plugin discovery: partial revert to bring back valgrind #1995
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
success |
MrAnno
reviewed
Apr 18, 2018
lib/plugin.c
Outdated
| { | ||
| FILE *discovery; | ||
| GList *candidate_plugins = NULL; | ||
| gchar module_name[4096]; | ||
| gint plugin_type; | ||
| gchar plugin_name[256]; | ||
|
|
||
| discovery = fdopen(input_fd, "r"); | ||
| discovery = fmemopen(module_descriptors->str, module_descriptors->len, "r"); |
There was a problem hiding this comment.
fmemopen is not really portable. We should add a compat functions for this.
furiel
changed the title
|
As discussed in an offline review, we will avoid fmemopen and increase the complexity of the revert instead. |
furiel
force-pushed
the
revert-fork-module-load
branch
2 times, most recently
from
0eb221f to
17a0a70
Compare
|
success |
1 similar comment
|
success |
furiel
changed the title
|
I pushed the complete revert of 03771db as discussed. Removed the wip flag. Ready to review. |
Kokan
approved these changes
Apr 26, 2018
furiel
force-pushed
the
revert-fork-module-load
branch
from
17a0a70 to
5e67fd6
Compare
|
Force push to resolve conflict. |
|
failure |
The original code scanned the plugins (shared libraries) in a separate process and sent the list of plugins to the master process, then exited. This means the main process could selectively load the available plugins. Unfortunately due to unknown reasons, syslog-ng fails to start with this version. We are reverting the patch, and later we might try to reintroduce this functionality when the root cause of the sympthom is found. Running valgrind with --run-libc-freeres=no would start syslog-ng, but causes a lot of false positive results (even a simple printf appears in a log). The problem seems to be connected with the fork() and exit(), but the root cause is yet to be known. Signed-off-by: Antal Nemes <antal.nemes@balabit.com>
furiel
force-pushed
the
revert-fork-module-load
branch
from
5e67fd6 to
f9f32c6
Compare
|
success |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Revert for 03771db.
The original code scanned the plugins (shared libraries) in a separate
process and sent the list of plugins to the master process, then
exited. This means the main process could selectively load the
available plugins. Unfortunately due to unknown reasons, syslog-ng
fails to start with valgrind with this version. We are reverting the patch, and
later we might try to reintroduce this functionality when the root
cause of the sympthom is found.
Running valgrind with --run-libc-freeres=no would start syslog-ng, but
causes a lot of false positive results (even a simple printf appears
in a log). The problem seems to be connected with the fork() and
exit(), but the root cause is yet to be known.
Fixes: #1953
CC: @juhaszviktor