Commit Graph

258217 Commits

Author SHA1 Message Date
Charlie Root
8e73e73197 improve best of 2 (still slow) 2019-08-23 15:30:30 -04:00
Oscar Tsalapatis
b4d0670a1f Move KQ scheduler flags to Non kernel 2019-08-02 16:34:03 -04:00
Oscar Zhao
2fdea6945b remove kq refcnt 2019-07-29 23:49:15 -04:00
Oscar Zhao
cb4c673500 WS + BON 2019-04-19 15:33:07 -04:00
Oscar Zhao
cb22931bab fix race in sleep 2019-04-05 15:23:07 -04:00
Charlie Root
b2e5289a2d kq domain + queue cache 2019-04-01 21:35:09 -04:00
Oscar Zhao
eb525254bc Brutal test 2019-03-14 03:49:26 -04:00
BuildTools
0d1463d912 inital commit 2019-03-05 01:55:45 -05:00
hselasky
23be74a4d3 Add new USB quirk.
PR:			235897
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2019-02-22 21:57:27 +00:00
bwidawsk
4442c99b63 nvdimm: Simple namespace support
Add support for simple NVDIMM v1.2 namespaces from the UEFI
version 2.7 specification. The combination of NVDIMM regions and
labels can lead to a wide variety of namespace layouts. Here we
support a simple subset of namespaces where each NVDIMM SPA range
is composed of a single region per member dimm.

Submitted by:	D Scott Phillips <d.scott.phillips@intel.com>
Discussed with:	kib
MFC after:	1 week
Sponsored by:	Intel Corporation
Differential Revision:	https://reviews.freebsd.org/D18736
2019-02-22 19:54:28 +00:00
bwidawsk
5063ce2b9b nvdimm: Read NVDIMM namespace labels
When attaching to NVDIMM devices, read and verify the namespace
labels from the special namespace label storage area. A later
change will expose NVDIMM namespaces derived from this label data.

Submitted by:	D Scott Phillips <d.scott.phillips@intel.com>
Discussed with:	kib
MFC after:	1 week
Sponsored by:	Intel Corporation
Differential Revision:	https://reviews.freebsd.org/D18735
2019-02-22 19:54:24 +00:00
bwidawsk
badbd8283d nvdimm: split spa dev into a separate entity
Separate code for exposing a device backed by a system physical
address range away from the NVDIMM spa code. This will allow a
future patch to add support for NVDIMM namespaces while using the
same device code.

Submitted by:	D Scott Phillips <d.scott.phillips@intel.com>
Reviewed by:	bwidawsk
MFC after:	1 week
Sponsored by:	Intel Corporation
Differential Revision:	https://reviews.freebsd.org/D18736
2019-02-22 19:54:21 +00:00
dab
c782cfe37d CID 1332000: Logically dead code in sys/dev/pms/RefTisa/tisa/sassata/sas/ini/itdio.c
A pointer is first tested for NULL. If non-NULL, another pointer is
set equal to the first. The second pointer is then checked for NULL
and an error path taken if so. This second test and the associated
path is dead code as the pointer value, having just been checked for
NULL, cannot be NULL at this point. Remove the dead code.

Reported by:	Coverity
Reviewed by:	daniel.william.ryan_gmail.com, vangyzen
Sponsored by:	Dell EMC Isilon
Differential Revision:	https://reviews.freebsd.org/D19165
2019-02-22 18:43:27 +00:00
dab
5a174e4632 Fix several Coverity-detected issues in newsyslog.
- CID 1394815, CID 1305673: Dereference before null check - memory was
  allocated and the allocation checked for NULL with a call to errx()
  if it failed. Code below that was guaranteed that the pointer was
  non-NULL, but there was another check for NULL at the exit of the
  function (after the memory had already been referenced). Eliminate
  the useless NULL check.

- CID 1007452: Resource leak - Storage intended to be allocated and
  returned to the caller was never freed. This was the result of a
  regression in the function signature introduced in r208648 (2010)
  (thanks for that find, @cem!). Fixed by altering the function
  signature and passing the allocated memory to the caller as
  intended. This also fixes PR158794.

- CID 1008620: Logically dead code in newsyslog.c - This was a direct
  result of CID 1007452. Since the memory allocated as described there
  was not returned to the caller, a subsequent check for the memory
  having been allocated was dead code. Returning the memory
  re-animates the code that is the subject of this CID.

- CID 1006131: Unused value - in parsing a configuration file, a
  pointer to the end of the last field was saved, but not used after
  that. Rewrite to use the pointer value. This could have been fixed
  by avoiding the assignment altogether, but this solutions more
  closely follows the pattern used in the preceding code.

PR:		158794
Reported by:	Coverity, Ken-ichi EZURA <k.ezura@gmail.com> (PR158794)
Reviewed by:	cem, markj
MFC after:	1 week
Differential Revision:	https://reviews.freebsd.org/D19105
2019-02-22 15:31:50 +00:00
imp
90eaca25c1 Rework logpage extensibility.
Move from using a linker set to a constructor function that's
called. This simplifies the code and is slightly more obvious.  We now
keep a list of page decoders rather than having an array we managed
before. Commands will move to something similar in the future.

Reviewed by: jhb@
Differential Revision: https://reviews.freebsd.org/D19275
2019-02-22 15:15:36 +00:00
dab
b9af2aaaa3 Revert r344389 & r343906 - Fix Coverity errors in newsyslog
It was pointed out that a couple of the "memory leak" CIDs that I
fixed were arguably Coverity errors rather than errors in the
newsyslog code and the cure was worse than the disease. Revert both
changes. The first change, which included fixes for other Coverity
errors, will be re-worked to omit the troublesome changes and then
re-committed with the remaining fixes.

Reported by:	bde
Sponsored by:	Dell EMC Isilon
2019-02-22 14:59:40 +00:00
bde
1e1408a8e7 Quick fix for building LINT on i386. A fix is needed on all arches and
this one should also work on amd64 and sparc64.

LINT was broken in r312910 with the removal of pc98 support, by changing
the pathname in UKBD_DFLT_KEYBAP from a removed pc98 file to a nonexistent
file.

There are many bugs nearby.  Some are:
- the error is not properly detected and handled by make(1), because
  kbdcontrol(8) exits with status 0 after failing to find the keymap file
- UKBD_DFLT_KEYBAP is supposed to be MI, and is in MI NOTES to try enforce
  this, but 5 out of 8 arches don't support it
- LINT seems to have been broken by this in only 7 out of 8 arches.  mips
  breaks test coverage instead, by killing this option in its MD NOTES.
  arm kills ukbd but that is not enough to configure an unsupported option
  used only by ukbd.
2019-02-22 11:52:40 +00:00
bde
559f0dfc7a Oops, actually fix warnings from -Wpointer-sign. 2019-02-22 07:00:23 +00:00
bde
426c5cade7 Fix warnings from -Wpointer-sign. 2019-02-22 06:57:46 +00:00
bde
d68a4690a6 Connect the restored dumb and sc terminal emulators to the kernel build.
Add or fix options to control static and dynamic configuration.  Keep
the default of scteken, but default to statically configuring all available
emulators (now 3 instead of 1).

The dumb emulator is almost usable.  libedit and libreadline handle
dumb terminals perfectly for at least shell history.  less(1) works
as well as possible except on exit.  But curses programs make messes.
The dumb emulator has strange color support, with 2 dumb colors for
normal output but fancy colorization for the cursor, mouse pointer and
(with a non-dumb initial emulator) for low-level console output.

Using the sc emulator instead of the default of scteken fixes at least
the following bugs:
- NUL is a printing character in cons25 but not in teken
- teken doesn't support fixed colors for "reverse" video.
- The best versions of sc are about 10 times faster than scteken (for
  printing to the frame buffer).  This version is only about 5 times
  faster.

Fix configuration features:
- make SC_DFLT_TERM (for setting the initial emulator) a normal option.

Add configuration features:
- negative options SC_NO_TERM_* for omitting emulators in the static config.
  Modules for emulators might work, but I don't know of any
- vidcontrol -e shows the available emulators
- vidcontrol -E <emulator> sets the active emulator.
2019-02-22 06:41:47 +00:00
ganbold
aedcdb3466 Add base to the warning threshold. 2019-02-22 03:11:27 +00:00
markj
04f9fc1a7a Commit a missing piece of r344452.
MFC with:	r344452
2019-02-21 22:56:54 +00:00
markj
57c4e2d807 Fix a tracepoint lookup race in fasttrap_pid_probe().
fasttrap hooks the userspace breakpoint handler; the hook looks up the
breakpoint address in a hash table of tracepoints.  It is possible for
the tracepoint to be removed by a different thread in between the
breakpoint trap and the hash table lookup, in which case SIGTRAP gets
delivered to the target process.  Fix the problem by adding a
per-process generation counter that gets incremented when a tracepoint
belonging to that process is removed.  Then, when a lookup fails, the
trapping instruction is restarted if the thread's counter doesn't match
that of the process.

Reviewed by:	cem
MFC after:	2 weeks
Sponsored by:	The FreeBSD Foundation
Differential Revision:	https://reviews.freebsd.org/D19273
2019-02-21 22:54:17 +00:00
sef
81d4ad979d Fix the usage error I introduced in r344192.
Specifically, I put the new option line in the wrong place, and then fixed
up the rest without realizing it.  This puts the usage statement back to
what it was, with an additional line for the new -V option.

Reported by:	mav
Sponsored by:	iXsystems Inc.
2019-02-21 22:49:39 +00:00
markj
af1684d0fe Disconnect fasttrap from the 32-bit powerpc build.
An upcoming bug fix requires 64-bit atomics, which aren't implemented on
powerpc.  The powerpc port of fasttrap is incomplete anyway and doesn't
get loaded by dtraceall.ko on powerpc because of a missing dependency;
it's presumed that it's effectively unused.

Discussed with:	jhibbits
MFC after:	2 weeks
2019-02-21 22:49:21 +00:00
emaste
55cc6970c9 scp: validate filenames provided by server against wildcard in client
OpenSSH-portable commits:

check in scp client that filenames sent during remote->local directory
copies satisfy the wildcard specified by the user.

This checking provides some protection against a malicious server
sending unexpected filenames, but it comes at a risk of rejecting wanted
files due to differences between client and server wildcard expansion rules.

For this reason, this also adds a new -T flag to disable the check.

reported by Harry Sintonen
fix approach suggested by markus@;
has been in snaps for ~1wk courtesy deraadt@

OpenBSD-Commit-ID: 00f44b50d2be8e321973f3c6d014260f8f7a8eda

Minor patch conflict (getopt) resolved.

Obtained from: OpenSSH-portable 391ffc4b9d31fa1f4ad566499fef9176ff8a07dc

scp: add -T to usage();

OpenBSD-Commit-ID: a7ae14d9436c64e1bd05022329187ea3a0ce1899

Obtained from: OpenSSH-portable 2c21b75a7be6ebdcbceaebb43157c48dbb36f3d8

PR:		234965
Approved by:	des
MFC after:	3 days
Obtained from:	OpenSSH-portable 391ffc4b9d, 2c21b75a7b
Sponsored by:	The FreeBSD Foundation
Differential Revision:	https://reviews.freebsd.org/D19076
2019-02-21 22:45:54 +00:00
jkim
55fd5feec9 MFV: r344447
Fix missing comma in array declaration.
2019-02-21 21:33:27 +00:00
jkim
0ac12a8077 Catch up with ACPICA 20190215.
PR:		235922
2019-02-21 21:15:49 +00:00
bde
543fd0f706 Fix the dumb and sc terminal emulators to compile and work.
First remove ifdefs of the unsupported option SC_DUMB_TERMINAL which
prevented building using both in the same kernel and broke regression
tests.  This option will be replaced by per-emulator supported options.

The dumb emulator rotted with KSE in r83366, but usually compiled since
it is ifdefed to nothing unless SC_DUMB_TERMINAL is defined.  The type
of an unused function parameter changed.

Both emulators rotted when 2 new methods were added while the emulators
were removed.  Only null methods are needed, but null function pointers
give panics instead.

The wildcard in the default for the unsupported option SC_DFLT_TERM
never really worked.  It tends to prefer the dumb emulator when multiple
emulators are configured.  Change it to prefer scteken for compatibility.
2019-02-21 19:19:30 +00:00
dim
992b29a96b Pull in r353299 from upstream lld trunk (by George Rimar):
Recommit r353293 "[LLD][ELF] - Set DF_STATIC_TLS flag for i386 target."

  With the following changes:
  1) Compilation fix:
  std::atomic<bool> HasStaticTlsModel = false; ->
  std::atomic<bool> HasStaticTlsModel{false};

  2) Adjusted the comment in code.

  Initial commit message:

  DF_STATIC_TLS flag indicates that the shared object or executable
  contains code using a static thread-local storage scheme.

  Patch checks if IE/LE relocations were used to check if the code uses
  a static model. If so it sets the DF_STATIC_TLS flag.

  Differential revision: https://reviews.llvm.org/D57749

Pull in r353378 from upstream lld trunk (by George Rimar):

  [LLD][ELF] - Set DF_STATIC_TLS flag for X64 target

  This is the same as D57749, but for x64 target.

  "ELF Handling For Thread-Local Storage" p41 says
  (https://www.akkadia.org/drepper/tls.pdf):
  R_X86_64_GOTTPOFF relocation is used for IE TLS models.
  Hence if linker sees this relocation we should add DF_STATIC_TLS flag.

  Differential revision: https://reviews.llvm.org/D57821

This adds support to lld for the DF_STATIC_TLS flag in shared objects,
which signals to the dynamic linker that the shared object requires
static thread local storage.

See also:	https://reviews.freebsd.org/D19072
MFC after:	1 week
2019-02-21 18:41:41 +00:00
bde
32731864c4 Move scterm_teken.c from 6 MD files lists to the MI files list so that it
is easier to configure.  It is MI, unlike some of the other syscons files
already in the MI list.

Move scvtb.c similarly.  It is needed whenever sc is configured, and is
more MI than most of the files already in the MI list.

This only changes the combined list for arm64 and mips.  These arches
already cannot build sc or even NOTES.
2019-02-21 17:31:33 +00:00
mav
df1afa210a Fix few issues in ioat(4) driver.
- Do not explicitly count active descriptors.  It allows hardware reset
to happen while device is still referenced, plus simplifies locking.
 - Do not stop/start callout each time the queue becomes empty.  Let it
run to completion and rearm if needed, that is much cheaper then to touch
it every time, plus also simplifies locking.
 - Decouple submit and cleanup locks, making driver reentrant.
 - Avoid memory mapped status register read on every interrupt.
 - Improve locking during device attach/detach.
 - Remove some no longer used variables.

Reviewed by:	cem
MFC after:	1 week
Sponsored by:	iXsystems, Inc.
Differential Revision:	https://reviews.freebsd.org/D19231
2019-02-21 16:47:36 +00:00
markj
00ee058700 Clear pointers to indicate that the respective locks are released.
This fixes a problem in r344231: vm_pageout_launder() may scan two
queues when swap is disabled.

Reported by:	pho
MFC with:	r344231
2019-02-21 15:44:32 +00:00
emaste
0514659c04 svn: support building with WITH_PIE
Subversion builds and links against its own .a archives using local
rules, so did not benefit from with the WITH_PIE library support added
in r344179.  Apply the same _pie suffix locally.

Reviewed by:	cem
Differential Revision:	https://reviews.freebsd.org/D19246
2019-02-21 15:08:25 +00:00
mav
621218a1ed Allow I/OAT of present Xeon E5/E7 to work thorugh PLX NTB.
Its a hack, we can't know/list all DMA engines, but this covers all
I/OAT of Xeon E5/E7 at least from Sandy Bridge till Skylake I saw.

MFC after:	1 week
Sponsored by:	iXsystems, Inc.
2019-02-21 14:10:14 +00:00
tuexen
e763f31429 The receive buffer autoscaling for TCP is based on a linear growth, which
is acceptable in the congestion avoidance phase, but not during slow start.
The MTU is is also not taken into account.
Use a method instead, which is based on exponential growth working also in
slow start and being independent from the MTU.

This is joint work with rrs@.

Reviewed by:		rrs@, Richard Scheffenegger
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D18375
2019-02-21 10:35:32 +00:00
andrew
27723fd7ce Allow the kcov buffer to be mmaped multiple times.
After r344391 this restriction is no longer needed.

Sponsored by:	DARPA, AFRL
2019-02-21 10:11:15 +00:00
andrew
ccea7c15ea Use KCOV_ENTRY_SIZE for the entry size.
Previously it was sizeof(uint64_t). While this is currently true, it may
not be on all future architectures.

Sponsored by:	DARPA, AFRL
2019-02-21 09:43:14 +00:00
tuexen
192c95e996 This patch addresses an issue brought up by bz@ in D18968:
When TCP_REASS_LOGGING is defined, a NULL pointer dereference would happen,
if user data was received during the TCP handshake and BB logging is used.

A KASSERT is also added to detect tcp_reass() calls with illegal parameter
combinations.

Reported by:		bz@
Reviewed by:		rrs@
MFC after:		3 days
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D19254
2019-02-21 09:34:47 +00:00
bde
61c06858a1 Restore syscons' terminal emulators. The trivial fixes to make them compile
will be committed later.

The "sc" emulator has the advantages of full support for cons25 and running
about 8 times faster than teken (for writing to the frame buffer).

The "dumb" emulator has the advantage of being simple.

Runtime choice of the emulator is good, but compile time choice is bad.
2019-02-21 08:37:39 +00:00
cem
0be3f7847b fuse: Fix a regression introduced in r337165
On systems with non-default DFLTPHYS and/or MAXBSIZE, FUSE would attempt to
use a buf cache block size in excess of permitted size.  This did not affect
most configurations, since DFLTPHYS and MAXBSIZE both default to 64kB.
The issue was discovered and reported using a custom kernel with a DFLTPHYS
of 512kB.

PR:		230260 (comment #9)
Reported by:	ken@
MFC after:	π/𝑒 weeks
2019-02-21 02:41:57 +00:00
sef
0a594400b5 * Handle SIGPIPE in gssd
We've got some cases where the other end of gssd's AF_LOCAL socket gets
closed, resulting in an error (and SIGPIPE) when it tries to do I/O to it.
Closing without cleaning up means the next time nfsd starts up, it hangs,
unkillably; this allows gssd to handle that particular error.

* Limit the retry cound in gssd_syscall to 5.
The default is INT_MAX, which effectively means forever.  And it's an
uninterruptable RPC call, so it will never stop.

The two changes mitigate the problem.

Reviewed by:	macklem
MFC after:	2 weeks
Sponsored by:	iXsystems Inc.
Differential Revision:	https://reviews.freebsd.org/D19153
2019-02-21 01:30:37 +00:00
jkim
67c6829380 MFV: r344395
Import ACPICA 20190215.
2019-02-20 23:53:39 +00:00
andrew
2a20bb1f08 Unwire the kcov buffer when freeing the info struct.
Without this the physical memory will not be returned to the kernel.

While here call vm_object_reference on the object when mmapping the buffer.
This removed the need for buggy tracking of if it has been mapped or not.

This fixes issues where kcov could use all the system memory.

Reported by:	tuexen
Reviewed by:	kib
Sponsored by:	DARPA, AFTL
Differential Revision:	https://reviews.freebsd.org/D19252
2019-02-20 22:41:14 +00:00
andrew
fc60fcc79e Call pmap_qenter for each page when creating the kcov buffer.
This removes the need to allocate a buffer to hold the vm_page_t objects
at the cost of extra IPIs on some architectures.

Reviewed by:	kib
Sponsored by:	DARPA, AFRL
Differential Revision:	https://reviews.freebsd.org/D19252
2019-02-20 22:32:28 +00:00
dab
229dccf125 Complete fix for CID 1007454, CID 1007453: Resource leak in newsyslog
The result of a strdup() was stored in a global variable and not freed
before program exit. This is a follow-up to r343906. That change
attempted to plug these resource leaks but managed to miss a code path
on which the leak still occurs. Plug the leak on that path, too.

MFC after:	3 days
Sponsored by:	Dell EMC Isilon
2019-02-20 22:05:44 +00:00
sef
5e2c80229c It turns out that setting the IV length is necessary with CCM in OpenSSL.
This adds that back.

Reviewed by:	cem
2019-02-20 21:24:56 +00:00
tsoome
f6f04d1a3a loader: really fix cd9660 dirmatch
The cd9660_open() does pass whole path to dirmatch() and we need to
compare only the current path component, not full path.

Additinally, skip over duplicate / (if any) and check if the last component
in the path was meant to be directory (having trailing /). If it is in fact
a file, error out.
2019-02-20 21:07:09 +00:00
dim
bb9abf23e4 Pull in r354515 from upstream libc++ trunk:
Fix the build with gcc when `-Wredundant-decls` is passed

  Summary:

  gcc warns that `__throw_runtime_error` is declared both in
  `<__locale>` and `<stdexcept>`, if `-Wredundant-decls` is passed on
  the command line; this is the case with FreeBSD when ${WARNS} == 6.

  Since `<__locale>` gets its first declaration via a transitive
  include of `<stdexcept>`, and the second declaration is after the
  first invocation of `__throw_runtime_error`, delete that second
  declaration.

  Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>

  Reviewers: kristina, MaskRay, EricWF, ldionne, ngie

  Reviewed By: EricWF

  Subscribers: krytarowski, brooks, emaste, dim, christof, jdoerfert, libcxx-commits

  Tags: #libc

  Differential Revision: https://reviews.llvm.org/D58425

Submitted by:	ngie
MFC after:	3 days
2019-02-20 21:06:11 +00:00
mmacy
700097c0a1 PFS: Bump NAMELEN and don't require clients to be sleepable
- debugfs consumers expect to be able to export names more than 48 characters

- debugfs consumers expect to be able to hold locks across calls and are able
  to handle allocation failures

Reviewed by:	hps@
MFC after:	1 week
Sponsored by:	iX Systems
Differential Revision:	https://reviews.freebsd.org/D19256
2019-02-20 20:55:02 +00:00