Commit Graph

246834 Commits

Author SHA1 Message Date
manu
cbfe9423a0 devfs: Avoid a malloc/free if we just need to increment the refcount
MFC after:	1 week
Sponsored by:	Gandi.net
2017-11-30 12:38:42 +00:00
trasz
410d6f52ca Make the ddb(4) "set" command list variables when called without parameters.
Obtained from:	CheriBSD
MFC after:	2 weeks
Sponsored by:	DARPA, AFRL
2017-11-30 12:22:15 +00:00
hselasky
923fadf4e8 Properly define the VLAN_XXX() function macros to avoid miscompilation when
used inside "if" statements comparing with another value.

Detailed explanation:
"if (a ? b : c != 0)" is not the same like "if ((a ? b : c) != 0)"
which is the expected behaviour of a function macro.

Affects:
toecore, linuxkpi and ibcore.

Reviewed by:	kib
MFC after:	3 days
Sponsored by:	Mellanox Technologies
2017-11-30 11:35:22 +00:00
delphij
3bf7697aa7 Prevent OOB access on corrupted msdos directories.
Submitted by:	Veo Zhang <veo live com>
MFC after:	2 weeks
2017-11-30 08:28:17 +00:00
fabient
9186fa18d0 Update man page for r325354.
Reported by:	bjk
2017-11-30 08:28:04 +00:00
eadler
873b8cdb15 Finish cleaning up license :\ 2017-11-30 08:23:40 +00:00
eadler
feafece16b Fix typo
Reported by: mjg, zrj@dragonflybsd.org
2017-11-30 07:24:33 +00:00
eadler
16818cfbda Fix pthread_condattr(3) type
Reported by: Sascha Wildner <swildner@dragonflybsd.org>
Reviewed by: mjg
2017-11-30 07:15:32 +00:00
eadler
241ac6f200 Remove advertising clause from pr(1)
Permitted by https://www.freebsd.org/copyright/license.html
Renumber 4->3 to be consistent with other files
2017-11-30 07:08:45 +00:00
imp
90013f9fc4 Use const pointers to avoid casting away constness.
The host_open interface was a legitimate mismatch to the userboot
function, while the other pointers didn't need to be non-const.

This makes the powerpc warning free again.

Sponsored by: Netflix
2017-11-30 05:01:56 +00:00
jkim
c1509f7c95 Add a tunable "debug.hwpstate_verify" to check P-state after changing it and
turn it off by default.  It is very inefficient to verify current P-state of
each core, especially for CPUs with many cores.  When multiple commands are
requested to the same power domain before completion of pending transitions,
the last command is executed according to the manual.  Because requests are
serialized by the caller, all cores will receive the same command for each
call.  Do not call sched_bind() and sched_unbind().  It is redundant because
the caller does it anyway.
2017-11-30 01:40:07 +00:00
jkim
ce7b988218 Fix style(9). 2017-11-29 23:52:31 +00:00
scottl
49fb5dd79b It's time to retire AHC_REG_PRETTY_PRINT and AHD_REG_PRETTY_PRINT from
the standard kernels.  They are still available as custom compile
options.
2017-11-29 23:41:49 +00:00
hselasky
7213d95f20 The sched_add() function is not only used when the thread is initially
started, but also by the turnstiles to mark a thread as runnable for
all locks, for instance sleepqueues do:
setrunnable()->sched_wakeup()->sched_add()

In r326218 code was added to allow booting from non-zero CPU numbers
by setting the ts_cpu field inside the ULE scheduler's sched_add()
function. This had an undesired side-effect that prior sched_pin() and
sched_bind() calls got disregarded. This patch fixes the
initialization of the ts_cpu field for the ULE scheduler to only
happen once when the initial thread is constructed during system
init. Forking will then later on ensure that a valid ts_cpu value gets
copied to all children.

Reviewed by:	jhb, kib
Discussed with:	nwhitehorn
MFC after:	1 month
Differential revision:	https://reviews.freebsd.org/D13298
Sponsored by:	Mellanox Technologies
2017-11-29 23:28:40 +00:00
arichardson
a0d914b8b9 Don't fail the build due to clang integer constant range warnings
This warning checks whether a constant is out of range of the integer
type. An example is `comparison of 'u_int' > 4294967295 is always false`
and in this case the warning makes sense.
However, when the type is a typedef that can be either 64 or 32 bits the
if condition is only tautological in some configurations so this should
not be a warning that fails the build.

Reviewed by:	dim
Approved by:	jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D12912
2017-11-29 21:16:14 +00:00
markj
40d9c0da1e Verify the object/vnode association after vget() in vm_pageout_clean().
It's theoretically possible for the vnode and object to be disassociated
while locks are dropped around the vget() call, in which case we
shouldn't proceed with laundering.

Noted and reviewed by:	kib
MFC after:	1 week
2017-11-29 19:47:09 +00:00
shurd
23eaeecabf Fix comment introduced in r326369
The code uses the set of all CPUs, it doesn't zero out the set.

Sponsored by:	Limelight Networks
2017-11-29 18:21:17 +00:00
shurd
0de624aa47 Ensure that ctx->ifc_cpus is always initialized
If a device didn't support MSI-X, ctx->ifc_cpus would not be initialized,
but the IRQ allocation routines still uses the value.  Move the
initialization to common code.

Sponsored by:	Limelight Networks
2017-11-29 18:14:57 +00:00
markj
1643870966 Remove some comments that became incorrect with r325530. 2017-11-29 14:34:05 +00:00
danfe
0c2b556f96 Fix several noticed style issues.
Reviewed by:	bde
Approved by:	bapt
2017-11-29 12:49:22 +00:00
kib
fcb3a5a8a1 Implement hardware mlx5(4) rx timestamps.
Driver support is only provided for ConnectX4/5.

System-time timestamp is calculated based on the free-running counter
timestamp provided by hardware.  Driver periodically samples the
counter to calibrate it against the system clock and uses linear
interpolation to convert.  Stability of the crystal which drives the
clock is +-50 ppm at the operational temperature, which makes the
algorithm good enough.

The calculation is somewhat delicate because all values are 64bit and
overflow the naive formula for linear interpolation.  The calculation
drops the least significant bits in advance, see the PREC shift in
mlx5_mbuf_tstmp().

Hardware stamps can be turned off by 'ifconfig mceN -hwrxtsmp'.  Buggy
firmware might result in small but visible errors in the reported
timestamps, detectable e.g. by nonsensical (negative) RTT values for
LAN pings.

Reviewed by:	gallatin, hselasky
Sponsored by:	Mellanox Technologies
Differential revision:	https://reviews.freebsd.org/D12638
2017-11-29 10:04:11 +00:00
hselasky
30eed323c8 Disallow TUN and TAP character device IOCTLs to modify the network device
type to any value. This can cause page faults and panics due to accessing
uninitialized fields in the "struct ifnet" which are specific to the network
device type.

MFC after:	1 week
Found by:	jau@iki.fi
PR:		223767
Sponsored by:	Mellanox Technologies
2017-11-29 09:40:11 +00:00
delphij
97f0932623 Remove unused include.
MFC after:	2 weeks
2017-11-29 09:18:24 +00:00
jhb
5567bb0dd7 Replace a reference to a license in another file with the license text.
The relevant file was recently renamed, so the reference was stale.
In addition, explicit licenses are more typical in our sources.
2017-11-29 05:07:54 +00:00
jhibbits
144fd2e765 Only check the page tables if within the KVA.
Devices aren't mapped within the KVA, and with the way 64-bit hashes the
addresses pte_vatopa() may not return a 0 physical address for a device.

MFC after:	1 week
2017-11-29 01:26:07 +00:00
imp
9afae0e118 Import lua 5.3.4 to contrib 2017-11-29 00:46:36 +00:00
jeff
990ca74cdc Eliminate kmem_arena and kmem_object in preparation for further NUMA commits.
The arena argument to kmem_*() is now only used in an assert.  A follow-up
commit will remove the argument altogether before we freeze the API for the
next release.

This replaces the hard limit on kmem size with a soft limit imposed by UMA.  When
the soft limit is exceeded we periodically wakeup the UMA reclaim thread to
attempt to shrink KVA.  On 32bit architectures this should behave much more
gracefully as we exhaust KVA.  On 64bit the limits are likely never hit.

Reviewed by:	markj, kib (some objections)
Discussed with:	alc
Tested by:	pho
Sponsored by:	Netflix / Dell EMC Isilon
Differential Revision:	https://reviews.freebsd.org/D13187
2017-11-28 23:40:54 +00:00
jeff
f93de233c6 Move domain iterators into the page layer where domain selection should take
place.  This makes the majority of the phys layer explicitly domain specific.

Reviewed by:	markj, kib (some objections)
Discussed with:	alc
Tested by:	pho
Sponsored by:	Netflix & Dell EMC Isilon
Differential Revision:	https://reviews.freebsd.org/D13014
2017-11-28 23:18:35 +00:00
imp
f6f0c520a8 Import Lua 3.5.4
Downloaded from https://www.lua.org/ftp/lua-5.3.4.tar.gz
2017-11-28 22:55:31 +00:00
asomers
7de0462df8 Fix fetching ntp leapfile after 325256
Submitted by:	Ronald Klop <ronald-lists@klop.ws>
Reviewed by:	asomers
MFC after:	3 days
X-MFC-With:	325256
2017-11-28 20:44:10 +00:00
arichardson
babd7f0f92 Fix fabs() for MIPS when used on -0.0
It would previously return negative zero for -0.0 since -0.0 does not
compare less than 0. The issue was discovered when running the libc++
test suite on softfloat MIPS64.

I have verified that both clang and GCC generate sensible code for the
builtin. For soft float they clear the sign bit using integer operations
and in hard float mode they use abs.d.

Reviewed by:	#mips, jhb, brooks, imp, emaste
Approved by:	jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D13135
2017-11-28 20:37:27 +00:00
glebius
7d6335dc2e Revert r325558 as it is incorrect. 2017-11-28 19:25:01 +00:00
glebius
a271a57a8d Fix obvious NULL pointer dereference from r310350. 2017-11-28 19:18:00 +00:00
brooks
c6fbed1a3a Disable vim syntax highlighting.
Vim's default pick doesn't understand that ';' is a comment character
and the result looks horrible.

Reviewed by:	emaste
2017-11-28 18:23:17 +00:00
gjb
fe40e02df7 Correct a comment after r326330.
MFC after:	3 days
MFC with:	r326330
Sponsored by:	The FreeBSD Foundation
2017-11-28 18:08:14 +00:00
gjb
157d6ec6df Add a comment to release/release.conf.sample documenting
EMBEDDEDPORTS. [1]

Remove and update stale documentation from release(7) while here.

PR:		206344 [1]
MFC after:	3 days
Sponsored by:	The FreeBSD Foundation
2017-11-28 18:02:58 +00:00
alc
c123c7433b When the swap pager allocates space on disk, it requests contiguous
blocks in a single call to blist_alloc().  However, when it frees
that space, it previously called blist_free() on each block, one at a
time.  With this change, the swap pager identifies ranges of
contiguous blocks to be freed, and calls blist_free() once per
range.  In one extreme case, that is described in the review, the time
to perform an munmap(2) was reduced by 55%.

Submitted by:	Doug Moore <dougm@rice.edu>
Reviewed by:	kib
MFC after:	1 week
Differential Revision:	https://reviews.freebsd.org/D12397
2017-11-28 17:46:03 +00:00
andrew
57c55a6746 Fix the direction we move the instruction to retuirn to after exiting the
kernel debugger. We should skip the breakpoint instruction, not execute the
instruction before it.

Pointy-hat to:	andrew
MFC after:	1 week
Sponsored by:	DARPA, AFRL
2017-11-28 16:26:56 +00:00
fsu
bcca10fed1 Sync bsd_to_linux_errno[] table with i386 and amd64 tables in the sys directory.
Additional fix for r326282.

MFC after: 1 week
Approved by: pfg
2017-11-28 16:25:46 +00:00
gjb
d59286dc30 Set DISTDIR and WRKDIRPREFIX when building ports within the
chroot(8) to avoid mtime changes within the ports checkout,
which can cause checksum differences.

MFC after:	3 days
Sponsored by:	The FreeBSD Foundation
2017-11-28 16:00:27 +00:00
trasz
c4c92c8ee7 Make kdb_reenter() silent when explicitly called from db_error().
This removes the useless backtrace on various ddb(4) user errors.

Reviewed by:	jhb@
Obtained from:	CheriBSD
MFC after:	2 weeks
Sponsored by:	DARPA, AFRL
Differential Revision:	https://reviews.freebsd.org/D13212
2017-11-28 12:53:55 +00:00
andrew
042b16bcf3 When we exit the kernel debugger having entered because of a breakpoint
instruction we need to jump over the instruction. Without this we will
execute the same instruction again and enter into the debugger again.

PR:		223917
Reported by:	emaste
MFC after:	1 week
Sponsored by:	DARPA, AFRL
2017-11-28 11:04:47 +00:00
kib
dbbac96e0d Fix index calculation for the page table pages for efirt 1:1 map.
Stop issuing pre-assigned number to enumerate all page table pages,
the assignment is incorrect.  Instead automatically calculate the next
unused index. This index in fact does not serve any purpose except to
be unique to satisfy vm_page_grab() interface, we do not look up the
page by the index later.

Reported and tested by:	emaste
Reviewed by:	andrew
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
PR:	223906
Differential revision:	https://reviews.freebsd.org/D13273
2017-11-28 09:34:43 +00:00
nwhitehorn
52d939886a Back out OF module installation in the event of failure. PS3 firmware gives
some ancient FDT version (2) that fails the init check in OFW_FDT. It is
still possible to make progress, but not while the OF layer is going crazy.
2017-11-28 06:31:39 +00:00
bryanv
dd1aa0e47b Correctly report the vmxnet3 link down media status
Reported by:	lew@perftech.com
MFC after:	1 week
2017-11-28 06:21:37 +00:00
nwhitehorn
9e2f8a9afd Remove assertion that a CPU be present before returning a PCPU for it. It
is up to the caller to check for a NULL return value. The assert was meant
to catch buggy code that did not check the return value. Some code, however,
was smart and used the return value to see if a CPU existed, which this
broke.

Requested by:	jhb@
2017-11-28 05:39:48 +00:00
brooks
8b5b436329 Update vis(3) the latest from NetBSD.
This adds VIS_DQ for compatiblity with OpenBSD.

Correct by an off-by-one error and a read buffer overflow detected using
asan.

MFC after:	1 day
2017-11-28 01:35:28 +00:00
landonf
dd47f09b69 siba(4): Add support for devices have neither PMU nor PWRCTL support.
Very early BHND Wi-Fi devices (e.g. BCM4318) do not support any form of
dynamic clock control; on these devices, any PMU requests that cannot be
met by the device's fixed clock state will return an appropriate error
code.

Approved by:	adrian (mentor, implicit)
Sponsored by:	The FreeBSD Foundation
2017-11-28 00:42:44 +00:00
landonf
720eab5d5c bhndb(4): Implement bridge support for the BCM4312 and other PCI_V0 chipsets.
Very early (PCI_V0) Broadcom PCI Wi-Fi chipsets have a few quirks when
compared to later PCI(e) core revisions:

- The standard static BAR0 mapping of the PCI core registers is discontiguous,
  with siba's cfg0 register block mapped distinctly from the other core
  registers.
- No dedicated ChipCommon register mapping is provided; instead, the
  single configurable register window must be used to access both
  ChipCommon and D11 core registers. The D11 core's operational semantics
  guarantee the safety of -- after disabling interrupts -- borrowing
  the single dynamic register window to perform the few ChipCommon
  operations required by a driver.

To support these early PCI devices:

- Allow defining multiple discontiguous BHNDB_REGWIN_T_CORE register
  windows that map a single port/region, and producing bridged resource
  allocations backed by those discontiguous windows.
- Support stealing existing register window allocations to fulfill indirect
  bhnd(4) bus I/O requests within address ranges tagged with
  BHNDB_ALLOC_FULFILL_ON_OVERCOMMIT.
- Fix an inverted test of bhndb_is_pcie_attached() that disabled
  PCI-only clock bring-up required by these devices.

Approved by:	adrian (mentor, implicit)
Sponsored by:	The FreeBSD Foundation
2017-11-28 00:12:14 +00:00
landonf
68a611c92f siba(4): fix workaround for BCM4312 D11 core backplane timeouts
The SIBA_QUIRK_PCIE_D11_SB_TIMEOUT quirk should match on all BCM4312
revisions, and backplane service timeouts must also be disabled.

Approved by:	adrian (mentor, implicit)
Sponsored by:	The FreeBSD Foundation
2017-11-27 22:38:44 +00:00