rpms/libqb
5
0
Fork 0
Code Issues Pull Requests Releases Activity

1.0.2-12 - Evolution of the previous (rhbz#1478089)

Browse Source 
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
This commit is contained in:
Jan Pokorný 2017-10-25 21:13:12 +02:00
parent 8bf1fa94fc
commit e3be0a5f65
No known key found for this signature in database
GPG Key ID: 61BBB23A9E8F8DE2
7 changed files with 405 additions and 224 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
01-Med-qblog.h-better-explanation-behaviour-of-QB_LOG_I.patch
View File

@ -1,4 +1,4 @@
From 3f9ccfc3b291196657edb2e920327f2d26804d6a Mon Sep 17 00:00:00 2001
From 47f007a9bbdbfcdb1df6736dddde64f9c6381b58 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Pokorn=C3=BD?= <jpokorny@redhat.com>
Date: Fri, 6 Oct 2017 17:17:26 +0200
Subject: [PATCH 1/6] Med: qblog.h: better explanation + behaviour of
@ -26,7 +26,7 @@ requires.
- after:
pacemakerd: utils.c:69: common:
Assertion `"non-empty callsite section" && QB_ATTR_SECTION_START != QB_ATTR_SECTION_STOP' failed.
Assertion `"non-empty implicit callsite section" && QB_ATTR_SECTION_START != QB_ATTR_SECTION_STOP' failed.
Restructuring of the assertion inspired by the suggestion of Ferenc
Wágner (for the subsequent commit, actually).
@ -42,11 +42,11 @@ libqb-target cross-combination matrix and the findings).
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
---
include/qb/qblog.h | 43 ++++++++++++++++++++++++++++++++++++++-----
1 file changed, 38 insertions(+), 5 deletions(-)
include/qb/qblog.h | 45 ++++++++++++++++++++++++++++++++++++++++-----
1 file changed, 40 insertions(+), 5 deletions(-)
diff --git a/include/qb/qblog.h b/include/qb/qblog.h
index 3cb4eef..7b6db8b 100644
index 3cb4eef..4954932 100644
--- a/include/qb/qblog.h
+++ b/include/qb/qblog.h
@@ -42,6 +42,10 @@ extern "C" {
@ -78,7 +78,7 @@ index 3cb4eef..7b6db8b 100644
* @par Configuring log targets.
* A log target can be syslog, stderr, the blackbox, stdout, or a text file.
* By default only syslog is enabled.
@@ -268,11 +283,29 @@ typedef void (*qb_log_filter_fn)(struct qb_log_callsite * cs);
@@ -268,11 +283,31 @@ typedef void (*qb_log_filter_fn)(struct qb_log_callsite * cs);
extern struct qb_log_callsite QB_ATTR_SECTION_START[];
extern struct qb_log_callsite QB_ATTR_SECTION_STOP[];
@ -87,32 +87,34 @@ index 3cb4eef..7b6db8b 100644
- void name(void); \
- void name(void) \
- { if (QB_ATTR_SECTION_START == QB_ATTR_SECTION_STOP) assert(0); } \
+/* optional on-demand self-check of (1) toolchain sanity (prerequisite for
+ the logging subsystem to work properly) and (2) non-void active use of
+/* optional on-demand self-check of 1/ toolchain sanity (prerequisite for
+ the logging subsystem to work properly) and 2/ non-void active use of
+ logging (satisfied with a justifying existence of a logging callsite as
+ defined with a @c qb_logt invocation) at the target (but see below), which
+ is presumably assured by it's author as of relying on this very macro
+ is supposedly assured by it's author(!) as of relying on this very macro
+ [technically, the symbols that happen to be resolved under the respective
+ identifiers do not necessarily come from the same compilation unit level
+ as when it's not the end program (or by induction, a library higher in
+ the symbol resolution fall-through) but a shared library, the former takes
+ a precedence unless it lacks use of logging on its own making its very
+ section empty and hence without such boundary symbols provided at the
+ respective level and thus deferring to lower priority symbol resolution
+ stop -- despite this fuzzily targeted attestation, the check remains
+ reasonable and should not harm anything];
+ identifiers do not necessarily originate in the same compilation unit as
+ when it's not the end executable (or by induction, a library positioned
+ earlier in the symbol lookup order) but a shared library, the former takes
+ a precedence unless that site comes short of exercising the logging,
+ making its callsite section empty and, in turn, without such boundary
+ symbols, hence making the resolution continue further in the lookup order
+ -- despite fuzzily targeted attestation, the check remains reasonable];
+ only effective when link-time ("run-time amortizing") callsite collection
+ is; as a side effect, it can ensure the boundary-denoting symbols for the
+ target collection area are kept alive with some otherwise unkind linkers;
+ may be extended in future for more in-depth self-validation */
+#define QB_LOG_INIT_DATA(name) \
+ void name(void); \
+ void name(void) \
+ { assert("non-empty callsite section" \
+ void name(void) { \
+ /* our own (target's) sanity, or possibly that of higher priority \
+ symbol resolution site (unless target equals end executable) \
+ or even the lower one if no such predecessor defines these */ \
+ assert("non-empty implicit callsite section" \
+ && QB_ATTR_SECTION_START != QB_ATTR_SECTION_STOP); } \
void __attribute__ ((constructor)) name(void);
#else
#define QB_LOG_INIT_DATA(name)
--
2.14.2
2.14.3

4
02-build-configure-run-attribute-section-test-through-r.patch
View File

@ -1,4 +1,4 @@
From 796ce1191ffd99f0bceba174355f695af8dc368a Mon Sep 17 00:00:00 2001
From 287ae050e5330202eca59adbb7e87e9d26d69e08 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Pokorn=C3=BD?= <jpokorny@redhat.com>
Date: Fri, 6 Oct 2017 17:17:26 +0200
Subject: [PATCH 2/6] build: configure: run attribute section test through
@ -52,5 +52,5 @@ index 22630ba..478194c 100644
gcc_has_attribute_section=${ac_cv_link_attribute_section}
fi
--
2.14.2
2.14.3

32
03-tests-new-sort-of-tests-dubbed-functional-cover-link.patch
View File

@ -1,4 +1,4 @@
From 3df0138ba57e274b8c8065739870b1529b95e087 Mon Sep 17 00:00:00 2001
From afb1e96c8b45c520f1e66416167ad943326ae4fe Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Pokorn=C3=BD?= <jpokorny@redhat.com>
Date: Fri, 6 Oct 2017 17:17:26 +0200
Subject: [PATCH 3/6] tests: new sort of tests dubbed "functional", cover
@ -47,12 +47,18 @@ libqb.spec for RPM packaging, whereas the regular one (b.) serves as
a building block for new log_test_mock.sh runner of said test matrix
-- it iterates through all the possible permutations of linker-imposed
implicit visibility of mentioned symbols between various affected
linkage sides (see 1. - 3. above) so as to demonstrate a/ the impact
of the problem (see table below), and subsequently b/ that the fix is
effective in all these situations (updated table will be provided as
well) once it lands.
link participants all making use of logging (see 1. - 3. above) so as
to demonstrate A/ the impact of the problem (see table below), and
subsequently B/ that the fix is effective in all these situations
(updated table will be provided as well) once it lands. This script
also allows convoluting the test matrix further, notably with on-demand
defusing the self-checks based on QB_LOG_INIT_DATA macro, which is
of significance as demonstrated below (and will become even more
important with upcoming patches in this series).
Current state for this matrix, in which participants 1. - 3. map like:
* * *
Current state for such matrix, in which participants 1. - 3. map like:
1. ~ libqb(Y)
2. ~ "direct"
3. ~ libX(Y) [a.k.a. interlib]
@ -82,16 +88,18 @@ can observe a somewhat simpler story (DEP ~ "depends"):
# x = b # DEP[*I] | DEP[*I] : DEP[*K] # BAD[*1] | BAD[*1] : BAD[*L] #
+=========+=========+=========+=========+=========+=========+=========+
* * *
[*1] client logging not working
[*2] interlib logging not working
[*3] both client and interlib logging not working
[*A] boils down to [*1], unless QB_LOG_INIT_DATA used on client side,
which fails on 'non-empty callsite section' assertion
which fails on 'non-empty implicit callsite section' assertion
[*B] boils down to [*1], unless QB_LOG_INIT_DATA used on interlib side,
which fails on 'non-empty callsite section' assertion
which fails on 'non-empty implicit callsite section' assertion
[*C] boils down to [*3], unless QB_LOG_INIT_DATA used on interlib side,
which fails on 'non-empty callsite section' assertion
which fails on 'non-empty implicit callsite section' assertion
[*D] boils down to [*3], unless QB_LOG_INIT_DATA used on interlib side,
which makes it boil down just to [*1] (hypothesis: mere internal
self-reference to the section's boundary symbols makes them
@ -113,6 +121,8 @@ can observe a somewhat simpler story (DEP ~ "depends"):
(sufficient?), which makes it, likely through self-reference
keepalive (see below) boil down just to [*1]
* * *
Note: as observed with [*D] case (libqb linked with ld.bfd < 2.29
whereas interlib and its client linked with ld.bfd = 2.29), the exact
availability of a working logging doesn't depend solely on the linkers
@ -154,7 +164,7 @@ Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
tests/functional/log_test_interlib_client.err | 4 +
tests/functional/log_test_interlib_client.sh | 33 +++
tests/functional/log_test_mock.sh | 310 ++++++++++++++++++++++++++
18 files changed, 767 insertions(+), 4 deletions(-)
16 files changed, 754 insertions(+), 2 deletions(-)
create mode 100644 tests/functional/GNUmakefile
create mode 100644 tests/functional/Makefile.am
create mode 100644 tests/functional/log.am
@ -1051,5 +1061,5 @@ index 0000000..145b9f4
+ "- results stored in '_results/<timestamp>_<input_name>[_<tag>]'" \
+ || do_proceed "$@"
--
2.14.2
2.14.3

316
04-Med-add-extra-run-time-client-libqb-checks-that-logg.patch
View File

@ -1,4 +1,4 @@
From 4f0fceaaf38a04e9f1c067a7b581265b9ceb81de Mon Sep 17 00:00:00 2001
From ca5c4425393c462670861a023608d30977a17de7 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Pokorn=C3=BD?= <jpokorny@redhat.com>
Date: Fri, 6 Oct 2017 17:17:26 +0200
Subject: [PATCH 4/6] Med: add extra run-time (client, libqb) checks that
@ -7,15 +7,24 @@ MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The table at the bottom concludes how the test matrix overview
Now that the previous commit provides a foundation for what exactly can
go wrong with ld.bfd = 2.29+ linker, let's start reconciling that with
some reasonable assurance that logging is not silently severed, because
realizing the logs are missing is otherwise bound to happen when the
logs are suddently pretty crucial analytical resource :-)
We'll proceed in two steps as detailed.
* * *
As a first step, the table below concludes how the test matrix overview
introduced with a message for the preceding commit (also introducing
log_test_mock.sh runner which got reused here) looks as of this
refreshed sanity check, once QB_LOG_INIT_DATA macro at hand gets
applied (meaning "for non-libqb logging participants" so as not
complicate the matrix further). That macro is nothing triggered
directly, it will just plant a constructor-like function (to be
invoked automatically early in the execution) that will run through
the checks (one original and couple of new ones as of this changeset).
refreshed sanity check, once QB_LOG_INIT_DATA macro at hand gets applied
(meaning "for non-libqb logging participants" so as not complicate the
matrix further). That macro is nothing triggered directly, it will just
plant a constructor-like function (to be invoked automatically early in
the execution) that will run through the checks (one original and couple
of new ones as of this changeset).
Note that for libqb users, this implies a new link dependency on libdl,
because they may opt-in for refreshed QB_LOG_INIT_DATA sanity check that
@ -48,28 +57,68 @@ the table stated in the preceding commit):
# x = b # BAD[*A] | BAD[*B] : BAD[*C] #<BAD[*E]>|<BAD[*F]>:<BAD[*G]>#
+=========+=========+=========+=========+=========+=========+=========+
whereas if we swap 2.29 for 2.29.1, i.e., X(b) .. ld.bfd = 2.29.1, we
can observe a somewhat simpler story (DEP ~ "depends"):
Woefully, nothing changes if we swap binutils 2.29 for 2.29.1, i.e.,
X(b) .. ld.bfd = 2.29.1, compared to previous state, i.e., the second
table from the previous commit is still applicable for that situation.
The added sanity checks are useful nonetheless, consider for example,
that attribute-section-less libqb is what gets run-time linked to an
attribute-section-full target. The most precise check we could use
-- a custom logger function applied in a self-test scheme -- is not
available at the point the macro-defined function gets invoked, simply
because qb_log_init hasn't been invoked by the time that constructor
gets triggered. However, what we can do is to add a non-trapping
libqb-residing reverse-testing of the client space that (and once it)
voluntarily initiates qb_log_init (delivering abruption all of a sudden
at some unanticipated, as opposed to a well-timed like with
constructors, execution point, seems pretty bad idea + libqb as
a library is a mere helper, not an undertaker :) -- this check then
only announces, via syslog (the only pre-enabled logging target),
the target's logging may be severed.
Woefully, nothing changes if we swap binutils 2.29.1 for 2.29, i.e.,
X(b) .. ld.bfd = 2.29.1. The second table from the previous commit
hence remains the same. The added sanity checks are useful nonetheless,
consider for example, that attribute-section-less libqb is what gets
run-time linked to an attribute-section-full target. The most precise
check we could use -- a custom logger function applied in a self-test
scheme -- is not available at the point the macro-defined function gets
invoked, simply because qb_log_init hasn't been invoked by the time
that constructor gets triggered. However, what we can do is to add
a non-trapping libqb-residing reverse-testing of the client space that
invokes qb_log_init (delivering abruption all of a sudden at some
unanticipated, as opposed to a well-timed like with constructors,
execution point, seems pretty bad idea + libqb as a library is a mere
helper, not an undertaker :) -- this check only announces, via syslog
(the only pre-enabled logging target), the target logging won't work.
* * *
After incorporating such a change (and extending log_test_mock.sh so as
to capture syslog stream within the container), nothing changes with the
table above, but desirably changes with "X(b) .. ld.bfd = 2.29.1" one:
Hence, as a promised second step, after incorporating the syslog
change (and extending log_test_mock.sh so as to capture syslog
stream within the container), not much changes with the table above,
i.e., X(b) .. ld.bfd = 2.29:
[*A] in addition, unless QB_LOG_INIT_DATA used on client side,
syslog carries this error:
"(LOG@QB) target chain supplied section not observed by libqb,
target's and/or libqb's linkage at fault and logging of the target
will not work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
[*C] in addition, unless QB_LOG_INIT_DATA used on interlib side,
syslog carries this error:
"(LOG@QB) target chain supplied section not observed by libqb,
target's and/or libqb's linkage at fault and logging of the target
will not work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
[*E] in addition, unless QB_LOG_INIT_DATA used on client side,
syslog carries this warning:
"(LOG@QB) target chain supplied section empty, target's and/or
libqb's linkage at fault and logging of the target will not
work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
[*F] in addition, unless QB_LOG_INIT_DATA used on interlib side,
syslog carries this warning:
"(LOG@QB) target chain supplied section empty, target's and/or
libqb's linkage at fault and logging of the target will not
work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
[*G] in addition, unless QB_LOG_INIT_DATA used on interlib side,
syslog carries this warning:
"(LOG@QB) target chain supplied section empty, target's and/or
libqb's linkage at fault and logging of the target will not
work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
but desirably changes with "X(b) .. ld.bfd = 2.29.1" one
(DEP ~ "depends"):
+=========+=========+=========+=========+=========+=========+=========+
#client(x)# libqb(a) usage # libqb(b) usage #
@ -80,6 +129,8 @@ table above, but desirably changes with "X(b) .. ld.bfd = 2.29.1" one:
# x = b #<DEP[*N]>| DEP[*I] :<DEP[*O]>#<BAD[*M]>|<BAD[*M]>:<BAD[*L]>#
+=========+=========+=========+=========+=========+=========+=========+
* * *
[*1] client logging not working
[*2] interlib logging not working
[*3] both client and interlib logging not working
@ -101,8 +152,8 @@ table above, but desirably changes with "X(b) .. ld.bfd = 2.29.1" one:
assertion
[*E] boils down to [*1], unless QB_LOG_INIT_DATA used on client side,
which then fails on
"target's callsite section self-observable, otherwise target's
and/or libqb's linkage at fault and logging would not work reliably"
"implicit callsite section self-observable, otherwise target's and/or
libqb's linkage at fault and logging would not work reliably"
assertion
[*F] boils down to [*3], unless QB_LOG_INIT_DATA used on interlib side,
which then fails on
@ -111,44 +162,49 @@ table above, but desirably changes with "X(b) .. ld.bfd = 2.29.1" one:
assertion
[*G] boils down to [*3], unless QB_LOG_INIT_DATA used on interlib side,
which then fails on
"target's callsite section self-observable, otherwise target's and/or
"implicit callsite section self-observable, otherwise target's and/or
libqb's linkage at fault and logging would not work reliably"
assertion
[*I] boils down to [*1], unless QB_LOG_INIT_DATA used on client side,
which makes it, likely through self-reference keepalive (see
below) work OK
below), work OK
[*J] boils down to [*2], unless QB_LOG_INIT_DATA used on interlib side,
which makes it, likely through self-reference keepalive (see
below) work OK
below), work OK
[*K] boils down to [*3]
in addition, syslog carries this error:
"target chain supplied section not observed by libqb, target's and/or
libqb's linkage at fault and logging of the target will not work
reliably"
"(LOG@QB) target chain supplied section not observed by libqb,
target's and/or libqb's linkage at fault and logging of the target
will not work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
[*L] boils down to [*3], unless QB_LOG_INIT_DATA used on interlib side
(sufficient?), which makes it, likely through self-reference
keepalive (see below) boil down just to [*1];
keepalive (see below), boil down just to [*1];
in addition, syslog carries this error:
"target chain supplied section not observed by libqb, target's and/or
libqb's linkage at fault and logging of the target will not work
reliably"
"(LOG@QB) target chain supplied section not observed by libqb,
target's and/or libqb's linkage at fault and logging of the target
will not work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
[*M] boils down to [*1];
in addition, syslog carries this error:
"target chain supplied section not observed by libqb, target's and/or
libqb's linkage at fault and logging of the target will not work
reliably"
"(LOG@QB) target chain supplied section not observed by libqb,
target's and/or libqb's linkage at fault and logging of the target
will not work reliably (unless qb_log_init function used unexpectedly
in no-logging context)"
logged by libqb proper
[*N] boils down to [*M], unless QB_LOG_INIT_DATA used on client side,
which makes it, likely through self-reference keepalive (see
below) work OK
below), work OK
[*O] boils down to [*K], unless QB_LOG_INIT_DATA used on both client
and interlib side, which makes it, likely through self-reference
keepalive (see below) work OK (it's expected that this a mere
keepalive (see below), work OK (it's expected that this a mere
composite of situations [*I] and [*J] with consequences as stated)
* * *
Note: the only problematic (i.e. not captured automatically by the
QB_LOG_INIT_DATA macro presumably utilized at every non-libqb logging
system participant in the form of a discrete compilation unit)
@ -158,27 +214,31 @@ course, be solvable as well, but presumably not in an easy way, and
that use case should not be as frequent.
Takeway: whenever your target (library or client program) actively
utilizes logging (meaning it emits at least a single log message),
YOU ARE strongly ENCOURAGED TO USE QB_LOG_INIT_DATA macro function
at (exactly) one of the source code files (presumably the main one)
per respective target's compilation unit. It will alleviate the
hassles possibly caused by downgrading libqb to the linker-vs-libqb
incompatibly compiled one or in similar circumstances arising merely
from the linker behaviour change, which the current build system/code
quake is all about.
utilizes logging (meaning it emits at least a single log message,
otherwise there's an imminent danger of possibly even run-terminating
false positive in the self-check mechanism!),
YOU ARE strongly ENCOURAGED TO USE QB_LOG_INIT_DATA macro function
at (exactly) one of the source code files (presumably the main one)
per respective target's compilation unit.
It will alleviate the hassles possibly caused by downgrading libqb
to the linker-vs-libqb incompatibly compiled one or in similar
circumstances arising merely from the linker behaviour change,
which the current build system/code shake is all about.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
---
configure.ac | 1 +
include/qb/qblog.h | 95 ++++++++++++++++++++++++-------
lib/libqb.pc.in | 3 +-
lib/log.c | 44 ++++++++++++++
tests/functional/Makefile.am | 3 +-
tests/functional/log.am | 2 +-
tests/functional/log_external/Makefile.am | 8 ++-
tests/functional/log_test_mock.sh | 13 ++++-
tests/functional/syslog-stdout.py | 49 ++++++++++++++++
9 files changed, 189 insertions(+), 29 deletions(-)
configure.ac | 1 +
include/qb/qblog.h | 110 ++++++++++++++++++++++--------
lib/libqb.pc.in | 3 +-
lib/log.c | 49 +++++++++++++
tests/functional/Makefile.am | 3 +-
tests/functional/log.am | 2 +-
tests/functional/log_external/Makefile.am | 8 ++-
tests/functional/log_test_mock.sh | 13 +++-
tests/functional/syslog-stdout.py | 51 ++++++++++++++
9 files changed, 204 insertions(+), 36 deletions(-)
create mode 100755 tests/functional/syslog-stdout.py
diff --git a/configure.ac b/configure.ac
@ -194,35 +254,69 @@ index 274b49c..5769e60 100644
fi
diff --git a/include/qb/qblog.h b/include/qb/qblog.h
index 7b6db8b..10cd62c 100644
index 4954932..766cffe 100644
--- a/include/qb/qblog.h
+++ b/include/qb/qblog.h
@@ -44,6 +44,7 @@ extern "C" {
@@ -1,9 +1,10 @@
/*
- * Copyright (C) 2010 Red Hat, Inc.
+ * Copyright 2017 Red Hat, Inc.
*
* All rights reserved.
*
* Author: Angus Salkeld <asalkeld@redhat.com>
+ * Jan Pokorny <jpokorny@redhat.com>
*
* libqb is free software: you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as published by
@@ -44,6 +45,7 @@ extern "C" {
#ifdef QB_HAVE_ATTRIBUTE_SECTION
#include <assert.h> /* possibly needed for QB_LOG_INIT_DATA */
+#include <dlfcn.h>
+#include <dlfcn.h> /* dynamic linking: dlopen, dlsym, dladdr, ... */
#endif
/**
@@ -283,33 +284,83 @@ typedef void (*qb_log_filter_fn)(struct qb_log_callsite * cs);
@@ -70,14 +72,14 @@ extern "C" {
*
* @note
* In practice, such a minimalistic approach hardly caters real use cases.
- * Following section discusses the customization. Moreover, it's quite
- * vital to instrument the target user of this logging subsystem with
- * @c QB_LOG_INIT_DATA() macro placed in the top file scope in an exactly
- * one source file (preferably the main one) to be mixed into the resulting
- * compilation unit. This is a self-defensive measure for when the
- * linker-assisted collection of callsite data silently fails, which
- * could otherwise go unnoticed, causing troubles down the road.
- *
+ * Following section discusses the customization. Moreover when employing
+ * the log module is bound to its active use (some log messages are assuredly
+ * emitted within the target compilation unit), it's quite vital to instrument
+ * the target side with @c QB_LOG_INIT_DATA() macro placed in the top file
+ * scope in exactly one source file (preferably the main one) to be mixed into
+ * the resulting compilation unit. This is a self-defensive measure for when
+ * the linker-assisted collection of callsite data silently fails, which could
+ * otherwise go unnoticed, causing troubles down the road.
*
* @par Configuring log targets.
* A log target can be syslog, stderr, the blackbox, stdout, or a text file.
@@ -283,35 +285,85 @@ typedef void (*qb_log_filter_fn)(struct qb_log_callsite * cs);
extern struct qb_log_callsite QB_ATTR_SECTION_START[];
extern struct qb_log_callsite QB_ATTR_SECTION_STOP[];
-/* optional on-demand self-check of (1) toolchain sanity (prerequisite for
- the logging subsystem to work properly) and (2) non-void active use of
-/* optional on-demand self-check of 1/ toolchain sanity (prerequisite for
- the logging subsystem to work properly) and 2/ non-void active use of
- logging (satisfied with a justifying existence of a logging callsite as
- defined with a @c qb_logt invocation) at the target (but see below), which
- is presumably assured by it's author as of relying on this very macro
- is supposedly assured by it's author(!) as of relying on this very macro
- [technically, the symbols that happen to be resolved under the respective
- identifiers do not necessarily come from the same compilation unit level
- as when it's not the end program (or by induction, a library higher in
- the symbol resolution fall-through) but a shared library, the former takes
- a precedence unless it lacks use of logging on its own making its very
- section empty and hence without such boundary symbols provided at the
- respective level and thus deferring to lower priority symbol resolution
- stop -- despite this fuzzily targeted attestation, the check remains
- reasonable and should not harm anything];
- identifiers do not necessarily originate in the same compilation unit as
- when it's not the end executable (or by induction, a library positioned
- earlier in the symbol lookup order) but a shared library, the former takes
- a precedence unless that site comes short of exercising the logging,
- making its callsite section empty and, in turn, without such boundary
- symbols, hence making the resolution continue further in the lookup order
- -- despite fuzzily targeted attestation, the check remains reasonable];
- only effective when link-time ("run-time amortizing") callsite collection
- is; as a side effect, it can ensure the boundary-denoting symbols for the
- target collection area are kept alive with some otherwise unkind linkers;
@ -232,9 +326,10 @@ index 7b6db8b..10cd62c 100644
+#define QB_NONAPI_LOG_INIT_DATA_EXTRA_ \
+ _Pragma(QB_PP_STRINGIFY(GCC warning QB_PP_STRINGIFY( \
+ without "_GNU_SOURCE" defined (directly or not) \
+ qb_log_init_data cannot check sanity of libqb proper)))
+ QB_LOG_INIT_DATA cannot check sanity of libqb proper)))
+#else
+#define QB_NONAPI_LOG_INIT_DATA_EXTRA_ \
+ { Dl_info work_dli; \
+ /* libqb sanity (locating libqb by it's relatively unique \
+ -- and currently only such per-linkage global one -- \
+ non-functional symbol, due to possible confusion otherwise) */ \
@ -251,24 +346,23 @@ index 7b6db8b..10cd62c 100644
+ assert("libqb's callsite section non-empty, otherwise libqb's \
+linkage at fault and logging would not work reliably" \
+ && work_s1 != work_s2); \
+ dlclose(work_handle); }
+ dlclose(work_handle); } }
+#endif /* _GNU_SOURCE */
+
+/**
+ * Optional on-demand self-check of (1) toolchain sanity (prerequisite for
+ * the logging subsystem to work properly) and (2) non-void active use of
+ * Optional on-demand self-check of 1/ toolchain sanity (prerequisite for
+ * the logging subsystem to work properly) and 2/ non-void active use of
+ * logging (satisfied with a justifying existence of a logging callsite as
+ * defined with a @c qb_logt invocation) at the target (but see below), which
+ * is presumably assured by it's author as of relying on this very macro
+ * is supposedly assured by it's author(!) as of relying on this very macro
+ * [technically, the symbols that happen to be resolved under the respective
+ * identifiers do not necessarily come from the same compilation unit level
+ * as when it's not the end program (or by induction, a library higher in
+ * the symbol resolution fall-through) but a shared library, the former takes
+ * a precedence unless it lacks use of logging on its own making its very
+ * section empty and hence without such boundary symbols provided at the
+ * respective level and thus deferring to lower priority symbol resolution
+ * stop -- despite this fuzzily targeted attestation, the check remains
+ * reasonable and should not harm anything];
+ * identifiers do not necessarily originate in the same compilation unit as
+ * when it's not the end executable (or by induction, a library positioned
+ * earlier in the symbol lookup order) but a shared library, the former takes
+ * a precedence unless that site comes short of exercising the logging,
+ * making its callsite section empty and, in turn, without such boundary
+ * symbols, hence making the resolution continue further in the lookup order
+ * -- despite fuzzily targeted attestation, the check remains reasonable];
+ * only effective when link-time ("run-time amortizing") callsite collection
+ * is; as a side effect, it can ensure the boundary-denoting symbols for the
+ * target collection area are kept alive with some otherwise unkind linkers.
@ -282,22 +376,23 @@ index 7b6db8b..10cd62c 100644
+ */
#define QB_LOG_INIT_DATA(name) \
void name(void); \
- void name(void) \
- { assert("non-empty callsite section" \
- && QB_ATTR_SECTION_START != QB_ATTR_SECTION_STOP); } \
+ void name(void) { \
void name(void) { \
+ void *work_handle; struct qb_log_callsite *work_s1, *work_s2; \
+ Dl_info work_dli; \
+ /* our own (target's) sanity */ \
/* our own (target's) sanity, or possibly that of higher priority \
symbol resolution site (unless target equals end executable) \
or even the lower one if no such predecessor defines these */ \
- assert("non-empty implicit callsite section" \
- && QB_ATTR_SECTION_START != QB_ATTR_SECTION_STOP); } \
+ if ((work_handle = dlopen(NULL, RTLD_LOCAL|RTLD_LAZY)) != NULL) { \
+ work_s1 = (struct qb_log_callsite *) \
+ dlsym(work_handle, QB_ATTR_SECTION_START_STR); \
+ work_s2 = (struct qb_log_callsite *) \
+ dlsym(work_handle, QB_ATTR_SECTION_STOP_STR); \
+ assert("target's callsite section self-observable, otherwise \
+ assert("implicit callsite section observable, otherwise \
+target's and/or libqb's linkage at fault and logging would not work \
+reliably" \
+ && work_s1 != NULL && work_s2 != NULL); } \
+ /* better targeted attestations when available */ \
+ QB_NONAPI_LOG_INIT_DATA_EXTRA_; \
+ /* finally, original, straightforward check */ \
+ assert("non-empty implicit callsite section, otherwise target's \
@ -324,7 +419,7 @@ index 8a8d0ba..37d27b7 100644
+Libs.private: @LIBS@
Cflags: -I${includedir}
diff --git a/lib/log.c b/lib/log.c
index bfd218f..3e9972e 100644
index bfd218f..777f378 100644
--- a/lib/log.c
+++ b/lib/log.c
@@ -853,6 +853,18 @@ qb_log_init(const char *name, int32_t facility, uint8_t priority)
@ -372,7 +467,7 @@ index bfd218f..3e9972e 100644
qb_log_callsites_register(QB_ATTR_SECTION_START, QB_ATTR_SECTION_STOP);
dl_iterate_phdr(_log_so_walk_callback, NULL);
_log_so_walk_dlnames();
@@ -884,6 +915,19 @@ qb_log_init(const char *name, int32_t facility, uint8_t priority)
@@ -884,6 +915,24 @@ qb_log_init(const char *name, int32_t facility, uint8_t priority)
_log_target_state_set(&conf[QB_LOG_SYSLOG], QB_LOG_STATE_ENABLED);
(void)qb_log_filter_ctl(QB_LOG_SYSLOG, QB_LOG_FILTER_ADD,
QB_LOG_FILTER_FILE, "*", priority);
@ -382,12 +477,17 @@ index bfd218f..3e9972e 100644
+ " not observed by libqb, target's and/or"
+ " libqb's linkage at fault and"
+ " logging of the target will not"
+ " work reliably");
+ " work reliably (unless qb_log_init"
+ " function used unexpectedly in"
+ " no-logging context)");
+ else if (preinit_err == preinit_err_target_empty) {
+ qb_util_log(LOG_WARNING, "(LOG@QB) target chain supplied"
+ " section empty, target's linkage"
+ " at fault and logging of the target"
+ " will not work reliably");
+ " section empty, target's and/or"
+ " libqb's linkage at fault and"
+ " logging of the target will not"
+ " work reliably (unless qb_log_init"
+ " function used unexpectedly in"
+ " no-logging context)");
+ }
}
@ -461,10 +561,10 @@ index 145b9f4..cdfce1f 100755
mock ${mock_args} --copyout \
diff --git a/tests/functional/syslog-stdout.py b/tests/functional/syslog-stdout.py
new file mode 100755
index 0000000..a0a761b
index 0000000..64baf4c
--- /dev/null
+++ b/tests/functional/syslog-stdout.py
@@ -0,0 +1,49 @@
@@ -0,0 +1,51 @@
+#!/usr/bin/python3
+# Copyright 2017 Red Hat, Inc.
+#
@ -485,6 +585,8 @@ index 0000000..a0a761b
+# You should have received a copy of the GNU Lesser General Public License
+# along with libqb. If not, see <http://www.gnu.org/licenses/>.
+
+"""Simple /dev/log to stdout forwarding"""
+
+import socket
+from atexit import register
+from os import remove
@ -515,5 +617,5 @@ index 0000000..a0a761b
+if __name__ == '__main__':
+ main(*argv)
--
2.14.2
2.14.3

170
05-High-bare-fix-for-libqb-logging-not-working-with-ld..patch
View File

@ -1,4 +1,4 @@
From dbe771c0ab30ded6f1d82b6ed9d61ac04560e913 Mon Sep 17 00:00:00 2001
From 0de4d2298da9c5880ff8c32ac2e75c3104d44816 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Pokorn=C3=BD?= <jpokorny@redhat.com>
Date: Fri, 6 Oct 2017 17:17:26 +0200
Subject: [PATCH 5/6] High: bare fix for libqb logging not working with
@ -7,7 +7,9 @@ MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
(or rather [read on]: "bare" fix)
(or rather [read on]: "bare" fix, now that we established means to
analyse the impact of the linker-dependent misbehaviour and to detect
some of its symptoms in preceding two commits, respectively)
Initially with the help of the internal test suite and the failing log
test, it was eventually discovered[1] that these binutils commits going
@ -43,7 +45,7 @@ https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=487b6440dad57440939fa
At least that change doesn't invalidate all the effort being put into
the original version of the changeset, only the configure script check
had to be refined so as not to miss the "orphan section magic not
working properly out of the box, without a bandaid" observation
working properly out of the box, without band aid" observation
(see the inline comment) -- the workaround arrangement needs
to be applied in that case as well.
@ -103,6 +105,8 @@ So further extensions were devised to cover both consumption models:
of consuming libqb outside of it's own playground is hardly
the Right Thing) if portability is important, nonetheless!
* * *
On the address of linker script workaround, there are linkers out there
that do not support the trick, for instance:
- ld.gold from binutils (but it has hardly ever been working with
@ -112,7 +116,7 @@ that do not support the trick, for instance:
system linker even in the most recent releases of FreeBSD, unless
GCC toolchain is used instead
If these are hit when (because) the compiler has already demonstrated it
supports section attribute, the build system configuration is forcibly
supports "section" attribute, the build system configuration is forcibly
stopped, simply to stay conceptually compatible with the prior state in
which the affinity to leverage that feature hasn't been called off
under any circumstances. One is, however, able to achieve exactly
@ -143,6 +147,60 @@ within libqb.so proper (at least with binutils 2.29).
* * *
To provide a high-level prioritized overview of what drove the approach:
- PRESERVATION OF BINARY COMPATIBILITY (ABI), which is achieved except
for a single "ABI nongracefulness" I am aware of but that's more
a consequence of slightly incorrect assumptions in the logic of
QB_LOG_INIT_DATA macro function predating this whole affair by
a long shot and which the patchset finally rectifies:
if in the run-time dynamic link, following is combined:
(. libqb, arbitrary variant: pre-/post-fix, binutils < / >= 2.29)
. an "intermediate" library (something that the end executable links
with) triggering QB_LOG_INIT_DATA macro and being built with
pre-fix libqb (and perhaps only with binutils < 2.29)
. end executable using no libqb's logging at all, but being built
with post-fix libqb (and arbitrary binutils < / >= 2.29)
then, unlike when executable is built with pre-fix libqb, the
special callsite data containing section in the ELF structure
of the executable is created + its boundary denoting symbols
defined within, despite the section being empty (did not happen
with pre-fix libqb), and because the symbols defined within the
target program have priority over that of shared libraries in the
symbol resolution fallback scheme, the assertion of QB_LOG_INIT_DATA
of the mentioned intermediate library will actually be evaluating
the inequality of boundaries for the section of the executable(!)
rather than it's own (or whatever higher prio symbols are hit,
presumably only present if the section at that level is non-empty,
basically a generalization of the story so far);
the problem then manifests as unability to run said executable
as it will fail because of the intermediate library inflicted
assertion (sadly with very unhelpful "Assertion `0' failed"
message);
fortunately, there's enough flexibility so as how to fix
this, either should be fine:
. have everything in the executable's library dependency closure
that links against libqb assurably (compile-time) linked with one
variant of libqb only (either all pre-fix or post-fix, mind the
apparent limitation of binutils' versions with the former)
. have the end executable (that does not use logging at all as
discussed precondition) linked using substitution like this:
s/-lqb/-l:libqb.so.0/ (you may need to adapt the number later)
and you may also need to add this CPPFLAG for the executable:
-DQB_KILL_ATTRIBUTE_SECTION
- as high level of isolation of the client space from the linker
(respectively toolchain) subtleties as possible (no new compilation
flags and such required, plus there's no way to hook any dynamic
computational ad-hoc decision when the compilation is about to
happen, anyway), and in turn, versatility is preserved as much as
possible
* * *
Finally, let's have a look how the already well-known test matrix
overview changes as of this commit, but first as a recap,
"X(Y)" denotes "X linked with linker Y":
@ -164,12 +222,19 @@ table stated in the preceding commit):
Everything is green \o/
* * *
Note: as of this fix, it is assumed that the non-green counterpart of
this table in the message for the preceding commit doubles as an
indicator how will mixing the logging participants wrt. linker+libqb
version work out, when "X(Y)" becomes read as "X linked with linker
Y under additional restriction on libqb version when compile-time link
is performed of the particular part":
this table in the message for the preceding commit (loosely though[!],
as the occurrence of empty callsite section can no longer be attributed
to something bad going on as of this fix that enforces its presence
unconditionally, whereas it would be suppressed when unused before
with kind linkers, hence some other conditions can be witnessed
especially when QB_LOG_INIT_DATA misused in no-logging context)
doubles as an indicator how will mixing the logging participants wrt.
linker+libqb version work out, when "X(Y)" becomes read as "X linked
with linker Y under additional restriction on libqb version when
compile-time link is performed of the particular part":
X(a) .. ld.bfd < 2.29 OR [arbitrary ld.bfd AND libqb after this fix)
X(b) .. ld.bfd = 2.29 (and likely on) AND libqb up to, but excluding
this fix
@ -186,6 +251,11 @@ Deficiencies:
- not solvable as long as we use the linker script, and there's
hardly any other way not requiring the libqb consumers to adapt
in any aspect
* as already mentioned, lacking compatibility with ld.gold linker and
won't foreseeably be (cf. https://bugzilla.redhat.com/1500898#c7)
- please stick with ld.bfd (i.e. default ld linker), which you
had to do in the past anyway (at least for compiling libqb
itself)
Open questions:
* should we enable attribute((__section__)) for powerpc and other minor
@ -193,7 +263,8 @@ Open questions:
and related to that, we need to figure out how static and dynamic call
sites can live together -- say what will happen to logging programs
compiled with libqb w/o attribute sections when the underlying libqb is
flipped to the one with them...
flipped to the one with them ... do we need a transition plan to be
spread across an extended period to keep it smooth?
* * *
@ -210,14 +281,14 @@ Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
---
configure.ac | 147 +++++++++++++++++++++++++++++++++++++++++------
include/qb/qbconfig.h.in | 2 +
include/qb/qblog.h | 37 ++++++++----
include/qb/qblog.h | 32 +++++++++--
lib/Makefile.am | 67 +++++++++++++++++++++
lib/libqb.pc.in | 8 +++
lib/log.c | 7 +++
lib/qblog_script.la.in | 15 +++++
lib/qblog_script.ld.in | 27 +++++++++
lib/qblog_script_noop.ld | 1 +
10 files changed, 281 insertions(+), 30 deletions(-)
9 files changed, 282 insertions(+), 24 deletions(-)
create mode 100644 lib/qblog_script.la.in
create mode 100644 lib/qblog_script.ld.in
create mode 100644 lib/qblog_script_noop.ld
@ -429,10 +500,22 @@ index c1852e1..b0bf428 100644
/* versioning info: MAJOR, MINOR, MICRO, and REST components */
#undef QB_VER_MAJOR
diff --git a/include/qb/qblog.h b/include/qb/qblog.h
index 10cd62c..07ae4c3 100644
index 766cffe..cd7256a 100644
--- a/include/qb/qblog.h
+++ b/include/qb/qblog.h
@@ -286,15 +286,17 @@ extern struct qb_log_callsite QB_ATTR_SECTION_STOP[];
@@ -79,7 +79,10 @@ extern "C" {
* scope in exactly one source file (preferably the main one) to be mixed into
* the resulting compilation unit. This is a self-defensive measure for when
* the linker-assisted collection of callsite data silently fails, which could
- * otherwise go unnoticed, causing troubles down the road.
+ * otherwise go unnoticed, causing troubles down the road, but alas it cannot
+ * discern misuse of @c QB_LOG_INIT_DATA() macro in no-logging context from
+ * broken callsite section handling assumptions owing to overboard fancy
+ * linker -- situation that the self-check aims to detect in the first place.
*
* @par Configuring log targets.
* A log target can be syslog, stderr, the blackbox, stdout, or a text file.
@@ -287,12 +290,13 @@ extern struct qb_log_callsite QB_ATTR_SECTION_STOP[];
/* Related to the next macro that is -- unlike this one -- a public API */
#ifndef _GNU_SOURCE
@ -440,26 +523,21 @@ index 10cd62c..07ae4c3 100644
+#define QB_NONAPI_LOG_INIT_DATA_EXTRA_(name) \
_Pragma(QB_PP_STRINGIFY(GCC warning QB_PP_STRINGIFY( \
without "_GNU_SOURCE" defined (directly or not) \
- qb_log_init_data cannot check sanity of libqb proper)))
+ qb_log_init_data cannot check sanity of libqb proper \
+ nor )))
- QB_LOG_INIT_DATA cannot check sanity of libqb proper)))
+ QB_LOG_INIT_DATA cannot check sanity of libqb proper \
+ nor of the target site originating this check alone)))
#else
-#define QB_NONAPI_LOG_INIT_DATA_EXTRA_ \
+#define QB_NONAPI_LOG_INIT_DATA_EXTRA_(name) \
{ Dl_info work_dli; \
/* libqb sanity (locating libqb by it's relatively unique \
-- and currently only such per-linkage global one -- \
non-functional symbol, due to possible confusion otherwise) */ \
+ { Dl_info work_dli; \
if (dladdr(dlsym(RTLD_DEFAULT, "facilitynames"), &work_dli) \
&& (work_handle = dlopen(work_dli.dli_fname, \
RTLD_LOCAL|RTLD_LAZY)) != NULL) { \
@@ -308,7 +310,22 @@ linkage at fault and logging would not work reliably" \
@@ -310,6 +314,23 @@ linkage at fault and logging would not work reliably" \
assert("libqb's callsite section non-empty, otherwise libqb's \
linkage at fault and logging would not work reliably" \
&& work_s1 != work_s2); \
- dlclose(work_handle); }
+ dlclose(work_handle); } \
+ /* finally, slightly modified original, straightforward check */ \
+ /* sanity of the target site originating this check alone */ \
+ if (dladdr(dlsym(RTLD_DEFAULT, QB_PP_STRINGIFY(name)), &work_dli) \
+ && (work_handle = dlopen(work_dli.dli_fname, \
+ RTLD_LOCAL|RTLD_LAZY)) != NULL) { \
@ -468,42 +546,26 @@ index 10cd62c..07ae4c3 100644
+ work_s2 = (struct qb_log_callsite *) \
+ dlsym(work_handle, QB_ATTR_SECTION_STOP_STR); \
+ assert("target's own callsite section observable, otherwise \
+target's own linkage at fault and logging would not work reliably" \
+target's own linkage at fault and logging would not work reliably \
+(unless QB_LOG_INIT_DATA macro used unexpectedly in no-logging context)"\
+ && work_s1 != NULL && work_s2 != NULL); \
+ assert("target's own callsite section non-empty, otherwise \
+target's own linkage at fault and logging would not work reliably" \
+target's own linkage at fault and logging would not work reliably \
+(unless QB_LOG_INIT_DATA macro used unexpectedly in no-logging context)"\
+ && work_s1 != work_s2); \
+ dlclose(work_handle); } }
dlclose(work_handle); } }
#endif /* _GNU_SOURCE */
/**
@@ -341,22 +358,18 @@ linkage at fault and logging would not work reliably" \
void name(void); \
void name(void) { \
void *work_handle; struct qb_log_callsite *work_s1, *work_s2; \
- Dl_info work_dli; \
- /* our own (target's) sanity */ \
+ /* our own (target's) sanity, or possibly that of higher level \
+ symbol resolution site */ \
if ((work_handle = dlopen(NULL, RTLD_LOCAL|RTLD_LAZY)) != NULL) { \
work_s1 = (struct qb_log_callsite *) \
dlsym(work_handle, QB_ATTR_SECTION_START_STR); \
work_s2 = (struct qb_log_callsite *) \
dlsym(work_handle, QB_ATTR_SECTION_STOP_STR); \
- assert("target's callsite section self-observable, otherwise \
+ assert("target's callsite section observable, otherwise \
@@ -354,8 +375,7 @@ linkage at fault and logging would not work reliably" \
target's and/or libqb's linkage at fault and logging would not work \
reliably" \
&& work_s1 != NULL && work_s2 != NULL); } \
- /* better targeted attestations when available */ \
- QB_NONAPI_LOG_INIT_DATA_EXTRA_; \
- /* finally, original, straightforward check */ \
- assert("non-empty implicit callsite section, otherwise target's \
-linkage at fault and logging would not work reliably" \
- && QB_ATTR_SECTION_START != QB_ATTR_SECTION_STOP); } \
+ QB_NONAPI_LOG_INIT_DATA_EXTRA_(name); } \
void __attribute__ ((constructor)) name(void);
#else
#define QB_LOG_INIT_DATA(name)
+ QB_NONAPI_LOG_INIT_DATA_EXTRA_(name); \
/* finally, original, straightforward check */ \
assert("non-empty implicit callsite section, otherwise target's \
linkage at fault and logging would not work reliably" \
diff --git a/lib/Makefile.am b/lib/Makefile.am
index 0bebeb5..ff5b3b4 100644
--- a/lib/Makefile.am
@ -618,7 +680,7 @@ index 37d27b7..65d3b8e 100644
Libs.private: @LIBS@
Cflags: -I${includedir}
diff --git a/lib/log.c b/lib/log.c
index 3e9972e..514103f 100644
index 777f378..b0c8735 100644
--- a/lib/log.c
+++ b/lib/log.c
@@ -40,6 +40,13 @@
@ -697,5 +759,5 @@ index 0000000..f037fca
@@ -0,0 +1 @@
+/* this is an empty linker script having a role of a NO-OP link object */
--
2.14.2
2.14.3

50
06-Low-fix-internal-object-symbol-s-leak-expose-run-tim.patch
View File

@ -1,4 +1,4 @@
From eb0c29d0213784597b7fdffb24e79d9f0774b8c7 Mon Sep 17 00:00:00 2001
From 4cd8bdcd3daca8781b5ee7bd8cdfdbcd5ba9cb92 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Pokorn=C3=BD?= <jpokorny@redhat.com>
Date: Fri, 6 Oct 2017 17:17:26 +0200
Subject: [PATCH 6/6] Low: fix internal object symbol's leak & expose run-time
@ -13,11 +13,13 @@ hence it's presumably safe to make it static as it's meant to be.
On the other hand, QB_LOG_INIT_DATA macro from qblog.h has already
started to depend on that symbol so as to locate the library handle
for libqb itself correctly. This is trivially fixed by finally exposing
library versioning info in run-time through objects a structure with
members corresponding to compile-time/static counterparts from
qbconfig.h header file, plus lower-cased equivalent of QB_VER_STR.
Better than to roll out a futile data object serving as an artificial
anchor for the above purpose, and this was a while due, afterall.
library versioning info in run-time ("online") as a structure with
members corresponding to compile-time ("offline") counterparts from
qbconfig.h header file, which are admittedly of very limited use
as opposed to the newly introduced dynamic info, plus lower-cased
equivalent of QB_VER_STR. Better than to roll out a futile data object
serving as an artificial anchor for the above purpose, and this was
due for a while, afterall.
In turn, also bump "current" and "age" of fields of the libtool's
"-version-info" versioning system.
@ -31,12 +33,12 @@ Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
lib/log.c | 10 ++++++----
lib/log_format.c | 2 +-
lib/util.c | 10 ++++++++++
tests/functional/log_client.c | 4 ++++
tests/functional/log_client.c | 5 +++++
tests/functional/log_test_client.err | 2 +-
9 files changed, 45 insertions(+), 11 deletions(-)
9 files changed, 46 insertions(+), 11 deletions(-)
diff --git a/include/qb/qbconfig.h.in b/include/qb/qbconfig.h.in
index b0bf428..19d14bf 100644
index b0bf428..7199b1b 100644
--- a/include/qb/qbconfig.h.in
+++ b/include/qb/qbconfig.h.in
@@ -32,7 +32,12 @@
@ -47,27 +49,26 @@ index b0bf428..19d14bf 100644
+/* versioning info: MAJOR, MINOR, MICRO, and REST components;
+ note that static compile-time info is not that useful as consulting
+ the respectively named members of qb_version struct constant under
+ @c qb_ver identifier (or @qb_ver_str equivalent of the local
+ @c qb_ver identifier (or @c qb_ver_str equivalent of the local
+ upper-cased value) directly from libqb in run-time (see qbutil.h),
+ but that was only introduced after v1.0.2 */
#undef QB_VER_MAJOR
#undef QB_VER_MINOR
#undef QB_VER_MICRO
diff --git a/include/qb/qblog.h b/include/qb/qblog.h
index 07ae4c3..a059fb5 100644
index cd7256a..4b18eba 100644
--- a/include/qb/qblog.h
+++ b/include/qb/qblog.h
@@ -294,10 +294,12 @@ extern struct qb_log_callsite QB_ATTR_SECTION_STOP[];
#else
@@ -299,9 +299,11 @@ extern struct qb_log_callsite QB_ATTR_SECTION_STOP[];
#define QB_NONAPI_LOG_INIT_DATA_EXTRA_(name) \
{ Dl_info work_dli; \
/* libqb sanity (locating libqb by it's relatively unique \
- -- and currently only such per-linkage global one -- \
- non-functional symbol, due to possible confusion otherwise) */ \
- if (dladdr(dlsym(RTLD_DEFAULT, "facilitynames"), &work_dli) \
+ non-functional symbols -- the two are mutually exclusive, the \
+ ordinarily latter was introduced by accident, the former is \
+ intentional -- due to possible confusion otherwise) */ \
{ Dl_info work_dli; \
- if (dladdr(dlsym(RTLD_DEFAULT, "facilitynames"), &work_dli) \
+ intentional -- due to possible confusion otherwise) */ \
+ if ((dladdr(dlsym(RTLD_DEFAULT, "qb_ver_str"), &work_dli) \
+ || dladdr(dlsym(RTLD_DEFAULT, "facilitynames"), &work_dli)) \
&& (work_handle = dlopen(work_dli.dli_fname, \
@ -109,7 +110,7 @@ index ff5b3b4..a428975 100644
source_to_lint = util.c hdb.c ringbuffer.c ringbuffer_helper.c \
array.c loop.c loop_poll.c loop_job.c \
diff --git a/lib/log.c b/lib/log.c
index 514103f..37f1e4e 100644
index b0c8735..32c9315 100644
--- a/lib/log.c
+++ b/lib/log.c
@@ -892,10 +892,12 @@ qb_log_init(const char *name, int32_t facility, uint8_t priority)
@ -123,7 +124,7 @@ index 514103f..37f1e4e 100644
+ the proper libqb's ones (locating libqb by it's relatively unique
+ non-functional symbols -- the two are mutually exclusive, the
+ ordinarily latter was introduced by accident, the former is
+ intentional -- due to possible confusion otherwise) */
+ intentional -- due to possible confusion otherwise) */
+ if ((dladdr(dlsym(RTLD_DEFAULT, "qb_ver_str"), &work_dli)
+ || dladdr(dlsym(RTLD_DEFAULT, "facilitynames"), &work_dli))
&& (work_handle = dlopen(work_dli.dli_fname,
@ -168,28 +169,29 @@ index 6181a25..a510bd1 100644
+
+const char *const qb_ver_str = QB_VER_STR;
diff --git a/tests/functional/log_client.c b/tests/functional/log_client.c
index c60d657..cb3e20b 100644
index c60d657..d5d8089 100644
--- a/tests/functional/log_client.c
+++ b/tests/functional/log_client.c
@@ -53,6 +53,10 @@ main(int32_t argc, char *argv[])
@@ -53,6 +53,11 @@ main(int32_t argc, char *argv[])
qb_log_tags_stringify_fn_set(my_tags_stringify);
qb_log_format_set(QB_LOG_STDERR, "[%5g|%p] %f:%l:%b");
+#if 0
+ printf("\n==%s==\n\n", qb_ver_str);
+ printf("\n==%s consists of: %d, %d, %d, %s==\n\n", qb_ver_str,
+ qb_ver.major, qb_ver.minor, qb_ver.micro, qb_ver.rest);
+#endif
+
#if 0
printf("--\n");
qb_log_callsites_dump();
diff --git a/tests/functional/log_test_client.err b/tests/functional/log_test_client.err
index cb0ff3c..29b9db0 100644
index cb0ff3c..7bc11b0 100644
--- a/tests/functional/log_test_client.err
+++ b/tests/functional/log_test_client.err
@@ -1,2 +1,2 @@
-[MAIN |debug] ../log_client.c:64:hello
+[MAIN |debug] ../log_client.c:68:hello
+[MAIN |debug] ../log_client.c:69:hello
[libqb|error] log_blackbox.c:196:qb_log_blackbox_print_from_file: No such file or directory (2)
--
2.14.2
2.14.3

15
libqb.spec
View File

@ -2,7 +2,7 @@
Name: libqb
Version: 1.0.2
Release: 11%{?dist}
Release: 12%{?dist}
Summary: An IPC library for high performance servers
Group: System Environment/Libraries
@ -77,21 +77,24 @@ developing applications that use %{name}.
%{_mandir}/man3/qb*3*
%changelog
* Wed Oct 25 2017 Jan Pokorný <jpokorny+rpm-libqb@redhat.com> - 1.0.2-12
- Evolution of the previous (rhbz#1478089)
* Wed Oct 18 2017 Jan Pokorný <jpokorny+rpm-libqb@redhat.com> - 1.0.2-11
- Evolution of the previous (rhbz#1478089, rhbz#1487787)
- Evolution of the previous (rhbz#1478089)
* Fri Oct 13 2017 Jan Pokorný <jpokorny+rpm-libqb@redhat.com> - 1.0.2-10
- Evolution of the previous (rhbz#1478089, rhbz#1487787)
- Evolution of the previous (rhbz#1478089)
* Mon Oct 09 2017 Jan Pokorný <jpokorny+rpm-libqb@redhat.com> - 1.0.2-9
- Evolution of the previous (rhbz#1478089, rhbz#1487787)
- Evolution of the previous (rhbz#1478089)
* Fri Oct 06 2017 Jan Pokorný <jpokorny+rpm-libqb@redhat.com> - 1.0.2-8
- Evolution of the previous (rhbz#1478089, rhbz#1487787)
- Evolution of the previous (rhbz#1478089)
- New test included in check phase (as per upsteam)
* Mon Sep 04 2017 Jan Pokorný <jpokorny+rpm-libqb@redhat.com> - 1.0.2-7
- Evolution of the previous (rhbz#1478089, rhbz#1487787)
- Evolution of the previous (rhbz#1478089)
* Fri Sep 01 2017 Jan Pokorný <jpokorny+rpm-libqb@redhat.com> - 1.0.2-6
- Better approach so as to mitigate changed treatment of orphaned sections