]> git.ipfire.org Git - thirdparty/chrony.git/log
thirdparty/chrony.git
15 years agoMultiply clientlog node table size when reallocating
Miroslav Lichvar [Tue, 24 Nov 2009 14:20:23 +0000 (15:20 +0100)] 
Multiply clientlog node table size when reallocating

15 years agoUse NULLs in select call instead of empty sets
Miroslav Lichvar [Mon, 23 Nov 2009 16:24:44 +0000 (17:24 +0100)] 
Use NULLs in select call instead of empty sets

15 years agoReduce adjtime calling
Miroslav Lichvar [Mon, 23 Nov 2009 16:23:43 +0000 (17:23 +0100)] 
Reduce adjtime calling

Don't call adjtime to determine remaining offset when there is no slewing
running.

15 years agoDon't read past buffer in find_ordered_entry_with_flags
Miroslav Lichvar [Wed, 18 Nov 2009 14:36:13 +0000 (15:36 +0100)] 
Don't read past buffer in find_ordered_entry_with_flags

15 years agoAvoid blocking read in rtc_linux
Miroslav Lichvar [Wed, 18 Nov 2009 11:53:31 +0000 (12:53 +0100)] 
Avoid blocking read in rtc_linux

15 years agoAdd spaces when catenating chronyc command line arguments
Miroslav Lichvar [Thu, 12 Nov 2009 15:48:09 +0000 (16:48 +0100)] 
Add spaces when catenating chronyc command line arguments

15 years agoUse non-zero exit code in chronyc to report errors
Miroslav Lichvar [Thu, 12 Nov 2009 15:43:34 +0000 (16:43 +0100)] 
Use non-zero exit code in chronyc to report errors

15 years agoImprove status checking and printing in chronyc
Miroslav Lichvar [Thu, 12 Nov 2009 14:36:12 +0000 (15:36 +0100)] 
Improve status checking and printing in chronyc

15 years agoCheck in chronyc that command in reply is same as requested
Miroslav Lichvar [Thu, 12 Nov 2009 14:06:40 +0000 (15:06 +0100)] 
Check in chronyc that command in reply is same as requested

15 years agoDon't set NTP source as reachable when reply doesn't have valid data
Miroslav Lichvar [Thu, 12 Nov 2009 11:40:45 +0000 (12:40 +0100)] 
Don't set NTP source as reachable when reply doesn't have valid data

This fixes using uninitialized sourcestats values when selecting source.

15 years agoAllow overriding system detection in configure
Miroslav Lichvar [Tue, 10 Nov 2009 15:25:59 +0000 (16:25 +0100)] 
Allow overriding system detection in configure

15 years agoDocument -4 and -6 options
Miroslav Lichvar [Mon, 9 Nov 2009 13:35:39 +0000 (14:35 +0100)] 
Document -4 and -6 options

15 years agoSupport LDFLAGS and CPPFLAGS in configure
Miroslav Lichvar [Wed, 4 Nov 2009 14:46:58 +0000 (15:46 +0100)] 
Support LDFLAGS and CPPFLAGS in configure

15 years agoAllow LDFLAGS to be used in linking
Jonathan Cameron [Wed, 4 Nov 2009 14:04:02 +0000 (15:04 +0100)] 
Allow LDFLAGS to be used in linking

15 years agoValidate sample times received in refclock
Miroslav Lichvar [Tue, 3 Nov 2009 15:40:12 +0000 (16:40 +0100)] 
Validate sample times received in refclock

15 years agoDon't forget to save last PPS sequence number
Miroslav Lichvar [Tue, 3 Nov 2009 15:36:02 +0000 (16:36 +0100)] 
Don't forget to save last PPS sequence number

15 years agoAvoid compiler warnings in util.c on 32-bit archs
Miroslav Lichvar [Fri, 30 Oct 2009 10:13:08 +0000 (11:13 +0100)] 
Avoid compiler warnings in util.c on 32-bit archs

15 years agoReplaced references to sunsite with references to tuxfamily.
John G. Hasler [Thu, 29 Oct 2009 21:14:39 +0000 (16:14 -0500)] 
Replaced references to sunsite with references to tuxfamily.

15 years agoUpdate COPYING and FSF address
Miroslav Lichvar [Wed, 28 Oct 2009 16:53:33 +0000 (17:53 +0100)] 
Update COPYING and FSF address

15 years agoUpdate copyrights
Miroslav Lichvar [Wed, 28 Oct 2009 16:53:10 +0000 (17:53 +0100)] 
Update copyrights

15 years agoUpdate documentation for refclock and IPv6 support
Miroslav Lichvar [Wed, 28 Oct 2009 15:53:03 +0000 (16:53 +0100)] 
Update documentation for refclock and IPv6 support

15 years agoRemove forgotten text in server directive description
Miroslav Lichvar [Wed, 28 Oct 2009 15:31:58 +0000 (16:31 +0100)] 
Remove forgotten text in server directive description

15 years agoAdd PPS API refclock driver
Miroslav Lichvar [Wed, 28 Oct 2009 11:23:00 +0000 (12:23 +0100)] 
Add PPS API refclock driver

15 years agoReselect source also according to distance
Miroslav Lichvar [Tue, 27 Oct 2009 13:28:19 +0000 (14:28 +0100)] 
Reselect source also according to distance

Reselect when a source with the same stratum is available and has
significantly better distance than the current source.

15 years agoMake default refclock refid from number of the source
Miroslav Lichvar [Tue, 27 Oct 2009 13:04:59 +0000 (14:04 +0100)] 
Make default refclock refid from number of the source

It should avoid having two or more refclocks with the same refid.

15 years agoAdd refclock option for delay
Miroslav Lichvar [Tue, 27 Oct 2009 13:02:16 +0000 (14:02 +0100)] 
Add refclock option for delay

This is useful when refclocks don't agree on time, increasing the
delay will widen the interval used in the source selection algorithm.

15 years agoReduce size of NTP sources hash table
Miroslav Lichvar [Tue, 13 Oct 2009 15:16:41 +0000 (17:16 +0200)] 
Reduce size of NTP sources hash table

IPv6 addressing significantly increased size of the table,
keep only pointers to get it back.

15 years agoInclude both refid and IP address in tracking and sourcestats reports
Miroslav Lichvar [Tue, 13 Oct 2009 14:04:24 +0000 (16:04 +0200)] 
Include both refid and IP address in tracking and sourcestats reports

ref_id is not sufficient for IPv6 addresses and ref_id is needed for
reference clocks.

15 years agoAdd -4 and -6 options to set address family when resolving names
Miroslav Lichvar [Tue, 13 Oct 2009 11:10:28 +0000 (13:10 +0200)] 
Add -4 and -6 options to set address family when resolving names

15 years agoAlways send timevals in cmdmon protocol in 64-bit format
Miroslav Lichvar [Fri, 9 Oct 2009 13:31:59 +0000 (15:31 +0200)] 
Always send timevals in cmdmon protocol in 64-bit format

This is to avoid incompatibility between 64/32-bit client/server.
While at it, convert all time values in the protocol to timeval
to avoid Y2K38 problem.

15 years agoAdd IPv6 support
Miroslav Lichvar [Fri, 9 Oct 2009 13:00:59 +0000 (15:00 +0200)] 
Add IPv6 support

15 years agoDon't use uninitialized values
Miroslav Lichvar [Wed, 7 Oct 2009 13:11:09 +0000 (15:11 +0200)] 
Don't use uninitialized values

This fixes a bunch or valgrind errors.

15 years agoDon't forget to free capability object
Miroslav Lichvar [Mon, 21 Sep 2009 10:45:33 +0000 (12:45 +0200)] 
Don't forget to free capability object

15 years agoDon't copy util functions in client.c
Miroslav Lichvar [Tue, 15 Sep 2009 16:08:58 +0000 (18:08 +0200)] 
Don't copy util functions in client.c

This requires moving croak() to logging.c and avoiding use
of log functions in util.c.

15 years agoFix string termination in refclock parameter parser
Miroslav Lichvar [Wed, 9 Sep 2009 15:12:43 +0000 (17:12 +0200)] 
Fix string termination in refclock parameter parser

16 years agoDon't lose remaining adjtime in initiate_slew
Miroslav Lichvar [Wed, 26 Aug 2009 15:58:57 +0000 (17:58 +0200)] 
Don't lose remaining adjtime in initiate_slew

initiate_slew is called also from set_frequency which doesn't read
the remaining adjtime. This wasn't a problem before commit 8c0f3f4
as offset_register was 0.0 and initiate_slew immediately returned.

16 years agoAdd SOCK refclock driver
Miroslav Lichvar [Wed, 19 Aug 2009 13:39:06 +0000 (15:39 +0200)] 
Add SOCK refclock driver

This adds a support for receiving samples over unix domain socket.
It's a better alternative to the SHM refclock, the resolution is not
limited to microseconds and it doesn't require polling.

16 years agoSwitch refclock driver parameter to string
Miroslav Lichvar [Wed, 19 Aug 2009 13:29:52 +0000 (15:29 +0200)] 
Switch refclock driver parameter to string

16 years agoAdd support for SO_TIMESTAMP control messages
Miroslav Lichvar [Fri, 7 Aug 2009 15:24:19 +0000 (17:24 +0200)] 
Add support for SO_TIMESTAMP control messages

SO_TIMESTAMP messages contain kernel time stamps for received packets,
this should improve accuracy and avoid the impact of CPU scheduling
latencies.

16 years agoSet reply source IP from query destination IP
Timo Teras [Fri, 7 Aug 2009 14:26:15 +0000 (16:26 +0200)] 
Set reply source IP from query destination IP

Currently, on multihomed host, when chrony is not bound to a specific
IP address, a query is sent to an interface and the default source IP
hint for the back route differs, the reply will have a source IP
different than where the query was destinied to. This will cause
problems because connection tracking firewalls will drop the replies
and most likely the client program will get confused too.

This patch uses the IP_PKTINFO mechanism to get the IP address where
received packets where targetted to and use that IP address as source
hint when sending a reply.

16 years agoSwitch to recvmsg/sendmsg to get access to control messages
Timo Teras [Fri, 7 Aug 2009 12:27:51 +0000 (14:27 +0200)] 
Switch to recvmsg/sendmsg to get access to control messages

16 years agoAdd editline support
Miroslav Lichvar [Fri, 17 Jul 2009 10:38:37 +0000 (12:38 +0200)] 
Add editline support

GNU readline recently changed license to GPLv3+ which makes it
incompatible with chrony (GPLv2). This patch adds support for editline
library (BSD license).

16 years agoFix compiler warnings in getdate.c
Miroslav Lichvar [Mon, 6 Jul 2009 09:49:26 +0000 (11:49 +0200)] 
Fix compiler warnings in getdate.c

16 years agoTry to minimize adjtime error
Miroslav Lichvar [Wed, 1 Jul 2009 13:49:05 +0000 (15:49 +0200)] 
Try to minimize adjtime error

The offset is rounded before calling adjtime and the error below
microsecond is accumulated.

16 years agoAdd median filter for refclocks
Miroslav Lichvar [Wed, 1 Jul 2009 13:40:50 +0000 (15:40 +0200)] 
Add median filter for refclocks

16 years agoAdd client support for Kiss-of-Death RATE
Miroslav Lichvar [Tue, 30 Jun 2009 15:47:11 +0000 (17:47 +0200)] 
Add client support for Kiss-of-Death RATE

This adds support for RATE code which can be used in reply from NTP
server to reduce client's polling.

16 years agoAdd support for ADJ_OFFSET_SS_READ mode
Miroslav Lichvar [Wed, 20 May 2009 15:43:09 +0000 (17:43 +0200)] 
Add support for ADJ_OFFSET_SS_READ mode

Also assume that kernels >= 2.6.27 don't need frequency scaling.

16 years agoAdd SHM refclock driver
Miroslav Lichvar [Tue, 5 May 2009 07:17:38 +0000 (09:17 +0200)] 
Add SHM refclock driver

16 years agoAdd support for reference clocks
Miroslav Lichvar [Tue, 5 May 2009 07:14:13 +0000 (09:14 +0200)] 
Add support for reference clocks

16 years agoMake sure skew isn't 0.0 to avoid segfaults
Miroslav Lichvar [Tue, 5 May 2009 06:52:33 +0000 (08:52 +0200)] 
Make sure skew isn't 0.0 to avoid segfaults

16 years agoDon't finalise from signal handler
Miroslav Lichvar [Wed, 29 Apr 2009 11:53:02 +0000 (13:53 +0200)] 
Don't finalise from signal handler

16 years agoReduce request timeout in chronyc
Miroslav Lichvar [Mon, 27 Apr 2009 14:43:36 +0000 (16:43 +0200)] 
Reduce request timeout in chronyc

16 years agoFix memlockall patch
Miroslav Lichvar [Wed, 11 Feb 2009 17:01:23 +0000 (18:01 +0100)] 
Fix memlockall patch

16 years agoFix compiler warnings in wrap_adjtimex.c
Miroslav Lichvar [Tue, 10 Feb 2009 17:06:25 +0000 (18:06 +0100)] 
Fix compiler warnings in wrap_adjtimex.c

16 years agoAdd mlockall and SCHED_FIFO support
John Hasler [Tue, 10 Feb 2009 16:59:57 +0000 (17:59 +0100)] 
Add mlockall and SCHED_FIFO support

The attached patch adds support for mlockall() as well as the SCHED_FIFO
real-time scheduler. It should result in reduced (and more consistent)
latency. Usage is documented in all the documents.

16 years agoFix leap sign in measurements log
Miroslav Lichvar [Wed, 31 Dec 2008 09:03:49 +0000 (10:03 +0100)] 
Fix leap sign in measurements log

16 years agoLeap second support
Miroslav Lichvar [Wed, 31 Dec 2008 08:59:20 +0000 (09:59 +0100)] 
Leap second support

Leap second status is accepted and forwarded to clients if majority
of selectable sources agree. The actual insertion/deletion is supported
only on Linux now.

16 years agoRetry name resolving after temporary failure few times before giving up
Miroslav Lichvar [Mon, 10 Nov 2008 13:42:41 +0000 (14:42 +0100)] 
Retry name resolving after temporary failure few times before giving up

This is a temporary fix to allow starting when resolv.conf is not ready yet
(e.g. when using NetworkManager). It may delay start up to 1022 seconds.

16 years agoFix resolving IP addresses into names on 64-bit big-endian machines
Miroslav Lichvar [Thu, 6 Nov 2008 16:35:22 +0000 (17:35 +0100)] 
Fix resolving IP addresses into names on 64-bit big-endian machines

16 years agoLinux capabilities support
Miroslav Lichvar [Wed, 5 Nov 2008 23:50:48 +0000 (23:50 +0000)] 
Linux capabilities support

Attached is a patch adding a linux capabilities support to chronyd. It
adds -u option which can be used to specify the user which chronyd
should switch to.

16 years agoFix errors detected by valgrind
Miroslav Lichvar [Wed, 5 Nov 2008 23:48:58 +0000 (23:48 +0000)] 
Fix errors detected by valgrind

I tried running chronyd in valgrind and the result was that there are four
places where memory is not initialized. A patch fixing the errors is in the
attachment.

16 years agoFix fault where chronyd enters an endless loop on x86_64
John Hasler [Tue, 29 Apr 2008 17:40:15 +0000 (12:40 -0500)] 
Fix fault where chronyd enters an endless loop on x86_64

John writes:
Here is a patch that should prevent the endless loop.  I've changed
UTI_NormaliseTimeval() to use divide/remainder instead of a loop.  It also
replaces some similar loops with calls to UTI_NormaliseTimeval() and fixes
an unrelated bug in UTI_DiffTimevals().

17 years agoFix IP addressing in chronyc
Thomas Zajic [Tue, 29 Jul 2008 22:35:42 +0000 (23:35 +0100)] 
Fix IP addressing in chronyc

Thomas wrote:
I found a bug in the chrony client (chronyc) that affects its ability to talk
to remote hosts over the control port (323/udp).

For example, running "chronyc -h 192.168.1.3 sources -v" would just sit there
and hang, and eventually timeout. I found out with tcpdump that chronyc
actually tries to connect to 255.168.1.3 instead of 192.168.1.3.

17 years agoFix for chronyc "sources" command on 64 bit machines
Goswin Brederlow [Sat, 29 Mar 2008 20:49:59 +0000 (20:49 +0000)] 
Fix for chronyc "sources" command on 64 bit machines

(Taken from
  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=348412
)

Attached is a patchlet to make the "sources" command of chrony output properly
signed numbers. The chronyd code (see e.g. ntp.h) properly uses int32_t and
friends to get the right number of bits per datatype while client.c just uses
short, int, long. But long will be 64 bit or 32 bit depending on the cpu.

17 years agoUpdate NEWS for 1.23 1.23
Richard P. Curnow [Sun, 2 Dec 2007 14:53:09 +0000 (14:53 +0000)] 
Update NEWS for 1.23

17 years agoDefine io_linux.h constants for x86_64
Richard P. Curnow [Sun, 2 Dec 2007 14:39:50 +0000 (14:39 +0000)] 
Define io_linux.h constants for x86_64

Based on thread from chrony-users, October 2007.

18 years agogit archive's --prefix arg needs a trailing / 1.23-pre1
Richard P. Curnow [Wed, 27 Jun 2007 22:57:03 +0000 (23:57 +0100)] 
git archive's --prefix arg needs a trailing /

18 years agoUse git-archive instead of the obsolete git-tar-tree in make_release
Richard P. Curnow [Wed, 27 Jun 2007 22:54:43 +0000 (23:54 +0100)] 
Use git-archive instead of the obsolete git-tar-tree in make_release

18 years agoMerge branch 'bu'
Richard P. Curnow [Tue, 26 Jun 2007 22:51:18 +0000 (23:51 +0100)] 
Merge branch 'bu'

18 years agoFix formatting from last patch
Richard P. Curnow [Tue, 26 Jun 2007 22:50:53 +0000 (23:50 +0100)] 
Fix formatting from last patch

18 years agoFix handling of stratum zero.
Bill Unruh [Tue, 26 Jun 2007 22:46:33 +0000 (23:46 +0100)] 
Fix handling of stratum zero.

Further to the discussion with John Hasler, here are new diffs which
handles the incoming stratum 0 claim of a remote server by redefining the
incoming stratum as one bigger than the Max if it is zero, as per the NTP
version 4 documentation.

If the incoming stratum is zero it sets it to NTP_MAX_STRATUM+1 . If our
current stratum is larger than the NTP_MAX_STRATUM, the outgoing stratum is
also set to zero as per the suggestions in the NTP docs.
Introduces the new NTP_INVALID_STRATUM of 0 for doing these tests or
setting the outgoing stratum.

It is unclear whether chrony wants to follow NTP in setting the outgoing
stratum to zero if it is unknown or invalid, rather than a number larger
than the max stratum. Setting it to zero seems silly, since zero is already
used to define the stratum of a hardware clock (GPS, atomic, etc). This
seems ripe for confusion. But the fact that the ntp docs state to do this,
and that ntp servers (eg ntp.ubc.ca) are already doing this (using 0 to
mean invalid) means that chrony has to handle it on the incoming packets
from the servers.

18 years agoHandle fluctuations in peer distance better.
Bill Unruh [Tue, 26 Jun 2007 22:45:04 +0000 (23:45 +0100)] 
Handle fluctuations in peer distance better.

18 years agoFix whitespace issue with last patch
Richard P. Curnow [Tue, 26 Jun 2007 22:43:28 +0000 (23:43 +0100)] 
Fix whitespace issue with last patch

18 years agoFix problems with rtc_linux.
Bill Unruh [Tue, 26 Jun 2007 22:42:11 +0000 (23:42 +0100)] 
Fix problems with rtc_linux.

2) Changes to rtc_linux.c which a) do a double read of /dev/rtc when the
PPM interupt is turned on after the wait time expires. The current read
does not block to the second, as it should, thus two reads are needed.

Also, changes so that at startup the system properly ignores the last
system time from the initial burst mode for setting the system time since
it can be way off. At present this last system time is included in the
regression, which throws it off until finally that sample is dropped.

18 years agoFix sign v zero extension error in handling IP address
Stefan Lucke [Tue, 26 Jun 2007 22:02:33 +0000 (23:02 +0100)] 
Fix sign v zero extension error in handling IP address

I switch to the git version of chrony. Accidently this version did not
talk to by lokal server at 192.168.192.4. Instead it continuosly tried
255.255.192.4 :-( .

Tracked that down to "char", "unsigned char" issue in nameserv.c:

18 years agoMerge branch 'vm'
Richard P. Curnow [Tue, 26 Jun 2007 21:13:40 +0000 (22:13 +0100)] 
Merge branch 'vm'

18 years agoconfigure: fix indentation from previous patch
Richard P. Curnow [Tue, 26 Jun 2007 21:11:19 +0000 (22:11 +0100)] 
configure: fix indentation from previous patch

18 years agoAllow RTC support to be excluded at compile time.
Vladimir Michl [Tue, 26 Jun 2007 21:08:49 +0000 (22:08 +0100)] 
Allow RTC support to be excluded at compile time.

Add a new option to configure script, allowing
to disable (and exclude) RTC module. It saves same memory.

18 years agoAdd support for Linux/arm
Vladimir Michl [Tue, 26 Jun 2007 21:06:39 +0000 (22:06 +0100)] 
Add support for Linux/arm

19 years agoFix format of "could not send to" message
Richard P. Curnow [Sat, 15 Apr 2006 22:57:42 +0000 (23:57 +0100)] 
Fix format of "could not send to" message

19 years agoRename round() function to avoid clash with builtin mandriva-1.22
Richard P. Curnow [Fri, 14 Apr 2006 23:34:04 +0000 (00:34 +0100)] 
Rename round() function to avoid clash with builtin

19 years agoMerge quashing of compile warnings
Richard P. Curnow [Fri, 14 Apr 2006 23:32:08 +0000 (00:32 +0100)] 
Merge quashing of compile warnings

19 years agoQuash a load of compile warnings
kevin lyda [Fri, 14 Apr 2006 16:48:43 +0000 (17:48 +0100)] 
Quash a load of compile warnings

Kevin Lyda writes:
I enclose the following patch which removes all but three of the warnings.  i
don't have any non-linux systems handy to test a fix to the round() function.
but having it return a double should be fine.

It doesn't actually fix anything, it just shuts up -Wall, so it's certainly an
optional type of patch.

19 years agoFix version.txt to remove compile warning
Richard P. Curnow [Fri, 14 Apr 2006 23:27:07 +0000 (00:27 +0100)] 
Fix version.txt to remove compile warning

19 years agoFix bogus "system time" report for 64 bit systems
Eric Lammerts [Thu, 13 Apr 2006 15:15:26 +0000 (11:15 -0400)] 
Fix bogus "system time" report for 64 bit systems

Eric Lammerts writes:
This is known as Debian bug #195620, which is almost three years old!

The problem is that a uint32_t which comes out of ntohl() (but
actually represents a signed value) is directly promoted to long.
Therefore no sign extension takes place.

Patch below solves the problem. There are other places where this
needs to be fixed, but I'll leave that to a less lazy person.

19 years agoRemove volatile keyword from function declarations
Richard P. Curnow [Fri, 14 Apr 2006 23:10:21 +0000 (00:10 +0100)] 
Remove volatile keyword from function declarations

This silences a lot of gcc-4 compile warnings

19 years agoFix linux_io.h for MIPS
Bernhard Weiss [Sat, 24 Sep 2005 16:17:56 +0000 (18:17 +0200)] 
Fix linux_io.h for MIPS

Bernard Weiss writes:
I managed to compile the chrony 1.21 package for the MIPS architecture.
For the package to compile I had to add the following lines to io_linux.h:

[patch]

These values are taken from the ioctl.h file of linux 2.4.30 for the MIPS arch
(__ASM_MIPS_IOCTL_H).

19 years agoFreeBSD support
NAKAMURA Takumi [Sun, 11 Sep 2005 11:09:56 +0000 (20:09 +0900)] 
FreeBSD support

NAKAMURA Takumi writes:

I tried to compile chrony-1.21 on FreeBSD 4.8-RELEASE & 5.4-RELEASE.
I modify two files, configure, sysinc.h.

configure:
    add label "FreeBSD-i386" to "BSD/386" line
sysincl.h:
    1. FreeBSD obsoletes alloca.h
    2. FreeBSD use stdlib.h instead of malloc.h, to use malloc(), free()

Attached file includes the above modifications.

19 years agoFlush chronyc output buffers.
Paul Elliott [Mon, 5 Dec 2005 07:16:26 +0000 (01:16 -0600)] 
Flush chronyc output buffers.

The following is a patch to chronyc that causes it
to flush the buffers to stderr and stdout after
executing each command. This is needed if
you are controling chronyc from a program (i.e. chronyc's
input and output descriptors are pipes which are being
written/read by another program) and
you do not want to block waiting for chronyc response
which is trapped in a buffer!

19 years agoAdd a .gitignore file
Richard P. Curnow [Sun, 29 Jan 2006 23:29:01 +0000 (23:29 +0000)] 
Add a .gitignore file

19 years agomanpage improvements.
Richard P. Curnow [Sun, 29 Jan 2006 23:23:52 +0000 (23:23 +0000)] 
manpage improvements.

Try to address this Debian bug:
  Bug#345034: chrony: man pages refer to wrong sections

19 years ago1st attempt at a git-compatible make_release script
Richard P. Curnow [Sun, 22 Jan 2006 23:54:19 +0000 (23:54 +0000)] 
1st attempt at a git-compatible make_release script

19 years agoTidy up io_linux.h so unknown architecture result in compile-time error again
Richard P. Curnow [Sun, 22 Jan 2006 23:38:43 +0000 (23:38 +0000)] 
Tidy up io_linux.h so unknown architecture result in compile-time error again

19 years agoMore architectures in io_linux.h (John Hasler)
Richard P. Curnow [Sun, 22 Jan 2006 23:36:16 +0000 (23:36 +0000)] 
More architectures in io_linux.h (John Hasler)

19 years agoFix gcc-4 compilation of addrfilt.c
Richard P. Curnow [Thu, 19 Jan 2006 21:56:19 +0000 (21:56 +0000)] 
Fix gcc-4 compilation of addrfilt.c

John Hasler sent in a patch to do this (which still wouldn't make it compile
for me).  This reminded me that I had tackled this myself when my distro moved
to gcc-4 a while back.  It turned out I had never even checked in the file from
the working copy I was using (!).  Anyway, here it is now.

19 years agoFix potential buffer overflow problem (Martin Simmons) 1.21
Richard Curnow [Mon, 8 Aug 2005 22:40:31 +0000 (22:40 +0000)] 
Fix potential buffer overflow problem (Martin Simmons)

19 years agoMore build_kit fixes
Richard Curnow [Tue, 19 Jul 2005 22:09:30 +0000 (22:09 +0000)] 
More build_kit fixes

19 years agoFirst stab at moving build_kit to release versioning based on tla
Richard Curnow [Tue, 19 Jul 2005 22:04:01 +0000 (22:04 +0000)] 
First stab at moving build_kit to release versioning based on tla

19 years agoRTC usage bails out cleanly if accessing it goes wrong (e.g. HPET without generic...
Richard Curnow [Tue, 19 Jul 2005 21:22:52 +0000 (21:22 +0000)] 
RTC usage bails out cleanly if accessing it goes wrong (e.g. HPET without generic RTC emulation)

19 years agoFurther fixes to avoid use of linux kernel header files
Richard Curnow [Wed, 6 Jul 2005 22:49:13 +0000 (22:49 +0000)] 
Further fixes to avoid use of linux kernel header files