[Git][NTPsec/ntpsec][master] Update and reorganize metafiles.
Eric S. Raymond
gitlab at mg.gitlab.com
Mon Sep 26 16:33:28 UTC 2016
Eric S. Raymond pushed to branch master at NTPsec / ntpsec
Commits:
7d117a77 by Eric S. Raymond at 2016-09-26T12:32:10-04:00
Update and reorganize metafiles.
The one real bug in BUGS went to the issue tracker.
- - - - -
3 changed files:
- − BUGS
- README
- devel/TODO
Changes:
=====================================
BUGS deleted
=====================================
--- a/BUGS
+++ /dev/null
@@ -1,42 +0,0 @@
-= Known bugs =
-
-* The current setup assumed pthreads. They are only used for DNS
- lookups. It looks like most of the code is setup to do the right
- thing, but include/ntp_workimpl.h unconditionally turns on USE_WORKER
-
- waf sets up HAVE_PTHREAD and HAVE_PTHREAD_H, but none of the code
- uses them.
-
- USE_WORKER should probably be renamed to USE_DNS or USE_DNS_WORKER
-
-
-* msyslog doesn't have its own mutex. Non-atomicity can cause scrambled
- logs in some cases involving asynchronous getaddrinfo lookups.
-
-* There is a concurrency bug in asynchronous getaddrinfo lookups which
- manifests as name lookups getting dropped (the workaround is to
- specify servers as IP addresses). Hal says:
-
---------------------------------------------------------------------------
-The bug is also in NTP classic.
-
-It's allocating a slot for the [DNS] response. It's screwing up in
-the need-more case.
-
-First an overview:
-
-There is one thread doing DNS lookups. All its info is in a
-blocking_child struct.
-
-There are 2 arrays. One for requests and another for responses. They
-are used as ring buffers.
-
-If the next slot isn't free, ensure_workresp_empty_slot (re)allocates more space
-But it doesn't leave the index pointing to a new slot.
-
-I don't think there is any simple fix. You can't make ring-buffers
-bigger by just adding more space on the end. The fingers may be
-pointing into the middle of the block.
-
-There is probably a similar bug on the request side.
---------------------------------------------------------------------------
=====================================
README
=====================================
--- a/README
+++ b/README
@@ -13,8 +13,6 @@ directory.
The base directory ./ contains the configuration files, source
directories and related stuff:
-BUGS:: Descriptions of known bugs.
-
INSTALL:: Generic installation instructions.
NEWS:: What's new in this release.
@@ -25,7 +23,8 @@ VERSION:: The version stamp, to be used by scripts and programs.
contrib/:: Directory containing contributed scripts, dragons dwell here.
-devel/:: Documentation aimed at developers.
+devel/:: Documentation and small tools aimed at developers.
+ Includes a hacking guide and a tour of the internals.
docs/:: Directory containing a complete set of documentation on
building and configuring a NTP server or client. The files
=====================================
devel/TODO
=====================================
--- a/devel/TODO
+++ b/devel/TODO
@@ -4,6 +4,12 @@
=== Code ===
+* Land TESTFRAME.
+
+* Land Daniel's protocol-machine refactoring.
+
+* Land Daniel's redesign of the restriction language.
+
* Can the KERNEL_PLL code be removed? Hal thinks it may no longer
have a point since processors are much faster and schedulers
smoother than when it was built. Might remove the need for
@@ -20,6 +26,11 @@
== After 1.0 release ==
+* Replace ntpq with a Python wrapper around the ntp Python module.
+ This work is already in progress.
+
+* Move ntpdig to Python. Should be a wrapper around an ntp.sntp module.
+
=== Slow convergence ===
[quote, Gary Miller]
@@ -32,7 +43,6 @@ Right now, after 10 mins, ntpd has 2,000 times the jitter as chronyd had
when I turned it off.
__________
-
== User interface ==
[quote, Gary Miller]
@@ -86,10 +96,6 @@ Neither is ideal, easy pickings for someone to code on.
ntpshark, a command interpreter written around the ntp Python module and
loosely modeled on wireshark.
-* Replace ntpq with a Python wrapper around the ntp Python module.
-
-* Move ntpdig to Python. Should be a wrapper around an ntp.sntp module.
-
* We might be able to eliminate a lot of the Linux runtime
droproot code by using file capabilities.
@@ -135,8 +141,7 @@ __________
970318: in hourly_stats(?), squawk if the magnitude of the drift is,
say, >400.
-
-== Simple Tasks for an intern ==
+== Simple tasks for an intern ==
[quote, ESR 2015-09-30]
View it on GitLab: https://gitlab.com/NTPsec/ntpsec/commit/7d117a77e489ded2d79825b31045ef81264d564a
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ntpsec.org/pipermail/vc/attachments/20160926/e983c13a/attachment.html>
More information about the vc
mailing list