Commit Graph

214812 Commits

Author SHA1 Message Date
John Baldwin
847bfa8eb7 Use the port device name for the iov device for Chelsio T4/T5 cards.
Chelsio T4/T5 adapters are multifunction cards.  The main driver uses
physical function 4 (PF4).  However, VF devices for SR-IOV are only
supported on physical functions 0 through 3, where PF0 creates VFs tied
to port 0, etc.  The t4iov/t5iov driver was previously added to
create VF devices for ports that are present on each adapter.  This
change uses the recently added pci_iov_attach_name() function to
name the character device in /dev/iov after the associated port on
the card (e.g. /dev/iov/cxl0 is used to create VFs that share the
cxl0 port).  With this in place, mark the t4iov/t5iov devices quiet
to prevent them from cluttering dmesg.

Reviewed by:	rstone
Sponsored by:	Chelsio Communications
Differential Revision:	https://reviews.freebsd.org/D7402
2016-08-03 17:11:08 +00:00
John Baldwin
0aee83cc1d Permit the name of the /dev/iov entry to be set by the driver.
The PCI_IOV option creates character devices in /dev/iov for each PF
device driver that registers support for creating VFs.  By default the
character device is named after the PF device (e.g. /dev/iov/foo0).
This change adds a variant of pci_iov_attach() called pci_iov_attach_name()
that allows the name of the /dev/iov entry to be specified by the
driver.

Reviewed by:	rstone
MFC after:	1 month
Sponsored by:	Chelsio Communications
Differential Revision:	https://reviews.freebsd.org/D7400
2016-08-03 17:09:12 +00:00
John Baldwin
1f095f7051 Apply the fix from r232612 to fixed function counters.
Reviewed by:	emaste
MFC after:	1 month
Differential Revision:	https://reviews.freebsd.org/D7397
2016-08-03 16:52:00 +00:00
Dag-Erling Smørgrav
af8ee1391d Disable DSA again.
MFC after:	3 days
2016-08-03 16:34:20 +00:00
Pedro F. Giffuni
b4939677e8 indent(1): accept offsetof(3) as a keyword.
Reference:
c470e5e2c9

Differential Revision: https://reviews.freebsd.org/D6966  (Partial)
Submitted by:	Piotr Stefaniak
2016-08-03 16:33:34 +00:00
Dag-Erling Smørgrav
9ded33068e Remove DSA from default cipher list and disable SSH1.
Upstream did this a long time ago, but we kept DSA and SSH1 in FreeBSD for
reasons which boil down to POLA.  Now is a good time to catch up.

MFC after:	3 days
Relnotes:	yes
2016-08-03 16:08:21 +00:00
Konstantin Belousov
ad600ac8e3 Remove ncl_printf(), use printf(9) directly. After r303710 the
function duplicates printf().

Correct function names in the messages [*].

Noted by:	bde [*]
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2016-08-03 15:58:20 +00:00
John Baldwin
2de70600fa Correct assertion on vcpuid argument to vm_gpa_hold().
PR:		208168
Submitted by:	Dave Cameron <daverabbitz@ihug.co.nz>
Reviewed by:	grehan
MFC after:	1 month
2016-08-03 15:20:10 +00:00
Konstantin Belousov
fa03524a9f Merge i386 and amd64 variants of mp_watchdog.c into x86/, there is no
difference between files.
For pc98, put x86/mp_x86.c into the same place as used by i386 file list.
Fix typo in comment.

Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2016-08-03 13:51:53 +00:00
Konstantin Belousov
83d7cf21ea Remove unneeded (recursing) Giant acquisition around vprintf(9).
Reviewed by:	rmacklem
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2016-08-03 11:49:17 +00:00
Mateusz Guzik
0453ade508 locks: fix sx compilation on mips after r303643
The kernel.h header is required for the SYSINIT macro, which apparently
was present on amd64 by accident.

Reported by:	kib
2016-08-03 09:15:10 +00:00
Andrey A. Chernov
20e37fa8ca Although the code amount is not big, move POSIX error processing into
two sepatate functions to make glob(3) code less obscure and more simple.
There is no needs to make them inline since it is error path which supposed
to not happes often.
2016-08-03 09:09:34 +00:00
Konstantin Belousov
29ffb32ccd Remove Giant asserts. Update comment.
Owning Giant in the init/uninit is accidental due to the moment where
VFS modules initialization is performed, and is not enforced by the
VFS interface.  The Giant lock does not prevent a parallel execution
of the code, it is VFS which implements the proper protocol.

Approved by:	des (pseudofs maintainer)
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2016-08-03 08:57:15 +00:00
Konstantin Belousov
6828ba639a Some style changes. Fix a typo in comment.
Approved by:	des (pseudofs maintainer)
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2016-08-03 08:53:29 +00:00
Konstantin Belousov
0c657d22eb Explain why swapgeom_close_ev() is delegated.
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2016-08-03 07:11:19 +00:00
Konstantin Belousov
e69ba32f88 Remove mention of the Giant from the fork_return() description.
Making emphasis on this lock in the core function comment is confusing
for the modern kernel.

Sponsored by:	The FreeBSD Foundation
MFC after:	3 days
2016-08-03 07:10:09 +00:00
Ed Schouten
e938ebbc0c Regenerate system call tables for r303699 and r303700. 2016-08-03 06:36:45 +00:00
Ed Schouten
1b42875d4b Re-add traling slash that was removed in r303699.
I must have accidentally pressed some random key in vim.
2016-08-03 06:35:58 +00:00
Ed Schouten
a813fdc6c3 mprotect(): Change prototype to comply to POSIX.
Our mprotect() function seems to take a "const void *" address to the
pages whose permissions need to be adjusted. POSIX uses "void *". Simply
stick to the POSIX one to prevent us from writing unportable code.

PR:		211423 (exp-run)
Tested by:	antoine@ (Thanks!)
2016-08-03 06:33:04 +00:00
Justin Hibbits
fc5b6d7242 Remove a duplicate PMC CPU number
CPU number for MPC85XX is a duplicate for E500.  Since we don't support MPC85XX
"uncore" registers right now, rather than renumbering it simply remove it, and
it will properly use the E500 CPU ID.
2016-08-03 01:46:55 +00:00
Justin Hibbits
6cedae09a2 Merge MPC85XX and QorIQ config options
Summary:
MPC85XX and QorIQ are very similar.  When the DPAA dTSEC driver was
added, QORIQ_DPAA was brought in as a config option to support the differences
in hardware register settings between QorIQ (e500mc-, e5500- based) SoCs and
QUICC (e500v1/e500v2-based) SoCs, particularly in the Local Access Window (LAW)
target settings.

Unify these settings using macros to hide details and ease porting, and use a
new function (mpc85xx_is_qoriq()) to distinguish between QorIQ and QUICC SoCs at
runtime.

An alternative to using the function could be to use a variable initialized at
platform attach time, which may incur less overhead at runtime.  Since it's not
in the critical path once booted, this optimization doesn't seem necessary at
first pass.

Reviewed by: nwhitehorn
MFC after:	1 week
Differential Revision: https://reviews.freebsd.org/D7294
2016-08-03 01:22:11 +00:00
Navdeep Parhar
515b36c5b5 cxgbe/t4_tom: Read the chip's DDP page sizes and save them in a
per-adapter data structure.  This replaces a global array with hardcoded
page sizes.

Sponsored by:	Chelsio Communications
2016-08-02 23:54:21 +00:00
Bryan Drewery
a8d890b4fb Add link for getnetgrent_r(3).
MFC after:	3 days
Sponsored by:	EMC / Isilon Storage Division
2016-08-02 23:46:32 +00:00
Baptiste Daroussin
0f21f5285b truss: fix uninitialized trussinfo->curthread in add_threads()/enter_syscall
trussinfo->curthread must be initialized before calling enter_syscall(),
it is used by t->proc->abi->fetch_args().
Without that truss is segfaulting and the attached program also crash.

Submitted by:	Nikita Kozlov (nikita@gandi.net)
Reviewed by:	jhb
MFC after:	3 days
Differential Revision:	https://reviews.freebsd.org/D7399
2016-08-02 22:47:06 +00:00
Warner Losh
08ed5b80da tools/build looks for _WITH_GETLINE in /usr/include/stdio.h to see if
we need to include it in -legacy or not. Since the ifdef was removed,
this broke building 10.x and older source trees on -current. Restore
just enough of _WITH_GETLINE to allow these older source trees to
still build and properly omit getline() from their -legacy library.
2016-08-02 21:55:23 +00:00
Ed Maste
bfbeda91fe kgzip.8: note that it is deprecated and clarify BUGS section
kgzip is only useful on i386 and has a number of other limitations.
We're better served by officially deprecating it and encouraging users
to use loader(8) instead.

Reviewed by:	dteske, wblock
Differential Revision:	https://reviews.freebsd.org/D7098
2016-08-02 21:48:26 +00:00
Kristof Provost
72a3cf0f41 pfctl: Use const where possible.
This adds const qualifiers where it is possible.

Submitted by:	Christian Mauderer <christian.mauderer@embedded-brains.de>
2016-08-02 20:32:02 +00:00
Ed Maste
dd2be8cb6e Move/add ARM ELF PHDR types to elf_common.h
Accidentally missed in r303674
2016-08-02 20:26:04 +00:00
Dimitry Andric
00f060000f Fix a segfault in bsdgrep when parsing the invalid extended regexps "?"
or "+" (these are invalid, because there is no preceding operand).

When bsdgrep attempts to emulate GNU grep in discarding and ignoring the
invalid ? or + operators, some later logic in tre_compile_fast() goes
beyond the end of the buffer, leading to a crash.

Fix this by bailing out, and reporting a bad pattern instead.

Reported by:	Steve Kargl
MFC after:	1 week
2016-08-02 20:25:22 +00:00
Ed Maste
20136ffc7b readelf: report ARM program and section header types
Reviewed by:	andrew
Sponsored by:	The FreeBSD Foundation
Differential Revision:	https://reviews.freebsd.org/D7390
2016-08-02 20:11:04 +00:00
Kristof Provost
9a5e33be57 pfctl: Match prototype of pfctl_load_hostid.
The prototype and the implementation of the pfctl_load_hostid used a
different data type for one of the parameters.

Submitted by:	Christian Mauderer <christian.mauderer@embedded-brains.de>
2016-08-02 19:54:40 +00:00
Ed Maste
4df4e8ac47 Add ELFOSABI_ARM_AEABI ELF OSABI constant
Reported by:	andrew
Sponsored by:	The FreeBSD Foundation
2016-08-02 18:42:32 +00:00
Conrad Meyer
ce601a2676 proc_init: Fix a few memory leaks of 'phdl'
In the normal case and correct failure cases, the 'phdl' pointer is passed to
callers to use or clean up as needed.  However, some failure cases returned
early, failing to export the phdl pointer.

This was introduced in the restructuring of r303533.

Reported by:	Coverity
CID:		1361070
Reviewed by:	markj
Sponsored by:	EMC / Isilon Storage Division
2016-08-02 18:13:50 +00:00
Kristof Provost
0cd7a91aa7 pfctl: Allow TOS bits to be cleared
TOS value 0 is valid, so use 256 as an invalid value rather than zero.
This allows users to enforce TOS == 0 with pf.

Reported by:	Radek Krejča <radek.krejca@starnet.cz>
2016-08-02 15:41:42 +00:00
Pedro F. Giffuni
249a8a8003 sed(1): Revert r303047 "cleanup" and therefore r303572.
While big, the change was meant to have no effect on behavior and instead
so far we have found two regressions: one in the etcupdate tests and
another one in the games/openttd port[1].

Revert to a known working state. We will likely have to split the patch in
functional parts before bringing back the changes.

PR:		195929
Reported by:	danfe, madpilot [1]
2016-08-02 15:35:53 +00:00
Andrew Turner
27c9d42531 Remove trailing whitespace from the arm64 pmap
Obtained from:	ABT Systems Ltd
MFC after:	3 weeks
Sponsored by:	The FreeBSD Foundation
2016-08-02 15:26:46 +00:00
Ruslan Bukin
98f50c44e3 Update RISC-V port to Privileged Architecture Version 1.9.
Sponsored by:	DARPA, AFRL
Sponsored by:	HEIF5
2016-08-02 14:50:14 +00:00
Andrey V. Elsukov
723758b7ce Fix NULL pointer dereference.
ro pointer can be NULL when IPSec consumes mbuf.

PR:		211486
MFC after:	3 days
2016-08-02 12:18:06 +00:00
Sepherosa Ziehau
05cde7efa6 tcp/lro: Implement hash table for LRO entries.
This significantly improves HTTP workload performance and reduces
HTTP workload latency.

Reviewed by:	rrs, gallatin, hps
Obtained from:	rrs, gallatin
Sponsored by:	Netflix (rrs, gallatin) , Microsoft (sephe)
Differential Revision:	https://reviews.freebsd.org/D6689
2016-08-02 06:36:47 +00:00
Mateusz Guzik
fa5000a4f3 locks: fix compilation for KDTRACE_HOOKS && !ADAPTIVE_* case
Reported by:	Michael Butler <imb protected-networks.net>
2016-08-02 03:05:59 +00:00
Mateusz Guzik
0412689595 locks: fix up ifdef guards introduced in r303643
Both sx and rwlocks had copy-pasted ADAPTIVE_MUTEXES instead of the correct
define.

MFC after:	1 week
2016-08-02 00:15:08 +00:00
Conrad Meyer
9809e9dc3a rtentry: Initialize rt_mtx with MTX_NEW
The "rtentry" zone does not use UMA_ZONE_ZINIT, so it is invalid to assume the
mutex's memory will be zero.  Without MTX_NEW, garbage backing memory may
trigger the "re-initializing a mutex" assertion.

PR:		200991
Submitted by:	Chang-Hsien Tsai <luke.tw AT gmail.com>
2016-08-01 23:07:31 +00:00
Conrad Meyer
bf4c239e47 opencrypto AES-ICM: Fix heap corruption typo
This error looks like it was a simple copy-paste typo in the original commit
for this code (r275732).

PR:		204009
Reported by:	Chang-Hsien Tsai <luke.tw AT gmail.com>
Sponsored by:	EMC / Isilon Storage Division
2016-08-01 22:57:03 +00:00
Conrad Meyer
5f00f45775 Fix ddb "show proc" to show full arguments
PR:		200052
Submitted by:	Chang-Hsien Tsai <luke.tw AT gmail.com>
2016-08-01 22:41:50 +00:00
John Baldwin
315048f2ad Store the offset of the KDOORBELL and GTS registers in the softc.
VF devices use a different register layout than PF devices.  Storing
the offset in a value in the softc allows code to be shared between the
PF and VF drivers.

Reviewed by:	np
Sponsored by:	Chelsio Communications
Differential Revision:	https://reviews.freebsd.org/D7389
2016-08-01 22:39:51 +00:00
Mark Johnston
49fe7b5029 ipoib: Bound the number of egress mbufs buffered during pathrec lookups.
In pathological situations where the master subnet manager becomes
unresponsive for an extended period, we may otherwise end up queuing all
of the system's mbufs while waiting for a response to a path record lookup.

This addresses the same issue as commit 1e85b806f9 in Linux.

Reviewed by:	cem, ngie
MFC after:	2 weeks
Sponsored by:	EMC / Isilon Storage Division
2016-08-01 22:22:11 +00:00
John Baldwin
2611037c1c Disable PCI hotplug support for slots with power controllers.
After further review of the spec, I do not think the current HotPlug
code handles slots with power controllers correctly.  In particular,
the power state of the slot is to be inferred from other events, not
from examining the state of the power control bit in SLOT_CTL.  For now,
disable PCI hotplug support on such slots.

PR:		211081
Tested by:	Jeffrey E Pieper <jeffrey.e.pieper@intel.com>
MFC after:	3 days
2016-08-01 22:19:23 +00:00
Mateusz Guzik
1ada904147 Implement trivial backoff for locking primitives.
All current spinning loops retry an atomic op the first chance they get,
which leads to performance degradation under load.

One classic solution to the problem consists of delaying the test to an
extent. This implementation has a trivial linear increment and a random
factor for each attempt.

For simplicity, this first thouch implementation only modifies spinning
loops where the lock owner is running. spin mutexes and thread lock were
not modified.

Current parameters are autotuned on boot based on mp_cpus.

Autotune factors are very conservative and are subject to change later.

Reviewed by:	kib, jhb
Tested by:	pho
MFC after:	1 week
2016-08-01 21:48:37 +00:00
Sean Bruno
e72a746a6e r293331 mistakingly failed to add an assignment of paddr to the rxbuf
but only in the NETMAP code.  This lead to the NETMAP code paths
passing nothing up to userland.

Submitted by:	Ad Schellevis <ad@opnsense.org>
Reported by:	Franco Fichtner <franco@opnsense.org>
MFC after:	1 day
2016-08-01 21:19:51 +00:00
Andrey V. Elsukov
0428336393 Do not invoke resize event if initial disk size is zero. Some disks
report the size only after first opening.  And due to the events are
asynchronous, some consumers can receive this event too late and
this confuses them. This partially restores previous behaviour, and
at the same time this should fix the problem, when already opened
provider loses resize event.

PR:		211028
MFC after:	3 weeks
2016-08-01 20:54:54 +00:00