhere.

PR: 221763
Reported by: andrew
MFC after: 2 weeks

6 years agoFix a harmless typo from r310786
Alan Somers [Fri, 29 Dec 2017 18:09:06 +0000 (18:09 +0000)]
Fix a harmless typo from r310786

I copy/pasted a reference to an undefined shell variable.

MFC after: 2 weeks

6 years agoFix ubldr. uboot/lib uses defines for the loader. It's part of the
Warner Losh [Fri, 29 Dec 2017 18:08:35 +0000 (18:08 +0000)]
Fix ubldr. uboot/lib uses defines for the loader. It's part of the
loader, but not compile as loader (it's building a library), so we
can't just include loader.mk for the defines. Move LOADER_DISK_SUPPORT
back to defs.mk for the moment.

Sponsored by: Netflix

6 years agostand/fdt: Consistently apply fdt_overlays
Kyle Evans [Fri, 29 Dec 2017 18:08:30 +0000 (18:08 +0000)]
stand/fdt: Consistently apply fdt_overlays

Overlays were previously not applied when U-Boot provides FDT or EFI
provides FDT, only when we load FDT from /boot/dtb given name from U-Boot.

Make all three paths lead to loading fdt_overlays and applying them, so that
fdt_overlays can be expected to Just Work.

Reviewed by: gonzo, imp, manu
Differential Revision: https://reviews.freebsd.org/D13664

6 years agoAfter r327168, the variable "vm_pageout_wanted" can be static.
Alan Cox [Fri, 29 Dec 2017 17:02:22 +0000 (17:02 +0000)]
After r327168, the variable "vm_pageout_wanted" can be static.

MFC after: 2 weeks

6 years agogeli: factor out some common code in the geli tests
Alan Somers [Fri, 29 Dec 2017 16:13:06 +0000 (16:13 +0000)]
geli: factor out some common code in the geli tests

No functional change.

MFC after: 2 weeks
Sponsored by: Spectra Logic Corp

6 years agoFix potential TOCTTOU bug in the geli tests
Alan Somers [Fri, 29 Dec 2017 16:06:10 +0000 (16:06 +0000)]
Fix potential TOCTTOU bug in the geli tests

This change mostly reverts r293436, which introduced the bug due to a belief
that geli(8) would allocate md(4) devices by itself. However, that belief is
incorrect. Instead of using linear probing to find available md(4) numbers,
it's best to use the existing attach_md function.

Reviewed by: ngie
MFC after: 2 weeks
Sponsored by: Spectra Logic Corp
Differential Revision: https://reviews.freebsd.org/D13666

6 years agoAdd a crude workaround for ports with flavors
Poul-Henning Kamp [Fri, 29 Dec 2017 15:38:43 +0000 (15:38 +0000)]
Add a crude workaround for ports with flavors

6 years agolocale: add cnr (Montenegrin)
Eitan Adler [Fri, 29 Dec 2017 15:04:49 +0000 (15:04 +0000)]
locale: add cnr (Montenegrin)

See https://www.loc.gov/standards/iso639-2/php/code_changes.php

6 years agolibc/locale: fix an off-by-one in newlocale
Eitan Adler [Fri, 29 Dec 2017 14:56:46 +0000 (14:56 +0000)]
libc/locale: fix an off-by-one in newlocale

Reported by: zrj@DragonFlyBSD.org

6 years agobsd-family-tree: add NetBSD 7.1.1; correct NetBSD abbrev
Eitan Adler [Fri, 29 Dec 2017 14:31:43 +0000 (14:31 +0000)]
bsd-family-tree: add NetBSD 7.1.1; correct NetBSD abbrev

Reported by: Herbert J. Skuhra <herbert@gojira.at>
Reported by: N.J. Mann <njm@njm.me.uk>

6 years ago- There is no need to keep the tuning error and re-tuning interrupts
Marius Strobl [Fri, 29 Dec 2017 12:48:19 +0000 (12:48 +0000)]
- There is no need to keep the tuning error and re-tuning interrupts
  enabled (though, no interrupt generation enabled for them) all the
  time as soon as (re-)tuning is supported; only enable them and let
  them generate interrupts when actually using (re-)tuning.
- Also disable all interrupts except SDHCI_INT_DATA_AVAIL ones while
  executing tuning and not just their signaling.

6 years agoCorrect the comment describing badrs which is bad router solicitiation,
Cy Schubert [Fri, 29 Dec 2017 07:23:18 +0000 (07:23 +0000)]
Correct the comment describing badrs which is bad router solicitiation,
not bad router advertisement.

MFC after: 3 days

6 years agoThe fix in r327273 turns a memory leak into freeing wild pointer.
Xin LI [Fri, 29 Dec 2017 06:22:05 +0000 (06:22 +0000)]
The fix in r327273 turns a memory leak into freeing wild pointer.
Fix this by freeing only the initialized pointer.

6 years agobsd-family-tree: add NetBSD 6.0.6
Eitan Adler [Fri, 29 Dec 2017 05:01:07 +0000 (05:01 +0000)]
bsd-family-tree: add NetBSD 6.0.6

This was a missing release. Released on the same day as 6.1.5 per
https://blog.netbsd.org/tnf/entry/netbsd_6_1_5_and

6 years agobsd-family-tree: add HardenedBSD
Eitan Adler [Fri, 29 Dec 2017 04:49:59 +0000 (04:49 +0000)]
bsd-family-tree: add HardenedBSD

This adds HardenedBSD which is a pseudo-fork of FreeBSD. It hasn't had a
release yet, but does does have active users and a community. As such
document it as a branch off of FreeBSD-stable. Ideally this adds enough
space so that future releases are easy enough to add.

6 years agocxgbe(4): Reduce duplication by consolidating minor variations of the
Navdeep Parhar [Fri, 29 Dec 2017 02:30:21 +0000 (02:30 +0000)]
cxgbe(4): Reduce duplication by consolidating minor variations of the
same code into a single routine.

Sponsored by: Chelsio Communications

6 years agoif_txpreg.h: drop clauses 3 and 4.
Pedro F. Giffuni [Fri, 29 Dec 2017 00:59:56 +0000 (00:59 +0000)]
if_txpreg.h: drop clauses 3 and 4.

Obtained from: OpenBSD (CVS 1.35)

6 years agodev/txp: Update if_txpreg.h to match OpenBSD's version.
Pedro F. Giffuni [Fri, 29 Dec 2017 00:27:12 +0000 (00:27 +0000)]
dev/txp: Update if_txpreg.h to match OpenBSD's version.

Resolve a minor mismatch: TXP_CMD_READ_VERSION instead of
TXP_CMD_VERSIONS_READ.
Curiously the later is defined but not used in OpenBSD.

Obtained from: OpenBSD (CVS 1.31-1.34)
MFC after: 1 week

6 years agoClean up the comment.
Konstantin Belousov [Thu, 28 Dec 2017 23:50:21 +0000 (23:50 +0000)]
Clean up the comment.

Reviewed by: alc, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D13671

6 years agoRemove ELF note for Open Firmware. It is marked optional in a single 1996
Nathan Whitehorn [Thu, 28 Dec 2017 23:49:53 +0000 (23:49 +0000)]
Remove ELF note for Open Firmware. It is marked optional in a single 1996
draft of a never-finalized standard (CHRP) and is irrelevant in practice
on FreeBSD since we load the kernel with loader(8) on Open Firmware
platforms anyway. Moreover, loader(8), which is directly loaded by Open
Firmware, has never had an equivalent note.

MFC after: 2 weeks

6 years agohumanize_number(3): fix math edge case in rounding large numbers
Bartek Rutkowski [Thu, 28 Dec 2017 22:57:34 +0000 (22:57 +0000)]
humanize_number(3): fix math edge case in rounding large numbers

Fix for remainder overflow, when in rare cases adding remainder to divider
exceeded 1 and turned the total to 1000 in final formatting, taking up
the space for the unit character.

The fix continues the division of the original number if the above case
happens -- added the appropriate check to the for loop performing
the division. This lowers the value shown, to make it fit into the buffer
space provided (1.0M for 4+1 character buffer, as used by ls).

Add test case for the reported bug and extend test program to support
providing buffer length (ls -lh uses 5, tests hard-coded 4).

PR: 224498
Submitted by: Pawel Biernacki <pawel.biernacki@gmail.com>
Reported by: Masachika Ishizuka <ish@amail.plala.or.jp>
Reviewed by: cem, kib
Approved by: cem, kib
MFC after: 1 week
Sponsored by: Mysterious Code Ltd.
Differential Revision: D13578

6 years agoIn vm_swapout_map_deactivate_pages(), it is enough to lock the map for read.
Konstantin Belousov [Thu, 28 Dec 2017 22:56:30 +0000 (22:56 +0000)]
In vm_swapout_map_deactivate_pages(), it is enough to lock the map for read.

Reviewed by: alc, markj (as part of the larger patch)
Tested by: pho (again, as part of the larger patch)
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D13671

6 years agoProbe Intel Denverton eMMC 5.0 controllers.
Marius Strobl [Thu, 28 Dec 2017 22:03:08 +0000 (22:03 +0000)]
Probe Intel Denverton eMMC 5.0 controllers.

6 years agoWith the advent of interrupt remapping, Intel has repurposed bit 11
Marius Strobl [Thu, 28 Dec 2017 21:46:09 +0000 (21:46 +0000)]
With the advent of interrupt remapping, Intel has repurposed bit 11
(now: Interrupt_Index[15]) and assigned the previously reserved bits
55:48 (Interrupt_Index[14:0] goes into 63:49 while Destination Field
used 63:56 and bit 48 now is Interrupt_Format) in the IO redirection
tables (see the VT-d specification, "5.1.5.1 I/OxAPIC Programming").
Thus, when not using interrupt remapping, ensure that all previously
reserved bits in the high part of the RTEs are zero instead of doing
a read-modify-write for their Destination Field bits only.
Otherwise, on machines based on Apollo Lake and its derivatives such
as Denverton, typically some of the previously preserved bits remain
set after boot when not employing interrupt remapping. The result is
that INTx interrupts are not getting delivered.
Note: With an AMD IOMMU, interrupt remapping apparently bypasses the
IO APIC altogether.

Submitted by: loos (modulo comment)
Reviewed by: jhb (modulo comment)

6 years agoRemove a few more dangling references to ie(4).
John Baldwin [Thu, 28 Dec 2017 21:35:53 +0000 (21:35 +0000)]
Remove a few more dangling references to ie(4).

6 years agoe1000: Add support for Ice Lake and Cannon Lake
Sean Bruno [Thu, 28 Dec 2017 21:26:40 +0000 (21:26 +0000)]
e1000: Add support for Ice Lake and Cannon Lake

Ths add initial support for Ice Lake and Cannon Lake ethernet devices.

This also addressed errata 1.5.4.4 for Sky Lake and Kabby Lake devices:
https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/i218-i219-ethernet-connection-spec-update.pdf?asset=9561

Submitted by: Kevin Bowling <kevin.bowling@kev009.com>
Relnotes: Yes
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D13660

6 years agostand/fdt: Avoid bailout when dtbo has no fixups
Kyle Evans [Thu, 28 Dec 2017 21:12:27 +0000 (21:12 +0000)]
stand/fdt: Avoid bailout when dtbo has no fixups

In the case of a simple dtbo where fragment uses target-path and the overlay
contains no references, /__fixups__ will not be included by either our dtc
or dtc from ports, but the file still has valid fragments to be applied.

Additional testing found that /__symbols__ might also be omitted if it's
empty, which is not necessarily an error.

Reviewed by: gonzo, imp
Differential Revision: https://reviews.freebsd.org/D13663

6 years agostand/fdt: Fix loading of multiple fdt_overlays
Kyle Evans [Thu, 28 Dec 2017 21:09:36 +0000 (21:09 +0000)]
stand/fdt: Fix loading of multiple fdt_overlays

fdt_load_dtb_overlays was written to unload previous overlay when a new
valid one is come across. fdt_apply_overlays further down is written to
iterate over all .dtbo's currently loaded and apply them one-by-one. Correct
fdt_load_dtb_overlays to stop dropping valid overlays that were previously
loaded and match expectations.

Reviewed by: gonzo, imp
Differential Revision: https://reviews.freebsd.org/D13659

6 years agoFix some typos.
Pedro F. Giffuni [Thu, 28 Dec 2017 20:40:56 +0000 (20:40 +0000)]
Fix some typos.

Obtained from: OpenBSD (CVS v1.5)

6 years agonetinet6/ip6_id.c: niels kindly dropped clause 3/4 from the license.
Pedro F. Giffuni [Thu, 28 Dec 2017 20:35:21 +0000 (20:35 +0000)]
netinet6/ip6_id.c: niels kindly dropped clause 3/4 from the license.

This bring back r327293 from OpenBSD, with the important difference that
we are now getting it from their ip6_id.c file.

Obtained from: OpenBSD (CVS v1.3)

6 years agoStart syncing changes from OpenBSD's ip6_id.c instead of ip_id.c.
Pedro F. Giffuni [Thu, 28 Dec 2017 20:26:51 +0000 (20:26 +0000)]
Start syncing changes from OpenBSD's ip6_id.c instead of ip_id.c.

correct non-repetitive ID code, based on comments from niels provos.
- seed2 is necessary, but use it as "seed2 + x" not "seed2 ^ x".
- skipping number is not needed, so disable it for 16bit generator (makes
  the repetition period to 30000)

Obtained from: OpenBSD (CVS rev. 1.2)
MFC after: 1 week

6 years agoRevert r327293
Pedro F. Giffuni [Thu, 28 Dec 2017 20:10:10 +0000 (20:10 +0000)]
Revert r327293
netinet6/ip6_id.c: niels kindly dropped clause 3/4 from the license.

I was looking at the wrong file. There is an important merge that must be
done before I can bring this change.

6 years agonetinet6/ip6_id.c: niels kindly dropped clause 3/4 from the license.
Pedro F. Giffuni [Thu, 28 Dec 2017 19:42:53 +0000 (19:42 +0000)]
netinet6/ip6_id.c: niels kindly dropped clause 3/4 from the license.

This file is supposed to be based on the OpenBSD CVS v1.6 but checking
the OpenBSD repository the license had already dropped the 2&3 clasues by
then. Catch up with the licensing.

Obtained from: OpenBSD (CVS 1.2)

6 years agoMinor formatting and style tweaks to some comments.
John Baldwin [Thu, 28 Dec 2017 18:51:35 +0000 (18:51 +0000)]
Minor formatting and style tweaks to some comments.

6 years agoRemove a stale reference to ie(4).
John Baldwin [Thu, 28 Dec 2017 17:55:58 +0000 (17:55 +0000)]
Remove a stale reference to ie(4).

This was missed in r304513.

Submitted by: kib

6 years agorpc.sprayd: Remove 3rd and 4th clauses in christos' license.
Pedro F. Giffuni [Thu, 28 Dec 2017 17:51:53 +0000 (17:51 +0000)]
rpc.sprayd: Remove 3rd and 4th clauses in christos' license.

Obtained from: NetBSD

6 years agorpc.sprayd: Bring some changes from NetBSD.
Pedro F. Giffuni [Thu, 28 Dec 2017 17:44:30 +0000 (17:44 +0000)]
rpc.sprayd: Bring some changes from NetBSD.

Most notable, other than some style issues:
CVS 1.11:
  do not use LOG_CONS.
CVS 1.13:
  consistently use exit instead of return in main().
  use LOG_WARNING instead of LOG_ERR for non critical errors.

Obtained from: NetBSD
MFC after: 2 weeks

6 years agosys/sys/chio.h: add NetBSD RCS ID.
Pedro F. Giffuni [Thu, 28 Dec 2017 14:26:33 +0000 (14:26 +0000)]
sys/sys/chio.h: add NetBSD RCS ID.

Make it easier to identify the point where we started diverging from
NetBSD. Newer versions of this file has been updated to a new license so
this information may become useful some day.

Obtained from: NetBSD

6 years agosys/i386/isa/elink*: sync with (older) NetBSD.
Pedro F. Giffuni [Thu, 28 Dec 2017 14:14:35 +0000 (14:14 +0000)]
sys/i386/isa/elink*: sync with (older) NetBSD.

Make it easier to identify the point where we started diverging from
NetBSD. Newer versions of these files have been updated to a new license
so this information may become useful some day.

Obtained from: NetBSD

6 years agoReuse kern_proc_vmmap_resident() for procfs_map resident count.
Konstantin Belousov [Thu, 28 Dec 2017 13:23:13 +0000 (13:23 +0000)]
Reuse kern_proc_vmmap_resident() for procfs_map resident count.

The existing algorithm in procfs_map() to calculate count of resident
pages in an entry is too primitive, resulting in too long run time for
large sparse mapping entries.  Re-use the kern_proc_vmmap_resident()
from kern_proc.c which only looks at the existing pages in the
iterations.

Also, this makes procfs to honor kern.proc_vmmap_skip_resident_count,
if user does not need this information.

Reported by: Glenn Weinberg <glenn.weinberg@intel.com>
PR: 224532
No objections from: des (procfs maintainer)
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D13595

6 years agoMake kern_proc_vmmap_resident() externally accesible, and move the
Konstantin Belousov [Thu, 28 Dec 2017 13:16:32 +0000 (13:16 +0000)]
Make kern_proc_vmmap_resident() externally accesible, and move the
vmmap_skip_res_cnt control check inside it.

Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D13595

6 years agoStyle. Remove useless return.
Konstantin Belousov [Thu, 28 Dec 2017 13:01:27 +0000 (13:01 +0000)]
Style.  Remove useless return.

Sponsored by: The FreeBSD Foundation
MFC after: 1 week

6 years agoComplete r327264 by fixing yet another return without cleanup.
Konstantin Belousov [Thu, 28 Dec 2017 12:59:48 +0000 (12:59 +0000)]
Complete r327264 by fixing yet another return without cleanup.

Sponsored by: The FreeBSD Foundation
MFC after: 1 week

6 years agoUpdate pci_vendors to 2017.12.20
Eitan Adler [Thu, 28 Dec 2017 08:25:47 +0000 (08:25 +0000)]
Update pci_vendors to 2017.12.20

6 years agoFix a few speelling errors
Eitan Adler [Thu, 28 Dec 2017 08:22:26 +0000 (08:22 +0000)]
Fix a few speelling errors

- man pages
- bin/sh

Reviewed by: jilles

6 years agoRevert r327250 as it broke the build for some armv6 kernel and all armv4/5
Emmanuel Vadot [Thu, 28 Dec 2017 07:31:14 +0000 (07:31 +0000)]
Revert r327250 as it broke the build for some armv6 kernel and all armv4/5

Reported by: ian

6 years agoClose fd when we're done dv-sending the file.
Warner Losh [Thu, 28 Dec 2017 05:34:38 +0000 (05:34 +0000)]
Close fd when we're done dv-sending the file.

CID: 978413

6 years agoFree mp on a couple of error paths.
Warner Losh [Thu, 28 Dec 2017 05:34:34 +0000 (05:34 +0000)]
Free mp on a couple of error paths.

CID: 978387

6 years agoClose the fp rather than the fd since we fdopen the fd so we don't
Warner Losh [Thu, 28 Dec 2017 05:34:29 +0000 (05:34 +0000)]
Close the fp rather than the fd since we fdopen the fd so we don't
leak the FILE *.

CID: 978383

6 years agoWhen bind fails, make sure we closed the socket we tried to bind the
Warner Losh [Thu, 28 Dec 2017 05:34:24 +0000 (05:34 +0000)]
When bind fails, make sure we closed the socket we tried to bind the
address to.

CID: 978244

6 years agoClose fd and fd1 before returning now that we're done with them.
Warner Losh [Thu, 28 Dec 2017 05:34:19 +0000 (05:34 +0000)]
Close fd and fd1 before returning now that we're done with them.

CID: 978234, 978236

6 years agoDon't leak outfile. Free it before we return from bincvt.
Warner Losh [Thu, 28 Dec 2017 05:34:14 +0000 (05:34 +0000)]
Don't leak outfile. Free it before we return from bincvt.

CID: 273685

6 years agoPlug memory leak by freeing wantedblk{32,64}.
Warner Losh [Thu, 28 Dec 2017 05:34:08 +0000 (05:34 +0000)]
Plug memory leak by freeing wantedblk{32,64}.

CID: 273655, 273656

6 years agoFree zbuf when kflag is true too.
Warner Losh [Thu, 28 Dec 2017 05:34:04 +0000 (05:34 +0000)]
Free zbuf when kflag is true too.

CID: 273376

6 years agoClose slave on fork error to prevent pty fd leak.
Warner Losh [Thu, 28 Dec 2017 05:33:59 +0000 (05:33 +0000)]
Close slave on fork error to prevent pty fd leak.

CID: 978209

6 years agoFree path before returnig.
Warner Losh [Thu, 28 Dec 2017 05:33:54 +0000 (05:33 +0000)]
Free path before returnig.

CID: 977827

6 years agoClose cpfrom in an error case.
Warner Losh [Thu, 28 Dec 2017 05:33:49 +0000 (05:33 +0000)]
Close cpfrom in an error case.

CID: 271771

6 years agoFree data after we're done with it.
Warner Losh [Thu, 28 Dec 2017 05:33:44 +0000 (05:33 +0000)]
Free data after we're done with it.

CID: 271595, 275337, 1368743

6 years agoFree inb on error return.
Warner Losh [Thu, 28 Dec 2017 05:33:39 +0000 (05:33 +0000)]
Free inb on error return.

CID: 270099

6 years agoFree some variables before they go out of scope.
Warner Losh [Thu, 28 Dec 2017 05:33:34 +0000 (05:33 +0000)]
Free some variables before they go out of scope.

CID: 92074, 270099

6 years agoPlug two resource leaks.
Warner Losh [Thu, 28 Dec 2017 05:33:29 +0000 (05:33 +0000)]
Plug two resource leaks.

CID: 92077, 92074

6 years agoUse proper failure path rather than just returning.
Warner Losh [Thu, 28 Dec 2017 05:33:24 +0000 (05:33 +0000)]
Use proper failure path rather than just returning.

CID: 1199354100689410068931006892