FreeBSD src
Go to file
Hans Petter Selasky 10c8755706 Fix for race leading to endless timer interrupts related to
configtimer().

During normal operation "state->nextcallopt" will always be less than
or equal to "state->nextcall" and checking only "state->nextcallopt"
before calling "callout_process()" is sufficient. However when
"configtimer()" is called a race might happen requiring both of these
binary times to be checked.

Short description of race:

1) A configtimer() call will reset both "state->nextcall" and
"state->nextcallopt" to the same binary time.

2) If a "callout_reset()" call happens between "configtimer()" and the
next "callout_process()" call, "state->nextcallopt" will get updated
and "state->nextcall" will remain at the current time. Refer to logic
inside cpu_new_callout().

3) getnextcpuevent() only respects "state->nextcall" and returns this
value over and over again, even if it is in the past, until "now >=
state->nextcallopt" becomes true. Then these two time variables are
corrected by a "callout_process()" call and the situation goes back to
normal.

The problem manifests itself in different ways. The common factor is
the timer process(es) consume all CPU on one or more CPU cores for a
long time, blocking other kernel processes from getting execution
time. This can be seen by very high interrupt counts as displayed by
"vmstat -i | grep timer" right after boot.

When EARLY_AP_STARTUP was enabled in r310177 the likelyhood of hitting
this bug apparently increased.

Example output from "vmstat -i" before patch:
cpu0:timer                          7591         69
cpu9:timer                      39031773     358089
cpu4:timer                          9359         85
cpu3:timer                          9100         83
cpu2:timer                          9620         88

Example output from "vmstat -i" after patch:
cpu0:timer                          4242         34
cpu6:timer                          5531         44
cpu3:timer                          6450         52
cpu1:timer                          4545         36
cpu9:timer                          7153         58

Before the patch cpu9 in the example above, was spinning in a loop in
order to reach 39 million interrupts just a few seconds after
bootup. After the patch the timer interrupt counts are more or less
consistent.

Discussed with:		mav @
Reported by:		several people
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-01-20 17:40:31 +00:00
bin chmod: Add SIGINFO handler 2017-01-08 06:50:53 +00:00
cddl Fix an unchecked return value in zfsd 2017-01-18 22:10:18 +00:00
contrib MFV r312333: zlib 1.2.11. 2017-01-17 05:55:47 +00:00
crypto Sync ^/vendor/NetBSD/tests/dist with upstream 2017-01-12 07:26:39 +00:00
etc Remove obsolete /usr/lib/debug/usr/lib/private dir 2017-01-20 03:14:18 +00:00
gnu Use SRCTOP-relative paths to other directories instead of .CURDIR-relative ones 2017-01-20 05:51:25 +00:00
include Commit more accepted upstream changes from <NetBSD>/tests/... 2017-01-14 02:26:46 +00:00
kerberos5 Conditionalize adding ${KRB5DIR}/lib/gssapi/krb5/gkrb5_err.et to ETSRCS 2017-01-02 19:03:01 +00:00
lib Mention sendfile(2) by popular demand. 2017-01-20 17:29:59 +00:00
libexec rtld: do not rely on a populated GOT on amd64 2017-01-16 14:49:29 +00:00
release Enable IPv6 networking on Amazon EC2. 2017-01-15 09:06:45 +00:00
rescue DIRDEPS_BUILD: Update dependencies. 2016-11-13 00:11:30 +00:00
sbin Fix build of devd with GCC 4.2 2017-01-19 16:59:55 +00:00
secure Conditionalize building libwrap support into sshd 2017-01-07 08:08:35 +00:00
share Refresh tmpfs(5) man page. 2017-01-19 18:26:06 +00:00
sys Fix for race leading to endless timer interrupts related to 2017-01-20 17:40:31 +00:00
targets Merge ^/head r308491 through r308841. 2016-11-19 16:05:55 +00:00
tests Import ACPICA 20170119. 2017-01-19 19:46:15 +00:00
tools Add a new socket option SO_TS_CLOCK to pick from several different clock 2017-01-16 17:46:38 +00:00
usr.bin Remove some unused code. 2017-01-20 16:01:01 +00:00
usr.sbin Remove ISCSI_MAX_DATA_SEGMENT_LENGTH, using negotiated value. 2017-01-20 17:14:10 +00:00
.arcconfig callsign isn't required anymore 2016-09-29 06:19:45 +00:00
.arclint phabricator related changes: 2015-04-20 20:33:22 +00:00
COPYRIGHT Bump copyright year. 2016-12-31 12:41:42 +00:00
LOCKS
MAINTAINERS Remove myself from kern_timeout.c yeah! 2016-07-27 20:37:32 +00:00
Makefile Add full softfloat and hardfloat support for RISC-V. 2016-11-16 15:21:32 +00:00
Makefile.inc1 Enable /usr/lib32 for o32 binaries on mips64. 2017-01-06 23:30:54 +00:00
Makefile.libcompat Enable /usr/lib32 for o32 binaries on mips64. 2017-01-06 23:30:54 +00:00
ObsoleteFiles.inc Fix typo from change 310985 in ObsoleteFiles.inc 2017-01-10 20:37:44 +00:00
README Vendor import of zlib 1.2.11. 2017-01-17 05:47:05 +00:00
UPDATING Deprecate kernel configuration option EM_MULTIQUEUE now that the em(4) 2017-01-12 14:38:18 +00:00

This is the top level of the FreeBSD source directory.  This file
was last revised on:
$FreeBSD$

For copyright information, please see the file COPYRIGHT in this
directory (additional copyright information also exists for some
sources in this tree - please see the specific source directories for
more information).

The Makefile in this directory supports a number of targets for
building components (or all) of the FreeBSD source tree.  See build(7)
and http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html
for more information, including setting make(1) variables.

The `buildkernel` and `installkernel` targets build and install
the kernel and the modules (see below).  Please see the top of
the Makefile in this directory for more information on the
standard build targets and compile-time flags.

Building a kernel is a somewhat more involved process.  See build(7), config(8),
and http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/kernelconfig.html
for more information.

Note: If you want to build and install the kernel with the
`buildkernel` and `installkernel` targets, you might need to build
world before.  More information is available in the handbook.

The kernel configuration files reside in the sys/<arch>/conf
sub-directory.  GENERIC is the default configuration used in release builds.
NOTES contains entries and documentation for all possible
devices, not just those commonly used.


Source Roadmap:
---------------

bin		System/user commands.

cddl		Various commands and libraries under the Common Development
		and Distribution License.

contrib		Packages contributed by 3rd parties.

crypto		Cryptography stuff (see crypto/README).

etc		Template files for /etc.

gnu		Various commands and libraries under the GNU Public License.
		Please see gnu/COPYING* for more information.

include		System include files.

kerberos5	Kerberos5 (Heimdal) package.

lib		System libraries.

libexec		System daemons.

release		Release building Makefile & associated tools.

rescue		Build system for statically linked /rescue utilities.

sbin		System commands.

secure		Cryptographic libraries and commands.

share		Shared resources.

sys		Kernel sources.

tests		Regression tests which can be run by Kyua.  See tests/README
		for additional information.

tools		Utilities for regression testing and miscellaneous tasks.

usr.bin		User commands.

usr.sbin	System administration commands.


For information on synchronizing your source tree with one or more of
the FreeBSD Project's development branches, please see:

  http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/synching.html