From kl_sadi at esi.dz Thu Jan 2 17:14:16 2025 From: kl_sadi at esi.dz (Lina SADI) Date: Thu, 2 Jan 2025 23:14:16 +0100 Subject: Tracing Ceph with LTTng-UST issue Message-ID: Hello, We are writing to you regarding an issue we encountered while attempting to trace Ceph with LTTng-UST. Below are the steps we have followed so far: 1. *Compiling Ceph with LTTng Support*: - We modified the debian/rules file to enable LTTng support using the following flags: -DWITH_LTTNG=ON and -DWITH_OSD_INSTRUMENT_FUNCTIONS=ON. - The compilation process took approximately 7 hours on a virtual machine running Debian 12 with a kernel version 6.1.0-27-amd64, 8 GB of RAM, and 2 CPU cores. - The build steps can be found in our GitHub repository. 2. *Deploying a Ceph Cluster*: - We deployed the cluster manually (1MON,1MGR,2OSDs). Please find the script here . 3. *Attempting to Trace with LTTng-UST*: - We tried listing the tracepoints using the following command : # lttng list -u - This command did not return any Ceph-related tracepoints: UST events: ------------- NONE 4. *Preloading liblttng-ust-fork.so*: - We attempted to preload liblttng-ust-fork.so before starting the Ceph daemons (MON, OSD, MGR) as follows: # systemctl edit ceph-mon at debian - We added the following configuration: [Service] Environment="LD_PRELOAD=/usr/lib/x86_64-linux-gnu/liblttng-ust-fork.so - After restarting the daemons, we executed lttng list -u, which displayed tracepoints related to the libraries (lttng_ust_lib:*), trace logs (lttng_ust_tracelog:*), and state dump events ( lttng_ust_statedump:*), instead of the expected Ceph-related tracepoints. *Expected Result*: UST events: ------------- PID: 100859 - Name: /path/to/ceph-osd pg:queue_op (loglevel: TRACE_DEBUG_LINE (13)) (type: tracepoint) osd:do_osd_op_post (loglevel: TRACE_DEBUG_LINE (13)) (type: tracepoint) osd:do_osd_op_pre_unknown (loglevel: TRACE_DEBUG_LINE (13)) (type: tracepoint) osd:do_osd_op_pre_copy_from (loglevel: TRACE_DEBUG_LINE (13)) (type: tracepoint) osd:do_osd_op_pre_copy_get (loglevel: TRACE_DEBUG_LINE (13)) (type: tracepoint) ... *Actual Result*: UST events: ------------- lttng_ust_lib:* lttng_ust_tracelog:* lttng_ust_statedump:* We would greatly appreciate any help you could provide to resolve this problem. -------------- next part -------------- An HTML attachment was scrubbed... URL: From kstewart at efficios.com Mon Jan 6 11:33:09 2025 From: kstewart at efficios.com (Kienan Stewart) Date: Mon, 6 Jan 2025 11:33:09 -0500 Subject: Tracing Ceph with LTTng-UST issue In-Reply-To: References: Message-ID: <023def9c-ff2a-421c-8e89-b6036fb14668@efficios.com> Hi Lina, I have a couple of questions about your tracing setup. On 1/2/25 5:14 PM, Lina SADI via lttng-dev wrote: > Hello, > > We are writing to you regarding an issue we encountered while attempting to > trace Ceph with LTTng-UST. Below are the steps we have followed so far: > > 1. > > *Compiling Ceph with LTTng Support*: > - We modified the debian/rules file to enable LTTng support using the > following flags: -DWITH_LTTNG=ON and > -DWITH_OSD_INSTRUMENT_FUNCTIONS=ON. > - The compilation process took approximately 7 hours on a virtual > machine running Debian 12 with a kernel version 6.1.0-27-amd64, 8 GB of > RAM, and 2 CPU cores. > - The build steps can be found in our GitHub > > repository. > 2. > > *Deploying a Ceph Cluster*: > - We deployed the cluster manually (1MON,1MGR,2OSDs). Please find the > script here > > . With which user a running lttng-sessiond? When you run `lttng list -u`, what user are you using? Is that user a member of the `tracing` group? If you are running lttng-sessiond with verbose logs and/or verbose consumers - do you see the traced application(s) connecting? > 3. > > *Attempting to Trace with LTTng-UST*: > - We tried listing the tracepoints using the following command : > > # lttng list -u > > - This command did not return any Ceph-related tracepoints: > > UST events: > ------------- > NONE > > 4. > > *Preloading liblttng-ust-fork.so*: > - We attempted to preload liblttng-ust-fork.so before starting the Ceph > daemons (MON, OSD, MGR) as follows: > > # systemctl edit ceph-mon at debian > > - We added the following configuration: > > [Service] > Environment="LD_PRELOAD=/usr/lib/x86_64-linux-gnu/liblttng-ust-fork.so > > > Does adding `LTTNG_UST_DEBUG=1` to the environment variables add information to the service logs that are relevant? > - After restarting the daemons, we executed lttng list -u, which > displayed tracepoints related to the libraries (lttng_ust_lib:*), > trace logs (lttng_ust_tracelog:*), and state dump events ( > lttng_ust_statedump:*), instead of the expected Ceph-related > tracepoints. > > *Expected Result*: > > UST events: > ------------- > PID: 100859 - Name: /path/to/ceph-osd > pg:queue_op (loglevel: TRACE_DEBUG_LINE (13)) (type: tracepoint) > osd:do_osd_op_post (loglevel: TRACE_DEBUG_LINE (13)) (type: tracepoint) > osd:do_osd_op_pre_unknown (loglevel: TRACE_DEBUG_LINE (13)) > (type: tracepoint) > osd:do_osd_op_pre_copy_from (loglevel: TRACE_DEBUG_LINE (13)) > (type: tracepoint) > osd:do_osd_op_pre_copy_get (loglevel: TRACE_DEBUG_LINE (13)) > (type: tracepoint) > ... > > > *Actual Result*: > > UST events: > ------------- > lttng_ust_lib:* > lttng_ust_tracelog:* > lttng_ust_statedump:* > > > We would greatly appreciate any help you could provide to resolve this problem. > thanks, kienan From martin at sr-research.com Wed Jan 15 09:23:00 2025 From: martin at sr-research.com (Martin Hicks) Date: Wed, 15 Jan 2025 09:23:00 -0500 Subject: [PATCH] lttng-modules fix for RHEL 9.5 kernels Message-ID: <20250115142300.vlupsqd2lr5bipfo@martin-pc> This fixes the asoc_snd_*, kfree_skb and udp related tracepoints for the RHEL 9.5 distro kernel. In all cases they've updated to the format of newer kernels. Signed-off-by: Martin Hicks --- include/instrumentation/events/asoc.h | 8 ++++++-- include/instrumentation/events/skb.h | 4 +++- include/instrumentation/events/udp.h | 4 +++- 3 files changed, 12 insertions(+), 4 deletions(-) diff --git a/include/instrumentation/events/asoc.h b/include/instrumentation/events/asoc.h index f6b2c76..31bedfe 100644 --- a/include/instrumentation/events/asoc.h +++ b/include/instrumentation/events/asoc.h @@ -120,7 +120,9 @@ LTTNG_TRACEPOINT_EVENT_INSTANCE_MAP(asoc_snd_soc_preg, snd_soc_preg_read, ) #endif -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0)) +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0) || \ + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) + LTTNG_TRACEPOINT_EVENT_CLASS(asoc_snd_soc_dapm_context, TP_PROTO(struct snd_soc_dapm_context *dapm, int val), @@ -187,7 +189,9 @@ LTTNG_TRACEPOINT_EVENT_INSTANCE_MAP(asoc_snd_soc_card, snd_soc_bias_level_done, ) #endif -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0)) +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0) || \ + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) + LTTNG_TRACEPOINT_EVENT_CLASS(asoc_snd_soc_dapm_basic, TP_PROTO(struct snd_soc_card *card, int event), diff --git a/include/instrumentation/events/skb.h b/include/instrumentation/events/skb.h index 5556ab6..becfd80 100644 --- a/include/instrumentation/events/skb.h +++ b/include/instrumentation/events/skb.h @@ -40,7 +40,9 @@ LTTNG_TRACEPOINT_ENUM(skb_drop_reason, ) #endif -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,11,0)) +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,11,0) || \ + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) + LTTNG_TRACEPOINT_EVENT_MAP(kfree_skb, skb_kfree, diff --git a/include/instrumentation/events/udp.h b/include/instrumentation/events/udp.h index 0a94b46..c228ef6 100644 --- a/include/instrumentation/events/udp.h +++ b/include/instrumentation/events/udp.h @@ -9,7 +9,9 @@ #include #include -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,10,0)) +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,10,0) || \ + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) + LTTNG_TRACEPOINT_ENUM(lttng_sk_family, TP_ENUM_VALUES( ctf_enum_value("AF_INET", AF_INET) -- 2.43.5 -- Martin Hicks SR Research Ltd. Fast, Accurate, Reliable Eye Tracking From mathieu.desnoyers at efficios.com Tue Jan 21 14:56:28 2025 From: mathieu.desnoyers at efficios.com (Mathieu Desnoyers) Date: Tue, 21 Jan 2025 14:56:28 -0500 Subject: [PATCH] lttng-modules fix for RHEL 9.5 kernels In-Reply-To: <20250115142300.vlupsqd2lr5bipfo@martin-pc> References: <20250115142300.vlupsqd2lr5bipfo@martin-pc> Message-ID: <7127ac1b-7cc5-4f98-afe8-1244c3a5056b@efficios.com> On 2025-01-15 09:23, Martin Hicks via lttng-dev wrote: > > This fixes the asoc_snd_*, kfree_skb and udp related tracepoints for the > RHEL 9.5 distro kernel. > > In all cases they've updated to the format of newer kernels. I notice that this patch appears to be based on stable-2.13 rather than the master branch. I will apply it to master and forward port it to master, as the cherry-pick seems trivial, but ideally please submit patches against the master branch of lttng-modules in the future. Thanks! Mathieu > > Signed-off-by: Martin Hicks > --- > include/instrumentation/events/asoc.h | 8 ++++++-- > include/instrumentation/events/skb.h | 4 +++- > include/instrumentation/events/udp.h | 4 +++- > 3 files changed, 12 insertions(+), 4 deletions(-) > > diff --git a/include/instrumentation/events/asoc.h b/include/instrumentation/events/asoc.h > index f6b2c76..31bedfe 100644 > --- a/include/instrumentation/events/asoc.h > +++ b/include/instrumentation/events/asoc.h > @@ -120,7 +120,9 @@ LTTNG_TRACEPOINT_EVENT_INSTANCE_MAP(asoc_snd_soc_preg, snd_soc_preg_read, > ) > #endif > > -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0)) > +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0) || \ > + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) > + > LTTNG_TRACEPOINT_EVENT_CLASS(asoc_snd_soc_dapm_context, > > TP_PROTO(struct snd_soc_dapm_context *dapm, int val), > @@ -187,7 +189,9 @@ LTTNG_TRACEPOINT_EVENT_INSTANCE_MAP(asoc_snd_soc_card, snd_soc_bias_level_done, > ) > #endif > > -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0)) > +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,9,0) || \ > + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) > + > LTTNG_TRACEPOINT_EVENT_CLASS(asoc_snd_soc_dapm_basic, > > TP_PROTO(struct snd_soc_card *card, int event), > diff --git a/include/instrumentation/events/skb.h b/include/instrumentation/events/skb.h > index 5556ab6..becfd80 100644 > --- a/include/instrumentation/events/skb.h > +++ b/include/instrumentation/events/skb.h > @@ -40,7 +40,9 @@ LTTNG_TRACEPOINT_ENUM(skb_drop_reason, > ) > #endif > > -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,11,0)) > +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,11,0) || \ > + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) > + > LTTNG_TRACEPOINT_EVENT_MAP(kfree_skb, > > skb_kfree, > diff --git a/include/instrumentation/events/udp.h b/include/instrumentation/events/udp.h > index 0a94b46..c228ef6 100644 > --- a/include/instrumentation/events/udp.h > +++ b/include/instrumentation/events/udp.h > @@ -9,7 +9,9 @@ > #include > #include > > -#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,10,0)) > +#if (LTTNG_LINUX_VERSION_CODE >= LTTNG_KERNEL_VERSION(6,10,0) || \ > + LTTNG_RHEL_KERNEL_RANGE(5,14,0,503,21,1, 5,15,0,0,0,0)) > + > LTTNG_TRACEPOINT_ENUM(lttng_sk_family, > TP_ENUM_VALUES( > ctf_enum_value("AF_INET", AF_INET) -- Mathieu Desnoyers EfficiOS Inc. https://www.efficios.com From martin at sr-research.com Wed Jan 22 08:08:52 2025 From: martin at sr-research.com (Martin Hicks) Date: Wed, 22 Jan 2025 08:08:52 -0500 Subject: [PATCH] lttng-modules fix for RHEL 9.5 kernels In-Reply-To: <7127ac1b-7cc5-4f98-afe8-1244c3a5056b@efficios.com> References: <20250115142300.vlupsqd2lr5bipfo@martin-pc> <7127ac1b-7cc5-4f98-afe8-1244c3a5056b@efficios.com> Message-ID: On Tue, Jan 21, 2025 at 2:56?PM Mathieu Desnoyers < mathieu.desnoyers at efficios.com> wrote: > On 2025-01-15 09:23, Martin Hicks via lttng-dev wrote: > > > > This fixes the asoc_snd_*, kfree_skb and udp related tracepoints for the > > RHEL 9.5 distro kernel. > > > > In all cases they've updated to the format of newer kernels. > > I notice that this patch appears to be based on stable-2.13 rather > than the master branch. > > I will apply it to master and forward port it to master, as the > cherry-pick seems trivial, but ideally please submit patches > against the master branch of lttng-modules in the future. > Ah, you're right. I'll do that in the future. Sorry about that. mh -- Martin Hicks SR Research Ltd. Fast, Accurate, Reliable Eye Tracking -------------- next part -------------- An HTML attachment was scrubbed... URL: From jeremie.galarneau at efficios.com Wed Jan 22 23:45:05 2025 From: jeremie.galarneau at efficios.com (=?iso-8859-1?Q?J=E9r=E9mie?= Galarneau) Date: Wed, 22 Jan 2025 23:45:05 -0500 Subject: [RELEASE] Babeltrace 2.1.0 Message-ID: Hi there! Five years to the day [1] after launching Babeltrace 2.0 ?Amqui?, we're releasing Babeltrace 2.1 ?Brossard?! ? ???????????????????????????????????????????????????????????????????????????????????? ? NOTE: An HTML version of those release notes is available here: ? ? . ? ???????????????????????????????????????????????????????????????????????????????????? BACKGROUND ?????????? The DiaMon Workgroup [2] worked on the CTF 2 specification [3] from 2016 until its publication in March 2024. CTF 2 is a major revision of CTF 1.8 [4], bringing many improvements, such as: ? Using JSON text sequences [5] for the metadata stream. ? Simplifying the metadata stream. ? Adding new field classes (bit array, bit map, boolean, LEB128 [6], BLOB, and optional) and improving existing ones. ? Supporting UTF-16 and UTF-32 string fields. ? Using predefined roles instead of reserved structure member names to identify meaningful fields. ? Adding the attribute and extension features to extend and customize the format. EfficiOS [7] is the primary organization driving the development of both the LTTng [8] and Babeltrace projects. As of this date, the LTTng project intends to add CTF 2 production support soon, potentially enabling the addition of new features that aren't currently possible due to the limitations of CTF 1.8. Consequently, it seemed important to us that Babeltrace support this format and that it be released before any tracer officially adopts CTF 2. Although the visible impacts for CLI users are quite minimal, adding CTF 2 support to the project required a significant amount of work on the core library (libbabeltrace2) and, consequently, on all the components of the project that depend on its C API, directly or indirectly: ? The Python bindings API. ? All the official plugins, especially `ctf`. ? The command-line interface. Of course, the documentation and testing of all those components were also updated and improved. WHAT'S NEW SINCE BABELTRACE 2.0? ???????????????????????????????? See babeltrace2-intro(7) [9] to learn more about Babeltrace 2. The following subsections list the user-visible changes. Library ??????? ? The most important change of libbabeltrace2 is the addition of a new Message Interchange Protocol (MIP) [10]. libbabeltrace2 2.0 only offers MIP 0 which covers the message contents you're used to since 2020. Starting from Babeltrace 2.1, _MIP 1_ is available and adds many functions to support CTF 2 features. Moreover, MIP 1 introduces some important new concepts to the clock class [11] API, making some functions exclusive to either MIP 0 or MIP 1. In summary, the specific MIP 1 changes are: ? Clock class [11], event class [12], stream class [13], and trace [14] objects may have a namespace, a name, and a unique ID (UID) [15]. The UUID property of clock class and trace objects is removed in favor of the new UID property. ? Important parts of the clock class API are changed. ? A bit array field class [16] may have flags [17]. ? An integer field class [18] may have field value hints [19]. ? There are new BLOB field classes [20] and fields [21]. ? The field location [22] system replaces the field path [23] one. See the new ?Add MIP 1 support to your component class? guide [24] to learn more about all the changes brought by MIP 1. ? Your component class must implement the ?get supported MIP versions? method [25] to support MIP 1. This method remains optional: without it, the component class instances only support MIP 0. ? There are new MIP version access functions: ? bt_clock_class_get_graph_mip_version() [26] ? bt_field_class_get_graph_mip_version() [27] ? bt_trace_class_get_graph_mip_version() [28] ? bt_self_message_iterator_get_graph_mip_version() [29] ? The new bt_get_greatest_operative_mip_version_with_restriction() [30] function is like the existing bt_get_greatest_operative_mip_version(), but you may specify MIP version restrictions (a set of allowed MIP version ranges). Python bindings ??????????????? The Python bindings wrap all the library changes in the same systematic way as in Babeltrace 2.0. More specifically: ? Add type hints to the public API to assist with static analysis of your applications. This is currently an ongoing effort, but we believe we've made significant progress in this direction to improve application code quality. The new type hints also serve as typing documentation. The `bt2` package has its own internal `typing` module to work with Python 3.4 as the official `typing` module [31] was introduced in Python 3.5. ? Add new names and functions to address naming inconsistencies: ??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? ? Deprecated name/function ? New name/function ? ??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? ? `_DynamicArrayWithLengthFieldFieldClass` ? `_DynamicArrayFieldClassWithLengthField` ? ? `_DynamicArrayWithLengthFieldFieldClassConst` ? `_DynamicArrayFieldClassWithLengthFieldConst` ? ? `_OptionWithBoolSelectorFieldClass` ? `_OptionFieldClassWithBoolSelectorField` ? ? `_OptionWithBoolSelectorFieldClassConst` ? `_OptionFieldClassWithBoolSelectorFieldConst` ? ? `_OptionWithIntegerSelectorFieldClass` ? `_OptionFieldClassWithIntegerSelectorField` ? ? `_OptionWithIntegerSelectorFieldClassConst` ? `_OptionFieldClassWithIntegerSelectorFieldConst` ? ? `_OptionWithSelectorFieldClassConst` ? `_OptionFieldClassWithSelectorField` ? ? `_OptionWithSelectorFieldClassConst` ? `_OptionFieldClassWithSelectorFieldConst` ? ? `_OptionWithSignedIntegerSelectorFieldClass` ? `_OptionFieldClassWithSignedIntegerSelectorField` ? ? `_OptionWithSignedIntegerSelectorFieldClassConst` ? `_OptionFieldClassWithSignedIntegerSelectorFieldConst` ? ? `_OptionWithUnsignedIntegerSelectorFieldClass` ? `_OptionFieldClassWithUnsignedIntegerSelectorField` ? ? `_OptionWithUnsignedIntegerSelectorFieldClassConst` ? `_OptionFieldClassWithUnsignedIntegerSelectorFieldConst` ? ? `_TraceClass.create_option_with_bool_selector_field_class() ? _TraceClass.create_option_field_class_with_bool_selector_field() ? ? `_TraceClass.create_option_with_bool_selector_field_class() ? _TraceClass.create_option_field_class_with_bool_selector_field() ? ? `_TraceClass.create_option_with_integer_selector_field_class() ? _TraceClass.create_option_field_class_with_integer_selector_field() ? ? `_TraceClass.create_option_without_selector_field_class() ? _TraceClass.create_option_field_class_without_selector_field() ? ? `_VariantFieldClassWithIntegerSelector` ? `_VariantFieldClassWithIntegerSelectorField` ? ? `_VariantFieldClassWithIntegerSelectorConst` ? `_VariantFieldClassWithIntegerSelectorFieldConst` ? ? `_VariantFieldClassWithoutSelector` ? `_VariantFieldClassWithoutSelectorField` ? ? `_VariantFieldClassWithoutSelectorConst` ? `_VariantFieldClassWithoutSelectorFieldConst` ? ? `_VariantFieldClassWithSignedIntegerSelector` ? `_VariantFieldClassWithSignedIntegerSelectorField` ? ? `_VariantFieldClassWithSignedIntegerSelectorConst` ? `_VariantFieldClassWithSignedIntegerSelectorFieldConst` ? ? `_VariantFieldClassWithUnsignedIntegerSelector` ? `_VariantFieldClassWithUnsignedIntegerSelectorField` ? ? `_VariantFieldClassWithUnsignedIntegerSelectorConst` ? `_VariantFieldClassWithUnsignedIntegerSelectorFieldConst` ? ? `ClockOffset` ? `ClockClassOffset` ? ??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? ? Make some types which are useful for type hints public: ? `_ClockClass` ? `_ClockClassConst` ? `_DoublePrecisionRealFieldClass` ? `_DoublePrecisionRealFieldClassConst` ? `_DynamicArrayFieldWithLengthField` ? `_DynamicArrayFieldWithLengthFieldConst` ? `_EventClass` ? `_EventClassConst` ? `_Field` ? `_FieldClass` ? `_FieldClassConst` ? `_FieldConst` ? `_FieldPathConst` ? `_FieldPathConst` ? `_FieldPathItem` ? `_InputPortConst` ? `_ListenerHandle` ? `_MessageConst` ? `_MessageIteratorConfiguration` ? `_OptionFieldWithBoolSelectorField` ? `_OptionFieldWithBoolSelectorFieldConst` ? `_OptionFieldWithSignedIntegerSelectorField` ? `_OptionFieldWithSignedIntegerSelectorFieldConst` ? `_OptionFieldWithUnsignedIntegerSelectorField` ? `_OptionFieldWithUnsignedIntegerSelectorFieldConst` ? `_OutputPortConst` ? `_PluginSet` ? `_PrivateQueryExecutor` ? `_SignedEnumerationFieldClassMappingConst` ? `_SinglePrecisionRealFieldClass` ? `_SinglePrecisionRealFieldClassConst` ? `_Stream` ? `_Stream` ? `_StreamClass` ? `_StreamClassConst` ? `_StreamConst` ? `_StructureFieldClassMember` ? `_StructureFieldClassMemberConst` ? `_Trace` ? `_TraceClass` ? `_TraceClassConst` ? `_TraceConst` ? `_UnsignedEnumerationFieldClassMappingConst` ? `_UserComponentInputPort` ? `_UserComponentInputPortMessageIterator` ? `_UserComponentOutputPort` ? `_UserFilterComponentConfiguration` ? `_UserSinkComponentConfiguration` ? `_UserSourceComponentConfiguration` ? `_Value` ? `_ValueConst` ? `_VariantFieldClassOption` ? `_VariantFieldClassOptionConst` ? `_VariantFieldClassWithIntegerSelectorFieldOption` ? `_VariantFieldClassWithIntegerSelectorFieldOptionConst` ? `_VariantFieldClassWithSignedIntegerSelectorFieldOption` ? `_VariantFieldClassWithSignedIntegerSelectorFieldOptionConst` ? `_VariantFieldClassWithUnsignedIntegerSelectorFieldOption` ? `_VariantFieldClassWithUnsignedIntegerSelectorFieldOptionConst` ? `_VariantFieldWithSignedIntegerSelectorField` ? `_VariantFieldWithSignedIntegerSelectorFieldConst` ? `_VariantFieldWithUnsignedIntegerSelectorField` ? `_VariantFieldWithUnsignedIntegerSelectorFieldConst` ? Make some types inherit `enum.Enum`: ? `ComponentClassType` ? `EventClassLogLevel` ? `IntegerDisplayBase` ? `FieldPathScope` ? `LoggingLevel` ? Add a `members` parameter to _TraceClass.create_structure_field_class() to be able to create a complete structure field class at once. Also make _StructureFieldClass.__iadd__() accept a tuple of three elements, including optional member user attributes: my_struct_fc = my_trace_cls.create_structure_field_class(members=[ ('uid', my_trace_cls.create_string_field_class()), ('count', my_int_fc, {'context': 'bz.994', 'max': 2**16 - 1}), ('names', my_trace_cls.create_dynamic_array_field_class( my_trace_cls.create_string_field_class())), ? Add an `options` parameter to _TraceClass.create_variant_field_class() to be able to create a complete variant field class at once. Also make the various __iadd__() methods of variant field class classes accept a tuple of three elements, including optional option user attributes: my_var_fc = my_trace_cls.create_variant_field_class(options=[ ('i8', my_trace_cls.create_signed_integer_field_class(8)), ('i16', my_trace_cls.create_signed_integer_field_class(16)), ('i24', my_trace_cls.create_signed_integer_field_class(24), { 'unorthodox': true, }), ('i32', my_trace_cls.create_signed_integer_field_class(32)), ('i64', my_trace_cls.create_signed_integer_field_class(64)), ]) ? Add a `mappings` parameter to _TraceClass.create_unsigned_enumeration_field_class() and _TraceClass.create_signed_enumeration_field_class() to be able to create a complete enumeration field class at once: my_enum_fc = my_trace_cls.create_unsigned_enumeration_field_class(32, mappings=[ ('meow', bt2.UnsignedIntegerRangeSet([34])), ('mix', bt2.UnsignedIntegerRangeSet([(3, 17), (99, 101)])), ('????', bt2.UnsignedIntegerRangeSet([(1, 2), 13])), ]) ? Add the `graph_mip_version` property to the `_ClockClassConst`, `_EventClassConst`, `_FieldConst`, `_FieldClassConst`, `_StreamClassConst`, `_TraceConst`, and `_TraceClassConst` classes. ? Add MIP 1 support: ? Add the `namespace`, `name`, and `uid` read-only properties to the `_ClockClassConst`, `_TraceConst`, `_StreamClassConst`, and `_EventClassConst` classes. Add the `namespace`, `name`, and `uid` parameters to the _TraceClass.__call__(), _UserComponent._create_clock_class(), _Trace.create_stream(), and _StreamClass.create_event_class() methods. ? Make the `precision` parameter of _UserComponent._create_clock_class() optional. The `_ClockClassConst.precision` property may now return `None`. ? Add the `ClockOrigin` class, which you can instantiate directly to make a custom clock origin. The `unix_epoch_clock_origin` and `unknown_clock_origin` names are two provided, well-known clock origins. ? Add the `_ClockClassConst.accuracy`, `_ClockClassConst.origin_is_known`, and `_ClockClassConst.origin` properties. ? Make the `_ClockClassConst.uuid` property and the `uuid` parameter of _UserComponent._create_clock_class() exclusive to MIP 0. ? Add the `_FieldLocationConst` and `FieldLocationScope` classes to wrap the field location API. Create a new field location with _TraceClass.create_field_location(). Field class creation methods of the `_TraceClass` class which accepted an optional length/selector field class parameter, except for _TraceClass.create_variant_field_class(), now also accept a length/selector field location parameter: ? _TraceClass.create_dynamic_array_field_class() ? _TraceClass.create_option_field_class_with_bool_selector_field() ? _TraceClass.create_option_field_class_with_unsigned_integer_selector_field() ? _TraceClass.create_option_field_class_with_signed_integer_selector_field() Both parameters of a given method are mutually exclusive. _TraceClass.create_variant_field_class() is now exclusive to MIP 0. With MIP 1, use one of the following specific variant field class creation methods: ? _TraceClass.create_variant_field_class_without_selector_field() ? _TraceClass.create_variant_field_class_with_unsigned_integer_selector_field() ? _TraceClass.create_variant_field_class_with_signed_integer_selector_field() Make the `_OptionFieldClassWithSelectorFieldConst.selector_field_path`, `_VariantFieldClassWithIntegerSelectorFieldConst.selector_field_path`, and `_DynamicArrayFieldClassWithLengthFieldConst.length_field_path` properties exclusive to MIP 0. Add the `_OptionFieldClassWithSelectorFieldConst.selector_field_location`, `_VariantFieldClassWithIntegerSelectorFieldConst.selector_field_location`, and `_DynamicArrayFieldClassWithLengthFieldConst.length_field_location` read-only properties to access length/selector field locations. ? Add the `_BlobFieldClassConst`, `_BlobFieldClass`, `_StaticBlobFieldClassConst`, `_StaticBlobFieldClass`, `_DynamicBlobFieldClassConst`, `_DynamicBlobFieldClass`, `_DynamicBlobFieldClassWithLengthFieldConst`, `_DynamicBlobFieldClassWithLengthField` classes to wrap the BLOB field class API. Those are similar to their array field class equivalents, but the only specific BLOB property is an optional IANA media type [32]. Create BLOB field classes with the _TraceClass.create_static_blob_field_class() and _TraceClass.create_dynamic_blob_field_class() methods. ? Add the `_BlobFieldConst`, `_BlobField`, `_StaticBlobFieldConst`, `_StaticBlobField`, `_DynamicBlobFieldConst`, `_DynamicBlobField`, `_DynamicBlobFieldWithLengthFieldConst`, and `_DynamicBlobFieldWithLengthField` classes to wrap the BLOB field API. A BLOB field has a `data` property which returns a `memoryview` [33] object and which you can set to an instance of `bytes`, `bytearray`, or `memoryview`. ? Add the `_BitArrayFieldClassFlagConst` class to wrap the bit array field class flag API. `_BitArrayFieldClassFlagConst` is very similar to `_UnsignedEnumerationFieldClassMappingConst`. Add the _BitArrayFieldClassConst.__getitem__(), _BitArrayFieldClassConst.__len__(), _BitArrayFieldClass.add_flag(), and _BitArrayFieldClass.__iadd__() methods to access the optional flags of a bit array field class. ? Add the `_BitArrayFieldConst.active_flag_labels` and _BitArrayFieldClassConst.active_flags_for_value_as_integer() utilities. Those are similar to `_EnumerationFieldConst.labels` and _EnumerationFieldClassConst.mappings_for_value(). ? Make the `name` parameter of the append_option() method of the various variant field class classes optional. The `_VariantFieldClassOptionConst.name` property may now return `None`. Make _VariantFieldClassConst.__iter__() yield an index instead of an option name with MIP 1. Make the __getitem__() method of the various variant field class classes accept an index instead of an option name with MIP 1. Add the _VariantFieldClassConst.option_with_name() method to get an option by name, or `None` if none. ? Make the `_TraceConst.uuid` property and the `uuid` parameter of _TraceClass.__call__() exclusive to MIP 0. Command-line interface ?????????????????????? ? Add the `--allowed-mip-versions` [34] option to the `run` [35] and `convert` [36] commands. This option makes the trace processing graph only honour a specific version (0 or 1) of the Message Interchange Protocol (MIP) [10] instead of allowing both versions. The main purpose of this option is testing the implementation of project plugins for specific MIP versions. ? Show the trace format when printing LTTng live sessions: net://localhost:4859/host/salut/meow (timer = 100, 10 stream(s), 3 client(s) connected, CTF 1.8) net://localhost:4859/host/salut/mix (timer = 10, 42 stream(s), 4 client(s) connected, CTF 2) Plugins ??????? This section shows noteworthy changes to the project plugins. In general: ? The version of all the project plugins is now 2.0.0 instead of none. For example, here's the output of `babeltrace2 help ctf`: ctf: Path: /usr/lib/babeltrace2/plugins/babeltrace-plugin-ctf.so Version: 2.0.0 Description: CTF input and output Author: EfficiOS License: MIT Source component classes: 2 Filter component classes: 0 Sink component classes: 1 This is useful for a user application which uses such a plugin to know its features and default behaviour (see bt_plugin_get_version() [37]). ????????????????????????????????????????????????????????????????? ? NOTE: The plugin version is unrelated to the project version. ? ????????????????????????????????????????????????????????????????? ? Make all component classes support MIP 1, albeit sometimes conditionally to the initialization parameters. Component class specifics: `source.ctf.fs`: Add full CTF 2 support. CTF 2 support is only available with MIP 1. As such: ? Merge physical CTF 2 traces into a single logical one, to support the LTTng recording session rotation feature [38], using the namespace, name, and unique ID of traces (if available). ? Set the log level of a libbabeltrace2 event class from the value of the `log-level` attribute, under either the `lttng.org,2009` or `babeltrace.org,2020` namespace, of its corresponding CTF event record class. ? Set the EMF URI of a libbabeltrace2 event class from the value of the `emf-uri` attribute, under either the `lttng.org,2009` or `babeltrace.org,2020` namespace, of its corresponding CTF event record class. ? If the namespace NS, name NAME, and unique ID UID of the trace are available, then the trace ID part of an output port name is: {namespace: `NS`, name: `NAME`, uid: `UID`} If the namespace of the trace isn't available, but the name NAME and unique ID UID are, then it's: {name: `NAME`, uid: `UID`} Otherwise, it's the absolute directory path of the trace. ? For the `babeltrace.support-info` query object, if the namespace NS, name NAME, and unique ID UID of the trace are available, then the `group` property is: namespace: NS, name: NAME, uid: UID If the namespace of the trace isn't available, but the name NAME and unique ID UID are, then it's: name: NAME, uid: UID Otherwise, the entry doesn't exist. When reading a CTF 1.8 trace with MIP 1: ? For an LTTng trace: ? Set the namespace of any libbabeltrace2 trace to `lttng.org,2009`. ? Set the origin of any libbabeltrace2 clock class to the Unix epoch. ? If the UUID of a CTF trace is available, then set the UID of the corresponding libbabeltrace2 trace to the textual representation of the trace UUID and its name to an empty string. ? Set the UID of a libbabeltrace2 clock class to the textual representation of the UUID of its corresponding CTF clock class, if available. ? Make the accuracy of any libbabeltrace2 clock class unknown. See babeltrace2-source.ctf.fs(7) [39] for more details. `source.ctf.lttng-live`: Add full CTF 2 support. CTF 2 support is only available with MIP 1. A `source.ctf.lttng-live` component can connect to an LTTng relay daemon using the LTTng 2.4 or the upcoming LTTng 2.15 protocol. The latter is required for CTF 2 support. The ?get supported MIP versions? method [25] always reports to support MIP versions 0 and 1. If you attempt to make a `source.ctf.lttng-live` component read a CTF 2 recording session within a trace processing graph configured with MIP 0, then the message iterator will fail at initialization or ?next? method call time. For the `sessions` query object, add the `trace-format` property to each map of the returned array. This property is either `ctf-1.8` or `ctf-2.0`. See babeltrace2-source.ctf.lttng-live(7) [40] for more details. `sink.ctf.fs`: Add full CTF 2 support. CTF 2 support is only available with MIP 1, while CTF 1.8 support is only available with MIP 0. The new `ctf-version` initialization parameter controls the CTF version of the resulting traces: `1` or `2` (the default starting from Babeltrace 2.1). With the `ctf-version` parameter set to `2`: ? The component sets the `sink.ctf.fs` attribute, under the `babeltrace.org,2020` namespace, of the preamble fragment to `true`. ? The structure member class names of packet context field classes having a CTF 2 role have the metadata stream UUID as a prefix to avoid clashes with user member classes. For example, the packet content length member class could be named `afe9ed0e-ce73-413f-8956-83663bab2047-content_size`. ? The component translates an event class log level to the `log-level` attribute under the `babeltrace.org,2020` namespace. ? The component translates an event class EMF URI to the `emf-uri` attribute under the `babeltrace.org,2020` namespace as a JSON string. ? When the component needs to generate a length/selector field class F for dynamic field classes _without_ a length/selector field: ? The structure member name of F is unspecified. ? F contains the attribute `is-key-only`, under the `babeltrace.org,2020` namespace, set to `true`. ? The component doesn't translate trace and stream user attributes: a CTF 2 metadata stream has no way to represent them. To write CTF 1.8 traces using the `babeltrace2` [41] CLI tool: $ babeltrace2 /path/to/input --component=sink.ctf.fs \ --params='path="/path/to/output", ctf-version="1"' See babeltrace2-sink.ctf.fs(7) [42] for more details. `filter.utils.muxer`: Improve performance by using a heap instead of an array to sort messages by time. Our own experiments show that this version is slightly more efficient than the Babeltrace 2.0.6 one for a four-stream CTF trace scenario, but the performance gain improves quickly as you add streams (?RT? means ?run time?; average of two runs): ??????????????????????????????????????????????????????????????????????????? ? Version ? RT (4 streams) (s) ? RT (8 streams) (s) ? RT (40 streams) (s) ? ??????????????????????????????????????????????????????????????????????????? ? 2.0.6 ? 39.99 ? 45.28 ? 24.14 ? ? 2.1.0 ? 39.06 ? 39.51 ? 12.16 ? ??????????????????????????????????????????????????????????????????????????? Thanks to the Argonne Leadership Computing Facility [43] for sponsoring this feature! ? See babeltrace2-filter.utils.muxer(7) [44] for more details. `sink.text.pretty`: ? Add the `print-enum-flags` boolean initialization parameter (false by default). This new parameter makes a `sink.text.pretty` component try to print enumeration field values as a `|`-delimited list of bit flags, if it applies. When printing the value of an enumeration field _without any corresponding mapping_, the component splits the value into its individual bit values: if _each set bit_ has a corresponding mapping, then it considers that those mappings are in fact bit flags. If a set bit has no corresponding mapping, then the component prints ``. For example (`block_rq*` Linux kernel event record; structured output to make it more readable): [13:15:49.024354958] (+0.000003868) wilbrod block_rq_complete: { cpu_id = 4 }, { dev = 8388624, sector = 375490176, nr_sector = 360, error = 0, rwbs = ( "RWBS_FLAG_READ" | "RWBS_FLAG_RAHEAD" : container = 12 ) } This parameter is false by default to avoid any unexpected behaviour. To try this using the `babeltrace2` [41] CLI tool: $ babeltrace2 /path/to/trace --component=sink.text.pretty \ --params=print-enum-flags=yes Thanks to Genevi?ve Bastien for contributing this feature! ? ? A `sink.text.pretty` component now prints a comma-delimited list of _all_ the mapping labels of a given enumeration field value within curly braces: enum_field = ( { "bit0", "range1to3" } : container = 1 ) Thanks to Genevi?ve Bastien for this contribution! ? ? Print the data of BLOB fields: my-event: { my-blob = { 23 69 6e 63 6c 75 64 65 20 3c } } See babeltrace2-sink.text.pretty(7) [45] for more details. `sink.text.details`: ? Don't write the value of a trace environment entry or of a value object when it's an empty string. This is mostly to avoid trailing whitespaces in the output. ? Print field locations in a manner similar to field paths, except structure field member names replace indexes: Length field location [Event payload: walk, deep, surprise] ? Print BLOB field classes: my-blob: Static BLOB (Length 32, Media type `application/mathematica`) ? Print BLOB fields: my-blob: Static BLOB: Length 32: 44 6f 5b 50 72 69 6e 74 5b 22 48 65 6c 6c 6f 2c 20 57 6f 72 6c 64 22 5d 2c 20 7b 35 7d 5d 20 20 ? Print namespace, name, and UID properties of traces, clock classes, stream classes, and event classes: Trace `radical` (Namespace `chrysler`, UID `imperial`): Default clock class: Namespace: lttng.org,2009 Name: monotonic UID: boot-id:7d19f00b-f1c7-49ea-a0a1-5f1a6a062a29 Stream class `real` (Namespace `scion`, UID `tc`, ID 48): Event class `genetic` (Namespace `manic`, UID `gt`, ID 334): Add the `with-stream-class-namespace` and `with-uid` boolean initialization parameters to control the visibility of stream class namespaces and UIDs (both true by default). ? Print sorted bit array field class flags: Bit array (8 flags): Flags: MAP_ANONYMOUS: [5] MAP_DENYWRITE: [11] MAP_FIXED: [4] MAP_GROWSDOWN: [8] MAP_NORESERVE: [6] MAP_POPULATE: [15] MAP_PRIVATE: [1] MAP_SHARED: [0] ? Print new clock class properties: unknown precision, accuracy, and origin. Default clock class: Namespace: lttng.org,2009 Name: monotonic UID: boot-id:7d19f00b-f1c7-49ea-a0a1-5f1a6a062a29 User attributes: lttng.org,2009: tag: provigo original-index: 4 Description: Monotonic Clock Frequency (Hz): 1,000,000,000 Precision (cycles): 0 Accuracy (cycles): 1000 Offset from origin (s): 1,564,079,206 Offset from origin (cycles): 484,157,338 Origin: Namespace: quebec.ca,2017 Name: referendum UID: 1995 ? Print the field value hints property of integer field classes: Unsigned integer (64-bit, Base 10, Expect small values) See babeltrace2-sink.text.details(7) [46] for more details. FUTURE WORK ??????????? Without guaranteeing anything for Babeltrace 2.2, here are some improvement ideas that often come up in our discussions: ? Make the plugin provider API public so that anyone may implement their own libbabeltrace2 bindings. Currently, there's an intimate relation between libbabeltrace2 and the official Python bindings. As of this date, a patch set [47], written by a generous external contributor and which adds this feature, is already under review. ? Make it easier to create an efficient filter message iterator which needs to modify or augment event messages. Currently, if you need to modify an event message, considering that you cannot change an event class once you use it, you need to create a new event class, which means you also need to create a new stream class and a new trace class. You also need to copy all the metadata and data objects you don't alter. This is what `filter.lttng-utils.debug-info` [48] does. We do realize that trace-transforming filter message iterators represent an important Babeltrace 2 use case. While it's possible to write one for Babeltrace 2.1, it is cumbersome to do and the result is not as efficient as we'd like. ? Add a filter component class to drop messages (like events) based on a filter expression, similar to the `--filter` [49] option of the `lttng enable-event` command. ? Optimize specific project component classes, especially `source.ctf.fs` [39] and `source.ctf.lttng-live` [40]. For `source.ctf.fs`, implement the ?seek ns from origin? method using packet indexes to make a message iterator seek a specific time faster. STATISTICS ?????????? Since Babeltrace 2.0.0: Commits: 1438 Changed files: 1741 Diff: 227,712 insertions; 88,376 deletions Significant contributors: Simon Marchi, Philippe Proulx, Michael Jeanson, Francis Deslauriers, Erica Bugden, J?r?mie Galarneau, Mathieu Desnoyers, and Genevi?ve Bastien Programming languages as of Babeltrace 2.1.0: ?????????????????????????????????????????? ? Language ? Effective lines of code ? ?????????????????????????????????????????? ? C/C++ header ? 50,437 ? ? C ? 84,504 ? ? C++ ? 78,620 ? ? Python 3 ? 12,035 ? ?????????????????????????????????????????? IMPORTANT LINKS ??????????????? Babeltrace 2.1.0 tarball: Babeltrace website: Mailing list for support and development: IRC channel: `#lttng` on `irc.oftc.net` Git repository: GitHub project: Continuous integration: Code review: REFERENCES ?????????? [1]: https://babeltrace.org/docs/release-notes/babeltrace-2.0.0-release-notes.html [2]: https://diamon.org/ [3]: https://diamon.org/ctf/ [4]: https://diamon.org/ctf/v1.8.3/ [5]: https://datatracker.ietf.org/doc/html/rfc7464 [6]: https://en.wikipedia.org/wiki/LEB128 [7]: https://www.efficios.com/ [8]: https://lttng.org/ [9]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-intro.7 [10]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-intro.7/#doc-mip [11]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-clock-cls.html [12]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-ev-cls.html [13]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-stream-cls.html [14]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-trace.html [15]: https://en.wikipedia.org/wiki/Unique_identifier [16]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-fc.html#api-tir-fc-ba [17]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-fc.html#api-tir-fc-ba-prop-flags [18]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-fc.html#api-tir-fc-int [19]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-fc.html#api-tir-fc-int-prop-hints [20]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-fc.html#api-tir-fc-blob [21]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-field.html#api-tir-field-blob [22]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-field-loc.html [23]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-field-path.html [24]: https://babeltrace.org/docs/v2.1/libbabeltrace2/guide-mip-0-to-mip-1.html [25]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-comp-cls-dev.html#api-comp-cls-dev-meth-mip [26]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-clock-cls.html#ga8da11f39e67bab58848c71671c8192f0 [27]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-fc.html#ga3a3816edfcb31177574ddea4f456d657 [28]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-tir-trace-cls.html#gaedadf9f5626f5466b30653b47027d3ae [29]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-self-msg-iter.html#gaa2767e72e6b86835898c9030884be77c [30]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-msg.html#gab39e215853c38692cbc62a53d6107df9 [31]: https://docs.python.org/3/library/typing.html [32]: https://datatracker.ietf.org/doc/html/rfc2046 [33]: https://docs.python.org/3/library/stdtypes.html#memoryview [34]: https://babeltrace.org/docs/v2.1/man1/babeltrace2-run.1/#doc-opt--allowed-mip-versions [35]: https://babeltrace.org/docs/v2.1/man1/babeltrace2-run.1/ [36]: https://babeltrace.org/docs/v2.1/man1/babeltrace2-convert.1/ [37]: https://babeltrace.org/docs/v2.1/libbabeltrace2/group__api-plugin.html#gacbb8a7f96cdb85e5f5361289a133c8c6 [38]: https://lttng.org/docs/v2.13/#doc-session-rotation [39]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-source.ctf.fs.7 [40]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-source.ctf.lttng-live.7 [41]: https://babeltrace.org/docs/v2.1/man1/babeltrace2.1/ [42]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-sink.ctf.fs.7 [43]: https://www.alcf.anl.gov/ [44]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-filter.utils.muxer.7 [45]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-sink.text.pretty.7 [46]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-sink.text.details.7 [47]: https://review.lttng.org/c/babeltrace/+/11716/ [48]: https://babeltrace.org/docs/v2.1/man7/babeltrace2-filter.lttng-utils.debug-info.7 [49]: https://lttng.org/man/1/lttng-enable-event/v2.13/#doc-opt--filter From mathieu.desnoyers at efficios.com Fri Jan 24 15:54:55 2025 From: mathieu.desnoyers at efficios.com (Mathieu Desnoyers) Date: Fri, 24 Jan 2025 15:54:55 -0500 Subject: [RELEASE] LTTng-modules 2.13.17 (Linux kernel tracer) Message-ID: <6c4edb5a-5506-4cb4-a251-ed77fcb385ae@efficios.com> Hi, This is a stable release announcement for the LTTng kernel tracer, a kernel tracer for the Linux kernel. The most relevant news about this release is added support for the Linux kernel v6.13 and for the RHEL 9.5 kernels by LTTng-modules v2.13.17. We are aware that this release of LTTng-modules fails to build against sles15sp5 and Linux v6.13 on powerpc32 in our CI. We will investigate and fix these in a future release, but we aim to release promptly after Linux v6.13, so considering that LTTng-modules 2.13.x has never successfully built against those kernels, those are not considered release blockers. Detailed change log: 2025-01-24 (Global Belly Laugh Day) LTTng modules 2.13.17 * fix: include linux/fs.h for 'struct file' definition (v4.5) * fix: add missing check for __must_check 'lttng_file_ref_put()' (v6.13) * Fix: uprobe consumer handler signature changed in Linux 6.13.0-rc1 * Fix: lookup_fdget_rcu removed in Linux 6.13.0-rc1 * Fix: f_count replaced with f_ref in Linux 6.13.0-rc1 * Fix: sched_stat_runtime changed in Linux 6.6.66 * lttng-modules fix for RHEL 9.5 kernels Thanks, Mathieu Project website: https://lttng.org Documentation: https://lttng.org/docs Download link: https://lttng.org/download -- Mathieu Desnoyers EfficiOS Inc. https://www.efficios.com