lttng-modules.git
4 years agosyscalls: Make the flags and mode fields of open[at] enumerations
Geneviève Bastien [Wed, 11 Mar 2020 16:33:04 +0000 (12:33 -0400)] 
syscalls: Make the flags and mode fields of open[at] enumerations

The open and openat system call have a flags and mode fields, whose values
are defined in the linux/fcntl.h file. These fields are now
enumerations that can be read as a bit field enum, to make the values more
readable / meaningful.

Here's an example babeltrace output of the open system call:

[...] syscall_entry_openat: { cpu_id = 0 }, { dfd = -100,
filename = "/tmp/edg0_383407",
flags = ( "O_RDWR" | "O_CREAT" | "O_TRUNC" : container = 578 ),
mode = ( "S_IWOTH" | "S_IROTH" | "S_IWGRP" | "S_IRGRP" |
   "S_IWUSR" | "S_IRUSR" : container = 438 ) }

Change-Id: Id7a516670b03e52fc75f9ff3c6ba8114c61a3865
Signed-off-by: Geneviève Bastien <gbastien@versatic.net>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoFix: update kvm instrumentation for Ubuntu 4.15.0-92
Michael Jeanson [Mon, 23 Mar 2020 18:48:24 +0000 (14:48 -0400)] 
Fix: update kvm instrumentation for Ubuntu 4.15.0-92

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: Ib367b9a0ce3846f45313906e710a9a6d644e3955

4 years agoRecord event as soon as one filter evaluates to TRUE
Francis Deslauriers [Fri, 20 Mar 2020 20:23:17 +0000 (16:23 -0400)] 
Record event as soon as one filter evaluates to TRUE

This is a tiny optimization but it can reduce tracing cost in cases
where multiple filters are attached to a probe.

Since we want to trace an event if any of its filter expressions
evaluates to TRUE, there is no need to keep on evaluating expressions
once one is found to be TRUE.

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I6b5b7697869d3e25dc24d38dc39c34ae1d49dcc9

4 years agoAdd UDP and ICMP packet header information to the tracepoint
Florian Walbroel via lttng-dev [Tue, 17 Mar 2020 09:03:19 +0000 (10:03 +0100)] 
Add UDP and ICMP packet header information to the tracepoint

* UDP transport header
* ICMP transport header

Signed-off-by: Florian Walbroel <walbroel@silexica.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agostatedump: introduce file_table_address
Mathieu Desnoyers [Tue, 10 Mar 2020 19:00:29 +0000 (15:00 -0400)] 
statedump: introduce file_table_address

Currently the LTTng-modules statedump simply iterates over all processes
in the system and assumes all threads share the same file descriptor
table, which is only true if threads were created with clone
CLONE_FILES.

Directly invoking clone without the CLONE_FILES creates threads which
belong to the same process, but have their own file descriptor table.

Therefore, model-wise, we cannot assume that all threads in a process
have the same fd table content.

Add a new "file_table_address" field to the lttng_statedump_process_state
event, which dumps the address of the thread's struct files_struct
pointer. This pointer is guaranteed to never be re-used while we hold
the RCU read-side lock (so for the entire iteration over
processes/threads).

For the lttng_statedump_file_descriptor event, remove the "pid" field
(which is semantically inaccurate) and add a "file_table_address" field,
which contains the struct files_struct address of the file table
containing the file descriptor.

An optimization is performed to eliminate most duplcated file table
content by skipping file table dump if the same file table address is
encountered consecutively while iterating over a process' threads.

This introduces a semantic change to the statedump fields, and will
therefore be introduced in lttng-modules 2.12 onwards, not backported as
a fix.

Fixes: #1245
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoRemove kernel version from syscall headers name
Michael Jeanson [Wed, 4 Mar 2020 20:45:18 +0000 (15:45 -0500)] 
Remove kernel version from syscall headers name

This will reduce the delta when updating the syscall headers and make it
easier to track the changes.

Add a comment in the headers to track the version they were generated
from.

Change-Id: I939e1ebd8ab18d880ff1394e771982b5491d3b19
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoAdd script to automate the syscall extraction process
Michael Jeanson [Wed, 4 Mar 2020 20:37:11 +0000 (15:37 -0500)] 
Add script to automate the syscall extraction process

Change-Id: Ic8e1e62058616f221569e1f5ed2ffb9d54275af5
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoUpdate lttng-syscalls-extractor for v5.6.0
Michael Jeanson [Wed, 4 Mar 2020 20:21:03 +0000 (15:21 -0500)] 
Update lttng-syscalls-extractor for v5.6.0

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I48e4d82484d6edcc7601a40d2f94e5c46ad5184b

4 years agoAdd a Makefile for the lttng-syscalls-extractor module
Michael Jeanson [Wed, 4 Mar 2020 20:06:41 +0000 (15:06 -0500)] 
Add a Makefile for the lttng-syscalls-extractor module

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I0c1cc650d8594ef01b4039bb3d2af94bfddea54d

4 years agoCleanup: Syscall headers scripts shellcheck warnings
Michael Jeanson [Fri, 14 Feb 2020 21:38:27 +0000 (16:38 -0500)] 
Cleanup: Syscall headers scripts shellcheck warnings

Change-Id: Ia616724d942f43f5fde2186bff28446120b17c43
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoRemove lttng-ftrace integration
Mathieu Desnoyers [Mon, 2 Mar 2020 16:26:39 +0000 (11:26 -0500)] 
Remove lttng-ftrace integration

The lttng-ftrace integration (LTTNG_KERNEL_FUNCTION instrumentation
type) was unused for a while now. The "function" probing is actually
done with kprobes and kretprobes (LTTNG_KERNEL_KPROBE and
LTTNG_KERNEL_KRETPROBE).

Remove it so a use of kallsyms_lookup_name() can be removed as well.
Note that in the future we could add back this support by using
register_ftrace_function() which is exported to kernel modules, but
considering that we have not been using this code for a while,
just remove the implementation for now.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoRemove dependency on kallsyms for splice_to_pipe (kernel 4.2+)
Mathieu Desnoyers [Mon, 2 Mar 2020 16:03:19 +0000 (11:03 -0500)] 
Remove dependency on kallsyms for splice_to_pipe (kernel 4.2+)

Upstream commit 2b514574f7e88 "net: af_unix: implement splice for stream
af_unix sockets" exported the "splice_to_pipe" symbol, so use it to
remove a dependency on kallsyms_lookup_name().

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoRemove dependency on kallsyms for irq_to_desc (kernel 3.4+)
Mathieu Desnoyers [Mon, 2 Mar 2020 15:52:54 +0000 (10:52 -0500)] 
Remove dependency on kallsyms for irq_to_desc (kernel 3.4+)

Upstream commit 3911ff30f5d "genirq: export handle_edge_irq() and
irq_to_desc()" exported the irq_to_desc symbol, so use it to remove a
dependency on kallsyms_lookup_name().

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoRemove work-around for signed tracepoint module tainting (kernel 3.15+)
Mathieu Desnoyers [Mon, 2 Mar 2020 15:40:26 +0000 (10:40 -0500)] 
Remove work-around for signed tracepoint module tainting (kernel 3.15+)

Upstream commit 66cc69e34e86a "Fix: module signature vs tracepoints: add
new TAINT_UNSIGNED_MODULE" fixed an issue where the kernel was
considering unsigned modules as tainting the kernel in the same way as a
force-loaded modules, which was causing the tracepoints within those
modules to be hidden.

This fix was merged in kernel 3.15, so there is no use in applying this
work-around starting from that kernel.

This removes a dependency on kallsyms_lookup_name() for the symbol
"tracepoint_module_notify".

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoChange integer base to hex for fields representing addresses
Francis Deslauriers [Fri, 21 Feb 2020 21:59:33 +0000 (16:59 -0500)] 
Change integer base to hex for fields representing addresses

It seems more expected to show addresses in hexadecimal.

I applied this to all fields named "ip" or "addr".

I haven't change the `addr` fields in the i2c.h file as I am not aware
of what is the convention when displaying I2C addresses.

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I6ed00b31e319e1d966ab138471248dcf3d2c7315

4 years agoFix: rcu: Fix data-race due to atomic_t copy-by-value (5.5.6, 5.4.22)
Mathieu Desnoyers [Tue, 25 Feb 2020 15:18:11 +0000 (10:18 -0500)] 
Fix: rcu: Fix data-race due to atomic_t copy-by-value (5.5.6, 5.4.22)

The following upstream commit has been backported to stable kernels
5.5.6 and 5.4.22:

  commit 6cf539a87a61a4fbc43f625267dbcbcf283872ed
  Author: Marco Elver <elver@google.com>
  Date:   Wed Oct 9 17:57:43 2019 +0200

    rcu: Fix data-race due to atomic_t copy-by-value

    This fixes a data-race where `atomic_t dynticks` is copied by value. The
    copy is performed non-atomically, resulting in a data-race if `dynticks`
    is updated concurrently.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: y2038: hide timeval/timespec/itimerval/itimerspec types (v5.6)
Michael Jeanson [Mon, 24 Feb 2020 20:11:03 +0000 (15:11 -0500)] 
fix: y2038: hide timeval/timespec/itimerval/itimerspec types (v5.6)

See upstream commit:

  commit c766d1472c70d25ad475cf56042af1652e792b23
  Author: Arnd Bergmann <arnd@arndb.de>
  Date:   Thu Feb 20 20:03:57 2020 -0800

    y2038: hide timeval/timespec/itimerval/itimerspec types

    There are no in-kernel users remaining, but there may still be users that
    include linux/time.h instead of sys/time.h from user space, so leave the
    types available to user space while hiding them from kernel space.

    Only the __kernel_old_* versions of these types remain now.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I986a813ad8b1c753ab1fa07f726b0cc481f049cb

4 years agofix: use timespec64 on kernels that have it
Michael Jeanson [Mon, 24 Feb 2020 19:50:20 +0000 (14:50 -0500)] 
fix: use timespec64 on kernels that have it

This fixes v5.6 which has hidden 'struct timespec' from kernel code and
makes 32bit archs y2038 compliant on v3.17 and newer.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I1daedc4a93cbbebb8f5a1d99c4619cb27f6a9e07

4 years agofix: move lttng_close_on_exec to proper wrapper
Michael Jeanson [Mon, 24 Feb 2020 20:15:36 +0000 (15:15 -0500)] 
fix: move lttng_close_on_exec to proper wrapper

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I330c6d63fc9c1af00c9a9471563d322d036b9a9c

4 years agofix: 'struct timex' removed upstream (v5.6)
Michael Jeanson [Mon, 24 Feb 2020 16:30:22 +0000 (11:30 -0500)] 
fix: 'struct timex' removed upstream (v5.6)

The 'timex' struct was remove in v5.6 and replaced by 2 variants, one
that is y2038 compliant and a compat version for 32bit archs.

Add this temporary fix while we update our syscalls tracepoint headers,
the type of this struct has limited importance since it's only used to
record the adress in the trace.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Change-Id: I085b22f282db57985f1c3d341e7c0866cb20e3c9

4 years agoFix: statedump: consistently check task_cred_xxx() return value for NULL
Mathieu Desnoyers [Thu, 20 Feb 2020 15:46:25 +0000 (10:46 -0500)] 
Fix: statedump: consistently check task_cred_xxx() return value for NULL

trace_lttng_statedump_process_user_ns() internally checks whether
user_ns is NULL. While this does not appear to be a possible return
value for task_cred_xxx(), err on the safe side and check for NULL here
as well to be consistent with the paranoid behavior of
trace_lttng_statedump_process_user_ns().

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoFix: statedump: check task_active_pid_ns return value for NULL
Mathieu Desnoyers [Thu, 20 Feb 2020 14:58:42 +0000 (09:58 -0500)] 
Fix: statedump: check task_active_pid_ns return value for NULL

The lttng-statedump checks the return value of task_active_pid_ns()
before each use within lttng_statedump_process_pid_ns(), but misses
the NULL check before dereferencing pid_ns->parent.

This race happens if a task exists in "dead" state while the statedump
iterates on that task.

Reported-by: Li Zhou <li.zhou@windriver.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoFix: lttng-events.c: variable may be used uninitialized
Francis Deslauriers [Tue, 4 Feb 2020 21:05:30 +0000 (16:05 -0500)] 
Fix: lttng-events.c: variable may be used uninitialized

Fixes the following warning:
/home/frdeso/projets/lttng/modules/lttng-events.c: In function ‘print_metadata_escaped_field’:
/home/frdeso/projets/lttng/modules/lttng-events.c:2563:5: error: ‘ret’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
     if (ret)
           ^

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: If2db5e1ad9782fb49d6d07026976d3d22f89f2ab

4 years agoCleanup: remove trailing white spaces across project
Francis Deslauriers [Mon, 17 Feb 2020 23:13:56 +0000 (18:13 -0500)] 
Cleanup: remove trailing white spaces across project

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I2de34e4913f9cefbd72a138f2d3dfb9d67ece311

4 years agosched: Make the sched_switch task state an enum
Geneviève Bastien [Wed, 12 Feb 2020 21:58:25 +0000 (16:58 -0500)] 
sched: Make the sched_switch task state an enum

This gives meaning to the task state value. Only the bit masks are
enumerated, as defined compositions are not exhaustive listing of all
possible values and there would be a lot of unknown. Interpretation of
combination of bit flags is left to the consumer of the event.

Change-Id: I83c5fbee9cba2701c7238c0ac6abd4c8a351b193
Signed-off-by: Geneviève Bastien <gbastien@versatic.net>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoblock: Make the rwbs field as a bit field enum
Geneviève Bastien [Tue, 11 Feb 2020 16:20:27 +0000 (11:20 -0500)] 
block: Make the rwbs field as a bit field enum

The rwbs value of block request events is in fact a series of bit fields
set to 0 or 1. The enumeration values are all powers of 2, trace readers
could interpret this as a bit field enum and show the result as a
concatenation of the corresponding flags. For example, with a matching
patch for babeltrace2's pretty sink, the output for a rwbs value of 12
would be:

[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 ) }

Consumers who have no notion of the bit field enum can still use the
integer value of the field as before.

Change-Id: I2e873f16e5a0507e52cfc85c70b471a9f4d5d5d1
Signed-off-by: Geneviève Bastien <gbastien@versatic.net>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: workqueue: add worker function to workqueue_execute_end tracepoint (v5.6)
Michael Jeanson [Tue, 11 Feb 2020 20:35:11 +0000 (15:35 -0500)] 
fix: workqueue: add worker function to workqueue_execute_end tracepoint (v5.6)

See upstream commit :

  commit 1c5da0ec7f20dfb56030fb93f7f52f48e12deb52
  Author: Daniel Jordan <daniel.m.jordan@oracle.com>
  Date:   Mon Jan 13 17:52:39 2020 -0500

    workqueue: add worker function to workqueue_execute_end tracepoint

    It's surprising that workqueue_execute_end includes only the work when
    its counterpart workqueue_execute_start has both the work and the worker
    function.

    You can't set a tracing filter or trigger based on the function, and
    postprocessing scripts interested in specific functions are harder to
    write since they have to remember the work from _start and match it up
    with the same field in _end.

    Add the function name, taking care to use the copy stashed in the
    worker since the work is no longer safe to touch.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: media: v4l2: abstract timeval handling in v4l2_buffer (v5.6)
Michael Jeanson [Tue, 11 Feb 2020 20:13:53 +0000 (15:13 -0500)] 
fix: media: v4l2: abstract timeval handling in v4l2_buffer (v5.6)

See upstream commit :

  commit 77cdffcb0bfb87fe3645894335cb8cb94917e6ac
  Author: Arnd Bergmann <arnd@arndb.de>
  Date:   Mon Dec 16 15:15:00 2019 +0100

    media: v4l2: abstract timeval handling in v4l2_buffer

    As a preparation for adding 64-bit time_t support in the uapi,
    change the drivers to no longer care about the format of the
    timestamp field in struct v4l2_buffer.

    The v4l2_timeval_to_ns() function is no longer needed in the
    kernel after this, but there is userspace code relying on
    it to be part of the uapi header.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: rcu: Remove kfree_rcu() special casing and lazy-callback (v5.6)
Michael Jeanson [Tue, 11 Feb 2020 20:08:04 +0000 (15:08 -0500)] 
fix: rcu: Remove kfree_rcu() special casing and lazy-callback (v5.6)

See upstream commit :

  commit 77a40f97030b27b3fc1640a3ed203870f0817f57
  Author: Joel Fernandes (Google) <joel@joelfernandes.org>
  Date:   Fri Aug 30 12:36:32 2019 -0400

    rcu: Remove kfree_rcu() special casing and lazy-callback handling

    This commit removes kfree_rcu() special-casing and the lazy-callback
    handling from Tree RCU.  It moves some of this special casing to Tiny RCU,
    the removal of which will be the subject of later commits.

    This results in a nice negative delta.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: rcu: Fix data-race due to atomic_t copy-by-value (v5.6)
Michael Jeanson [Tue, 11 Feb 2020 19:51:01 +0000 (14:51 -0500)] 
fix: rcu: Fix data-race due to atomic_t copy-by-value (v5.6)

See upstream commit :

  commit 6cf539a87a61a4fbc43f625267dbcbcf283872ed
  Author: Marco Elver <elver@google.com>
  Date:   Wed Oct 9 17:57:43 2019 +0200

    rcu: Fix data-race due to atomic_t copy-by-value

    This fixes a data-race where `atomic_t dynticks` is copied by value. The
    copy is performed non-atomically, resulting in a data-race if `dynticks`
    is updated concurrently.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: btrfs: make btrfs_ordered_extent naming consistent (v5.6)
Michael Jeanson [Tue, 11 Feb 2020 19:47:23 +0000 (14:47 -0500)] 
fix: btrfs: make btrfs_ordered_extent naming consistent (v5.6)

See upstream commit :

  commit bffe633e00fb6b904817137fc17a44b42efcd985
  Author: Omar Sandoval <osandov@fb.com>
  Date:   Mon Dec 2 17:34:19 2019 -0800

    btrfs: make btrfs_ordered_extent naming consistent with btrfs_file_extent_item

    ordered->start, ordered->len, and ordered->disk_len correspond to
    fi->disk_bytenr, fi->num_bytes, and fi->disk_num_bytes, respectively.
    It's confusing to translate between the two naming schemes. Since a
    btrfs_ordered_extent is basically a pending btrfs_file_extent_item,
    let's make the former use the naming from the latter.

    Note that I didn't touch the names in tracepoints just in case there are
    scripts depending on the current naming.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: KVM: x86: Use gpa_t for cr2/gpa to fix TDP support on 32-bit (v5.6)
Michael Jeanson [Tue, 11 Feb 2020 19:41:29 +0000 (14:41 -0500)] 
fix: KVM: x86: Use gpa_t for cr2/gpa to fix TDP support on 32-bit (v5.6)

See upstream commit :

  commit 736c291c9f36b07f8889c61764c28edce20e715d
  Author: Sean Christopherson <sean.j.christopherson@intel.com>
  Date:   Fri Dec 6 15:57:14 2019 -0800

    KVM: x86: Use gpa_t for cr2/gpa to fix TDP support on 32-bit KVM

    Convert a plethora of parameters and variables in the MMU and page fault
    flows from type gva_t to gpa_t to properly handle TDP on 32-bit KVM.

    Thanks to PSE and PAE paging, 32-bit kernels can access 64-bit physical
    addresses.  When TDP is enabled, the fault address is a guest physical
    address and thus can be a 64-bit value, even when both KVM and its guest
    are using 32-bit virtual addressing, e.g. VMX's VMCS.GUEST_PHYSICAL is a
    64-bit field, not a natural width field.

    Using a gva_t for the fault address means KVM will incorrectly drop the
    upper 32-bits of the GPA.  Ditto for gva_to_gpa() when it is used to
    translate L2 GPAs to L1 GPAs.

    Opportunistically rename variables and parameters to better reflect the
    dual address modes, e.g. use "cr2_or_gpa" for fault addresses and plain
    "addr" instead of "vaddr" when the address may be either a GVA or an L2
    GPA.  Similarly, use "gpa" in the nonpaging_page_fault() flows to avoid
    a confusing "gpa_t gva" declaration; this also sets the stage for a
    future patch to combing nonpaging_page_fault() and tdp_page_fault() with
    minimal churn.

    Sprinkle in a few comments to document flows where an address is known
    to be a GVA and thus can be safely truncated to a 32-bit value.  Add
    WARNs in kvm_handle_page_fault() and FNAME(gva_to_gpa_nested)() to help
    document such cases and detect bugs.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: proc: decouple proc from VFS with "struct proc_ops" (v5.6)
Michael Jeanson [Tue, 11 Feb 2020 16:20:41 +0000 (11:20 -0500)] 
fix: proc: decouple proc from VFS with "struct proc_ops" (v5.6)

See upstream commit :

  commit d56c0d45f0e27f814e87a1676b6bdccccbc252e9
  Author: Alexey Dobriyan <adobriyan@gmail.com>
  Date:   Mon Feb 3 17:37:14 2020 -0800

    proc: decouple proc from VFS with "struct proc_ops"

    Currently core /proc code uses "struct file_operations" for custom hooks,
    however, VFS doesn't directly call them.  Every time VFS expands
    file_operations hook set, /proc code bloats for no reason.

    Introduce "struct proc_ops" which contains only those hooks which /proc
    allows to call into (open, release, read, write, ioctl, mmap, poll).  It
    doesn't contain module pointer as well.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoVersion 2.12.0-rc1 v2.12.0-rc1
Mathieu Desnoyers [Wed, 5 Feb 2020 15:36:41 +0000 (10:36 -0500)] 
Version 2.12.0-rc1

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoFix: lttng-syscalls.c: marking wrong syscall probe as unregistered
Francis Deslauriers [Tue, 28 Jan 2020 22:19:20 +0000 (17:19 -0500)] 
Fix: lttng-syscalls.c: marking wrong syscall probe as unregistered

When calling `lttng_syscalls_unregister()` we currently mark as
unregistered the wrong syscall probe type.

Concretely, when we unregister "sys_exit" we wrongfully mark that
sys_enter is unregistered and vice versa.

Given than currently entry and exit probes are always enabled together
(except on internal errors), the effect of this bug is not user-visible.

Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I64acf6a941a3d1fa1bf8be424f834ddb7fb92ace

4 years agoVersion 2.12.0-pre v2.12.0-pre
Mathieu Desnoyers [Tue, 7 Jan 2020 18:09:38 +0000 (13:09 -0500)] 
Version 2.12.0-pre

Integrate the missing 2.11.0-rc1 entries to the changelog. This
was caused by forking to stable-2.11 the commit before the v2.11.0-rc1
tag. In the future, we will try to keep the rc1 tag as the common
ancestor between master and stable branches.

This also updates the lttng-modules version to 2.12.0-pre, which is
a state that will stay until the v2.12.0-rc1 tag.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoBump LTTNG_MODULES_ABI_MINOR_VERSION to 5
Jérémie Galarneau [Fri, 20 Dec 2019 03:11:51 +0000 (22:11 -0500)] 
Bump LTTNG_MODULES_ABI_MINOR_VERSION to 5

New operations were added to the lttng-modules ABI as part of the 2.12
release cycle to support UID tracking and the "session clear"
functionality.

This will allow future LTTng-tools versions to check for those
capabilities.

Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: use user ns wrapper code in new id trackers
Michael Jeanson [Tue, 17 Dec 2019 17:11:11 +0000 (12:11 -0500)] 
fix: use user ns wrapper code in new id trackers

These wrappers are required to translate kuid on kernels prior to v3.5.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: function prototype in wrapper/mm.h
Michael Jeanson [Tue, 17 Dec 2019 17:11:10 +0000 (12:11 -0500)] 
fix: function prototype in wrapper/mm.h

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoID tracker: implement vpid/uid/vuid/gid/vgid trackers
Mathieu Desnoyers [Tue, 17 Dec 2019 15:17:01 +0000 (10:17 -0500)] 
ID tracker: implement vpid/uid/vuid/gid/vgid trackers

Co-developed-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agolttng-abi: Document ioctl numbers reserved by lttng-abi-old.h
Mathieu Desnoyers [Tue, 17 Dec 2019 14:27:54 +0000 (09:27 -0500)] 
lttng-abi: Document ioctl numbers reserved by lttng-abi-old.h

Document the ioctl numbers reserved by lttng-abi-old.h in the
(relatively) new lttng-abi.h, so they are taken into account when
assining numbers to new lttng ioctl commands.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agolttng-clear: stop tracing required
Mathieu Desnoyers [Wed, 23 Oct 2019 16:56:59 +0000 (12:56 -0400)] 
lttng-clear: stop tracing required

Require that tracing is stopped when buffers are cleared. Update
comments and warning checks to that effect.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agosunrpc: introduce lttng_get_clid helper
Mathieu Desnoyers [Thu, 12 Dec 2019 15:39:38 +0000 (10:39 -0500)] 
sunrpc: introduce lttng_get_clid helper

Introduce the lttng_get_clid helper to always check for NULL pointer
when getting the client id. While not always strictly needed depending
on the tracepoint callsite, prefer robustness of instrumentation and
always check for NULL rather than play whack-a-mole.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoFix: sunrpc: use signed integer for client id
Mathieu Desnoyers [Thu, 12 Dec 2019 15:29:37 +0000 (10:29 -0500)] 
Fix: sunrpc: use signed integer for client id

Within include/linux/sunrpc/clnt.h:struct rpc_cltn, the cl_clid field
is an unsigned integer, which is the type expected by the tracepoint
signature.

However, looking into net/sunrpc/clnt.c:rpc_alloc_clid(), its allocation
considers negative signed integer as errors.

Therefore, in order to properly show "-1" in the trace output (rather
than MAX_INT) when called with a NULL task->tk_client, move to a
signed integer as backing type for the client_id field.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoFix: sunrpc: null rpc_clnt dereference in rpc_task_queued tracepoint
Mathieu Desnoyers [Thu, 12 Dec 2019 15:29:02 +0000 (10:29 -0500)] 
Fix: sunrpc: null rpc_clnt dereference in rpc_task_queued tracepoint

Based on upstream Linux commit:

commit 0be283f676a1e7b208db0c992283197ef8b52158
Author: Benjamin Coddington <bcodding@redhat.com>
Date:   Tue Jan 23 09:32:35 2018 -0500

    SUNRPC: Fix null rpc_clnt dereference in rpc_task_queued tracepoint

    Backchannel tasks will not have a reference to the rpc_clnt.  Return -1 for
    cl_clid in that case.

Signed-off-by: Benjamin Coddington <bcodding@redhat.com>
Signed-off-by: Trond Myklebust <trondmy@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoFix: SUNRPC: Fix oops when trace sunrpc_task events in nfs client
Quanyang Wang [Thu, 5 Dec 2019 07:35:32 +0000 (15:35 +0800)] 
Fix: SUNRPC: Fix oops when trace sunrpc_task events in nfs client

See upstream commit :

commit 2ca310fc4160ed0420da65534a21ae77b24326a8
Author: Ditang Chen <chendt.fnst@cn.fujitsu.com>
Date: Fri, 7 Mar 2014 13:27:57 +0800
Subject: SUNRPC: Fix oops when trace sunrpc_task events in nfs client

When tracking sunrpc_task events in nfs client, the clnt pointer may be NULL.

[  139.269266] BUG: unable to handle kernel NULL pointer dereference at 0000000000000004
[  139.269915] IP: [<ffffffffa026f216>] ftrace_raw_event_rpc_task_running+0x86/0xf0 [sunrpc]
[  139.269915] PGD 1d293067 PUD 1d294067 PMD 0
[  139.269915] Oops: 0000 [#1] SMP
[  139.269915] Modules linked in: nfsv4 dns_resolver nfs lockd sunrpc fscache sg ppdev e1000
serio_raw pcspkr parport_pc parport i2c_piix4 i2c_core microcode xfs libcrc32c sd_mod sr_mod
cdrom ata_generic crc_t10dif crct10dif_common pata_acpi ahci libahci ata_piix libata dm_mirror
dm_region_hash dm_log dm_mod
[  139.269915] CPU: 0 PID: 59 Comm: kworker/0:2 Not tainted 3.10.0-84.el7.x86_64 #1
[  139.269915] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[  139.269915] Workqueue: rpciod rpc_async_schedule [sunrpc]
[  139.269915] task: ffff88001b598000 ti: ffff88001b632000 task.ti: ffff88001b632000
[  139.269915] RIP: 0010:[<ffffffffa026f216>]  [<ffffffffa026f216>] ftrace_raw_event_rpc_task_running+0x86/0xf0 [sunrpc]
[  139.269915] RSP: 0018:ffff88001b633d70  EFLAGS: 00010206
[  139.269915] RAX: ffff88001dfc5338 RBX: ffff88001cc37a00 RCX: ffff88001dfc5334
[  139.269915] RDX: ffff88001dfc5338 RSI: 0000000000000000 RDI: ffff88001dfc533c
[  139.269915] RBP: ffff88001b633db0 R08: 000000000000002c R09: 000000000000000a
[  139.269915] R10: 0000000000062180 R11: 00000020759fb9dc R12: ffffffffa0292c20
[  139.269915] R13: ffff88001dfc5334 R14: 0000000000000000 R15: 0000000000000000
[  139.269915] FS:  0000000000000000(0000) GS:ffff88001fc00000(0000) knlGS:0000000000000000
[  139.269915] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[  139.269915] CR2: 0000000000000004 CR3: 000000001d290000 CR4: 00000000000006f0
[  139.269915] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  139.269915] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[  139.269915] Stack:
[  139.269915]  000000001b633d98 0000000000000246 ffff88001df1dc00 ffff88001cc37a00
[  139.269915]  ffff88001bc35e60 0000000000000000 ffff88001ffa0a48 ffff88001bc35ee0
[  139.269915]  ffff88001b633e08 ffffffffa02704b5 0000000000010000 ffff88001cc37a70
[  139.269915] Call Trace:
[  139.269915]  [<ffffffffa02704b5>] __rpc_execute+0x1d5/0x400 [sunrpc]
[  139.269915]  [<ffffffffa0270706>] rpc_async_schedule+0x26/0x30 [sunrpc]
[  139.269915]  [<ffffffff8107867b>] process_one_work+0x17b/0x460
[  139.269915]  [<ffffffff8107942b>] worker_thread+0x11b/0x400
[  139.269915]  [<ffffffff81079310>] ? rescuer_thread+0x3e0/0x3e0
[  139.269915]  [<ffffffff8107fc80>] kthread+0xc0/0xd0
[  139.269915]  [<ffffffff8107fbc0>] ? kthread_create_on_node+0x110/0x110
[  139.269915]  [<ffffffff815d122c>] ret_from_fork+0x7c/0xb0
[  139.269915]  [<ffffffff8107fbc0>] ? kthread_create_on_node+0x110/0x110
[  139.269915] Code: 4c 8b 45 c8 48 8d 7d d0 89 4d c4 41 89 c9 b9 28 00 00 00 e8 9d b4 e9
e0 48 85 c0 49 89 c5 74 a2 48 89 c7 e8 9d 3f e9 e0 48 89 c2 <41> 8b 46 04 48 8b 7d d0 4c
89 e9 4c 89 e6 89 42 0c 0f b7 83 d4
[  139.269915] RIP  [<ffffffffa026f216>] ftrace_raw_event_rpc_task_running+0x86/0xf0 [sunrpc]
[  139.269915]  RSP <ffff88001b633d70>
[  139.269915] CR2: 0000000000000004
[  140.946406] ---[ end trace ba486328b98d7622 ]---

Signed-off-by: Quanyang Wang <quanyang.wang@windriver.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: ext4: Reserve revoke credits for freed blocks (v5.5)
Michael Jeanson [Tue, 10 Dec 2019 16:41:14 +0000 (11:41 -0500)] 
fix: ext4: Reserve revoke credits for freed blocks (v5.5)

See upstream commit:

  commit 83448bdfb59731c2f54784ed3f4a93ff95be6e7e
  Author: Jan Kara <jack@suse.cz>
  Date:   Tue Nov 5 17:44:29 2019 +0100

    ext4: Reserve revoke credits for freed blocks

    So far we have reserved only relatively high fixed amount of revoke
    credits for each transaction. We over-reserved by large amount for most
    cases but when freeing large directories or files with data journalling,
    the fixed amount is not enough. In fact the worst case estimate is
    inconveniently large (maximum extent size) for freeing of one extent.

    We fix this by doing proper estimate of the amount of blocks that need
    to be revoked when removing blocks from the inode due to truncate or
    hole punching and otherwise reserve just a small amount of revoke
    credits for each transaction to accommodate freeing of xattrs block or
    so.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: btrfs: tracepoints: constify all pointers (v5.5)
Michael Jeanson [Tue, 10 Dec 2019 16:41:12 +0000 (11:41 -0500)] 
fix: btrfs: tracepoints: constify all pointers (v5.5)

See upstream commit:

  commit 1d2e7c7c3ed73cc510a4dc093df2a935092ff5ad
  Author: David Sterba <dsterba@suse.com>
  Date:   Thu Oct 17 13:28:57 2019 +0200

    btrfs: tracepoints: constify all pointers

    We don't modify the data passed to tracepoints, some of the declarations
    are already const, add it to the rest.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: btrfs block group struct refactor (v5.5)
Michael Jeanson [Tue, 10 Dec 2019 16:41:13 +0000 (11:41 -0500)] 
fix: btrfs block group struct refactor (v5.5)

See upstream commits:

  commit 32da5386d9a4fd5c1155cecf703df104d918954c
  Author: David Sterba <dsterba@suse.com>
  Date:   Tue Oct 29 19:20:18 2019 +0100

    btrfs: rename btrfs_block_group_cache

    The type name is misleading, a single entry is named 'cache' while this
    normally means a collection of objects. Rename that everywhere. Also the
    identifier was quite long, making function prototypes harder to format.

  commit b3470b5dbe1300dea94191ae4b7d070be9a5cdc9
  Author: David Sterba <dsterba@suse.com>
  Date:   Wed Oct 23 18:48:22 2019 +0200

    btrfs: add dedicated members for start and length of a block group

    The on-disk format of block group item makes use of the key that stores
    the offset and length. This is further used in the code, although this
    makes thing harder to understand. The key is also packed so the
    offset/length is not properly aligned as u64.

    Add start (key.objectid) and length (key.offset) members to block group
    and remove the embedded key.  When the item is searched or written, a
    local variable for key is used.

  commit bf38be65f3703d5ef3661c0a2802bc28e76b8f19
  Author: David Sterba <dsterba@suse.com>
  Date:   Wed Oct 23 18:48:11 2019 +0200

    btrfs: move block_group_item::used to block group

    For unknown reasons, the member 'used' in the block group struct is
    stored in the b-tree item and accessed everywhere using the special
    accessor helper. Let's unify it and make it a regular member and only
    update the item before writing it to the tree.

    The item is still being used for flags and chunk_objectid, there's some
    duplication until the item is removed in following patches.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agofix: y2038: itimer: change implementation to timespec64 (v5.5)
Michael Jeanson [Tue, 10 Dec 2019 16:41:11 +0000 (11:41 -0500)] 
fix: y2038: itimer: change implementation to timespec64 (v5.5)

See upstream commit:

  commit bd40a175769d411b2a37e1c087082ac7ee2c15bb
  Author: Arnd Bergmann <arnd@arndb.de>
  Date:   Thu Nov 7 15:27:39 2019 +0100

    y2038: itimer: change implementation to timespec64

    There is no 64-bit version of getitimer/setitimer since that is not
    actually needed. However, the implementation is built around the
    deprecated 'struct timeval' type.

    Change the code to use timespec64 internally to reduce the dependencies
    on timeval and associated helper functions.

    Minor adjustments in the code are needed to make the native and compat
    version work the same way, and to keep the range check working after
    the conversion.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 years agoUpdate .gitignore from upstream
Michael Jeanson [Tue, 10 Dec 2019 16:41:10 +0000 (11:41 -0500)] 
Update .gitignore from upstream

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoAdd missing include for kernels between 3.8 and 3.15
Michael Jeanson [Mon, 28 Oct 2019 20:28:43 +0000 (16:28 -0400)] 
Add missing include for kernels between 3.8 and 3.15

This is required at least for:
 v3.13.x
 v3.11.x
 v3.9.x
 v3.8.x

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: LTTNG_KERNEL_ADD_CALLSITE: Handle unknown event type
Mathieu Desnoyers [Fri, 18 Oct 2019 19:23:59 +0000 (15:23 -0400)] 
Fix: LTTNG_KERNEL_ADD_CALLSITE: Handle unknown event type

Return -ENOSYS for unknown event type (similarly to other commands)
rather than falling-through the switch statement.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agonet: Add entry/exit tracepoints for all receive paths
Geneviève Bastien [Wed, 19 Jun 2019 19:03:49 +0000 (15:03 -0400)] 
net: Add entry/exit tracepoints for all receive paths

In the linux kernel, there are entry and exit tracepoints on
the various paths of network reception.

Those tracepoints are useful to bound the network reception such that
all events happening between the entry and exit can be linked to this
network reception (for example, wakeups). They can also be used to
perform network stack latency analysis.

Acked-by: Julien Desfossez <jdesfossez@digitalocean.com>
Signed-off-by: Geneviève Bastien <gbastien@versatic.net>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agostatedump: Add thread ID (tid) to interrupt
Geneviève Bastien [Mon, 1 Oct 2018 17:26:31 +0000 (13:26 -0400)] 
statedump: Add thread ID (tid) to interrupt

Threaded IRQs have a 'thread' field set in the the action structure,
defining which process to wakeup when the IRQ happens.

Having this information will allow to know which process are IRQ
handling process and the analyses can track what happens in those
processes to the IRQ that caused them to wake up.

[ Edit by Mathieu Desnoyers: Rename "thread" field to "tid" to stay
  consistent with the rest of lttng-modules tracepoints. ]

Signed-off-by: Geneviève Bastien <gbastien+lttng@versatic.net>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agometadata: Add the product uuid to the 'env' section
Geneviève Bastien [Thu, 1 Feb 2018 17:18:26 +0000 (12:18 -0500)] 
metadata: Add the product uuid to the 'env' section

The product UUID is taken from the DMI system information and can be
used to identify a machine's hardware (virtual or physical).

Signed-off-by: Geneviève Bastien <gbastien+lttng@versatic.net>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoCleanup: statedump process state event pid namespace fields
Michael Jeanson [Tue, 19 Feb 2019 21:14:38 +0000 (16:14 -0500)] 
Cleanup: statedump process state event pid namespace fields

Now that we have namespace specific events in the statedump,
remove the duplicated information in the process state event
and make it a single event instead of recursing on the pid
namespace hierarchy.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoAdd namespaces statedump
Michael Jeanson [Tue, 12 Feb 2019 14:47:34 +0000 (09:47 -0500)] 
Add namespaces statedump

Add a statedump event for each type of namespace.

The pid ns was already implemented as part of the lttng_statedump_process_state
event, move the "vtid" and "vpid" fields to the new
lttng_statedump_process_pid_ns event.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoAdd uid/gid contexts
Michael Jeanson [Tue, 12 Feb 2019 14:47:00 +0000 (09:47 -0500)] 
Add uid/gid contexts

Add a context for each available kernel user and group IDs

  * uid : real user ID
  * euid : effective user ID
  * suid : saved set-user ID

These are the IDs as seen in the initial user namespace, see
credentials(7) for details on each type.

Also add a "virtual" version of each type with the "v" prefix that
are the IDs as seen in the current user namespace.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoAdd namespace contexts
Michael Jeanson [Tue, 12 Feb 2019 14:43:24 +0000 (09:43 -0500)] 
Add namespace contexts

Add a context for each available kernel namespace which currently are :
cgroup, ipc, mnt, net, pid, user and uts. The id chosen to identify the
namespaces is the inode number of the file representing each one of them
in the proc filesystem. This was instroduced in v3.8.0 in this commit :

  commit 98f842e675f96ffac96e6c50315790912b2812be
  Author: Eric W. Biederman <ebiederm@xmission.com>
  Date:   Wed Jun 15 10:21:48 2011 -0700

    proc: Usable inode numbers for the namespace file descriptors.

    Assign a unique proc inode to each namespace, and use that
    inode number to ensure we only allocate at most one proc
    inode for every namespace in proc.

    A single proc inode per namespace allows userspace to test
    to see if two processes are in the same namespace.

    ...

Prior to this there is no unique identifier for a namespace that is
available to both the kernel and userspace. Enabling any of these
contexts on a kernel that is too old or doesn't have the proper features
enabled will fail and return -ENOSYS.

Per namespace particularities :

  - Cgroup
    - Introduced in 4.6.0
    - CONFIG_CGROUPS

  - IPC
    - Introduced in 2.6.25
    - CONFIG_IPC_NS

  - MNT
    - Introduced in 2.6.20
    - The mnt_namespace struct is defined in a private header

  - NET
    - Introduced in 2.6.24
    - CONFIG_NET_NS

  - PID
    - Introduced in 2.6.20
    - CONFIG_PID_NS

  - User
    - Introduced in 2.6.23
    - CONFIG_USER_NS

  - UTS
    - Introduced in 2.6.19
    - CONFIG_UTS_NS

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoREADME.md: Document LTTNG_TRACEPOINT_EVENT
Mathieu Desnoyers [Tue, 15 Oct 2019 19:58:16 +0000 (15:58 -0400)] 
README.md: Document LTTNG_TRACEPOINT_EVENT

Suggested-by: David Boles <blioboles@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoREADME.md: cleanup formatting for bullet lists
Mathieu Desnoyers [Tue, 15 Oct 2019 19:56:23 +0000 (15:56 -0400)] 
README.md: cleanup formatting for bullet lists

Suggested-by: David Boles <biblioboles@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: btrfs: move basic block_group definitions to their own header (v5.4)
Michael Jeanson [Mon, 30 Sep 2019 19:15:34 +0000 (15:15 -0400)] 
Fix: btrfs: move basic block_group definitions to their own header (v5.4)

  commit aac0023c2106952538414254960c51dcf0dc39e9
  Author: Josef Bacik <josef@toxicpanda.com>
  Date:   Thu Jun 20 15:37:44 2019 -0400

    btrfs: move basic block_group definitions to their own header

    This is prep work for moving all of the block group cache code into its
    own file.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoCleanup: Silence gcc fall-through warning
Mathieu Desnoyers [Tue, 17 Sep 2019 18:12:07 +0000 (14:12 -0400)] 
Cleanup: Silence gcc fall-through warning

Use a comment pattern recognized by gcc 7.4.0 to silence the
fall-through warning.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: update sched prev_state instrumentation for upstream kernel
Gabriel-Andrew Pollo-Guilbert [Mon, 16 Sep 2019 17:57:38 +0000 (13:57 -0400)] 
Fix: update sched prev_state instrumentation for upstream kernel

Introduced in upstream Linux kernel 4.14.

    commit efb40f588b4370ffaeffafbd50f6ff213d954254
    Author: Peter Zijlstra <peterz@infradead.org>
    Date:   Fri Sep 22 18:19:53 2017 +0200

        sched/tracing: Fix trace_sched_switch task-state printing

Introduced in upstream Linux kernel 4.15.
Backported in 13f12749af15 (4.14.64).

    commit 3f5fe9fef5b2da06b6319fab8123056da5217c3f
    Author: Thomas Gleixner <tglx@linutronix.de>
    Date:   Wed Nov 22 13:05:48 2017 +0100

        sched/debug: Fix task state recording/printout

Introduced in upstream Linux kernel 4.20.
Backported in e1e5fa73e466 (4.14.102) and fd8152818f11 (4.19.9).

    commit 3054426dc68e5d63aa6a6e9b91ac4ec78e3f3805
    Author: Pavankumar Kondeti <pkondeti@codeaurora.org>
    Date:   Tue Oct 30 12:24:33 2018 +0530

        sched, trace: Fix prev_state output in sched_switch tracepoint

Signed-off-by: Gabriel-Andrew Pollo-Guilbert <gabriel.pollo-guilbert@efficios.com>
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: gcc-9.1 stack frame size warning
Mathieu Desnoyers [Fri, 6 Sep 2019 01:01:47 +0000 (21:01 -0400)] 
Fix: gcc-9.1 stack frame size warning

gcc-9.1.0 warns about lttng_session_ioctl taking a too large frame size.

lttng-modules/lttng-abi.c:622:1: warning: the frame size of 2240 bytes
is larger than 2048 bytes [-Wframe-larger-than=]

Combine the variables used in the various case of the switch so they are
not duplicated on the stack by the compiler.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoImplement ring buffer clear
Mathieu Desnoyers [Wed, 30 May 2018 00:11:43 +0000 (02:11 +0200)] 
Implement ring buffer clear

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoMake bitfield.h C++-friendly
Simon Marchi [Tue, 20 Aug 2019 01:51:52 +0000 (21:51 -0400)] 
Make bitfield.h C++-friendly

This patch changes bitfield.h to be usable in C++11.

It will probably never be compiled as C++ in the context of
lttng-modules, but this is just to keep things sync'ed across projects.

Signed-off-by: Simon Marchi <simon.marchi@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoIntroduce LTTNG_KERNEL_SESSION_SET_CREATION_TIME
Jonathan Rajotte [Tue, 13 Aug 2019 18:14:25 +0000 (14:14 -0400)] 
Introduce LTTNG_KERNEL_SESSION_SET_CREATION_TIME

Add trace_creation_datetime to the metadata env field

This allows a viewer to get more information regarding the creation time
of a trace. This information is normally inferred from the trace
hierarchy.

The ABI expects an ISO8601 compliant string value. We leave the
formatting to the trace controller.

Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoAdd metadata env fields
Jonathan Rajotte [Tue, 13 Aug 2019 18:14:24 +0000 (14:14 -0400)] 
Add metadata env fields

Add the following fields:

  - tracer_buffering_scheme
      The buffering scheme used by the tracer. lttng-modules sole
      buffering scheme is "global".
  - trace_name
      The name of the trace. Use the session name.

Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoIntroduce LTTNG_KERNEL_SESSION_SET_NAME
Jonathan Rajotte [Tue, 13 Aug 2019 18:14:23 +0000 (14:14 -0400)] 
Introduce LTTNG_KERNEL_SESSION_SET_NAME

The tracer controller (lttng-sessiond) can now inform the kernel tracer
of the name of the created session. This will allow the tracer to
propagate the information inside the trace metadata under a "env" field.

This information is normally inferred from the generated folder
structure where a trace rests.

Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: do not use diagnostic pragma when GCC version is lower than 4.6.0
Jonathan Rajotte [Thu, 4 Jul 2019 20:02:13 +0000 (16:02 -0400)] 
Fix: do not use diagnostic pragma when GCC version is lower than 4.6.0

Officially the diagnostic pragma are supported starting in 4.6. [1]
But they were present before 4.6 with limitation which we cannot
honour easily.

[1] https://gcc.gnu.org/gcc-4.6/changes.html

Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: missing define when not building with gcc
Jonathan Rajotte [Thu, 4 Jul 2019 20:02:12 +0000 (16:02 -0400)] 
Fix: missing define when not building with gcc

Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: lttng-tracepoint module notifier should return NOTIFY_OK
Mathieu Desnoyers [Mon, 24 Jun 2019 13:43:45 +0000 (09:43 -0400)] 
Fix: lttng-tracepoint module notifier should return NOTIFY_OK

Module notifiers should return NOTIFY_OK on success rather than the
value 0. The return value 0 does not seem to have any ill side-effects
in the notifier chain caller, but it is preferable to respect the API
requirements in case this changes in the future.

Notifiers can encapsulate a negative errno value with
notifier_from_errno(), but this is not needed by the LTTng tracepoint
notifier.

The approach taken in this notifier is to just print a console warning
on error, because tracing failure should not prevent loading a module.
So we definitely do not want to stop notifier iteration. Returning
an error without stopping iteration is not really that useful, because
only the return value of the last callback is returned to notifier chain
caller.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: Don't print ring-buffer's records count when it is not used
Jérémie Galarneau [Tue, 11 Jun 2019 22:34:32 +0000 (18:34 -0400)] 
Fix: Don't print ring-buffer's records count when it is not used

The teardown of a ring buffer causes a number of diagnostic messages
to be printed using printk. One of those contains the "records
count", which is only updated when lttng-modules is built with
LTTNG_RING_BUFFER_COUNT_EVENTS defined.

Move the "records count" printing to a different function and stub it
out when LTTNG_RING_BUFFER_COUNT_EVENTS is not defined
(default configuration).

This eliminates messages of the following form from the dmesg output
when an LTTng session is torn down.

[...] ring buffer relay-discard, cpu 0: 0 records written, 0 records overrun

Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: do not set quiescent state on channel destroy
Mathieu Desnoyers [Tue, 11 Jun 2019 21:44:56 +0000 (23:44 +0200)] 
Fix: do not set quiescent state on channel destroy

Setting the quiescent state to true for each stream at channel
destruction is not useful: there are no readers left anyway at
that stage.

The side-effect perceived of setting this quiescent state on
destroy is that the metadata stream ends up with an empty last
packet (due to flush_empty performed when setting the quiescent state)
which is never consumed. This shows up in the lttng-modules error
reporting.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: ring_buffer_frontend.c: init read timer with uninitialized flags
Mathieu Desnoyers [Tue, 4 Jun 2019 18:59:26 +0000 (14:59 -0400)] 
Fix: ring_buffer_frontend.c: init read timer with uninitialized flags

For the config->alloc RING_BUFFER_ALLOC_GLOBAL (metadata channel), the
read timer flags argument is uninitialized.

Found by Coverity:
CID 1401114 (#1 of 1): Uninitialized scalar variable (UNINIT)
6. uninit_use_in_call: Using uninitialized value flags when calling init_timer_key.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoIntroduce callstack stackwalk implementation header
Mathieu Desnoyers [Wed, 22 May 2019 01:16:15 +0000 (21:16 -0400)] 
Introduce callstack stackwalk implementation header

Introduce a new implementation header for the stackwalk-based API, added
in Linux 5.2 and gradually integrated within each architecture.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoPrepare callstack common code for stackwalk
Mathieu Desnoyers [Wed, 22 May 2019 01:15:44 +0000 (21:15 -0400)] 
Prepare callstack common code for stackwalk

Prepare the callstack common code for stackwalk implementation,
moving more legacy code to the legacy implementation header.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoIntroduce callstack legacy implementation header
Mathieu Desnoyers [Tue, 21 May 2019 21:51:23 +0000 (17:51 -0400)] 
Introduce callstack legacy implementation header

Split the callstack code: keep boilerplate code within the
C implementation file, and move the parts which depend on the
"legacy" (pre-stackwalk) stacktrace kernel API to a separate
implementation header.

This is a preparation step to introduce a new implementation
header for the stackwalk API, added in Linux 5.2 and gradually
integrated within each architecture.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agofix: random: only read from /dev/random after its pool has received 128 bits (v5.2)
Michael Jeanson [Tue, 21 May 2019 20:33:14 +0000 (16:33 -0400)] 
fix: random: only read from /dev/random after its pool has received 128 bits (v5.2)

See upstream commit:

  commit eb9d1bf079bb438d1a066d72337092935fc770f6
  Author: Theodore Ts'o <tytso@mit.edu>
  Date:   Wed Feb 20 16:06:38 2019 -0500

    random: only read from /dev/random after its pool has received 128 bits

    Immediately after boot, we allow reads from /dev/random before its
    entropy pool has been fully initialized.  Fix this so that we don't
    allow this until the blocking pool has received 128 bits.

    We do this by repurposing the initialized flag in the entropy pool
    struct, and use the initialized flag in the blocking pool to indicate
    whether it is safe to pull from the blocking pool.

    To do this, we needed to rework when we decide to push entropy from the
    input pool to the blocking pool, since the initialized flag for the
    input pool was used for this purpose.  To simplify things, we no
    longer use the initialized flag for that purpose, nor do we use the
    entropy_total field any more.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agofix: mm: move recent_rotated pages calculation to shrink_inactive_list() (v5.2)
Michael Jeanson [Tue, 21 May 2019 20:33:13 +0000 (16:33 -0400)] 
fix: mm: move recent_rotated pages calculation to shrink_inactive_list() (v5.2)

See upstream commit:

  commit 886cf1901db962cee5f8b82b9b260079a5e8a4eb
  Author: Kirill Tkhai <ktkhai@virtuozzo.com>
  Date:   Mon May 13 17:16:51 2019 -0700

    mm: move recent_rotated pages calculation to shrink_inactive_list()

    Patch series "mm: Generalize putback functions"]

    putback_inactive_pages() and move_active_pages_to_lru() are almost
    similar, so this patchset merges them ina single function.

    This patch (of 4):

    The patch moves the calculation from putback_inactive_pages() to
    shrink_inactive_list().  This makes putback_inactive_pages() looking more
    similar to move_active_pages_to_lru().

    To do that, we account activated pages in reclaim_stat::nr_activate.
    Since a page may change its LRU type from anon to file cache inside
    shrink_page_list() (see ClearPageSwapBacked()), we have to account pages
    for the both types.  So, nr_activate becomes an array.

    Previously we used nr_activate to account PGACTIVATE events, but now we
    account them into pgactivate variable (since they are about number of
    pages in general, not about sum of hpage_nr_pages).

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agofix: mm/vmscan: simplify trace_reclaim_flags and trace_shrink_flags (v5.2)
Michael Jeanson [Tue, 21 May 2019 20:33:12 +0000 (16:33 -0400)] 
fix: mm/vmscan: simplify trace_reclaim_flags and trace_shrink_flags (v5.2)

See upstream commit:

  commit 60b62ff7cc4217ac3de76535fa4c1510a798dbcb
  Author: Yafang Shao <laoar.shao@gmail.com>
  Date:   Mon May 13 17:23:08 2019 -0700

    mm/vmscan: simplify trace_reclaim_flags and trace_shrink_flags

    trace_reclaim_flags and trace_shrink_flags are almost the same.
    We can simplify them to avoid redundant code.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agofix: mm/vmscan: drop may_writepage and classzone_idx from direct reclaim begin templa...
Michael Jeanson [Tue, 21 May 2019 20:33:11 +0000 (16:33 -0400)] 
fix: mm/vmscan: drop may_writepage and classzone_idx from direct reclaim begin template (v5.2)

See upstream commit:

  commit 3481c37ffa1de58ef140d0fe9eabf56305e74666
  Author: Yafang Shao <laoar.shao@gmail.com>
  Date:   Mon May 13 17:19:14 2019 -0700

    mm/vmscan: drop may_writepage and classzone_idx from direct reclaim begin template

    There are three tracepoints using this template, which are
    mm_vmscan_direct_reclaim_begin,
    mm_vmscan_memcg_reclaim_begin,
    mm_vmscan_memcg_softlimit_reclaim_begin.

    Regarding mm_vmscan_direct_reclaim_begin,
    sc.may_writepage is !laptop_mode, that's a static setting, and
    reclaim_idx is derived from gfp_mask which is already show in this
    tracepoint.

    Regarding mm_vmscan_memcg_reclaim_begin,
    may_writepage is !laptop_mode too, and reclaim_idx is (MAX_NR_ZONES-1),
    which are both static value.

    mm_vmscan_memcg_softlimit_reclaim_begin is the same with
    mm_vmscan_memcg_reclaim_begin.

    So we can drop them all.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agofix: timer/trace: Improve timer tracing (v5.2)
Michael Jeanson [Tue, 21 May 2019 20:33:10 +0000 (16:33 -0400)] 
fix: timer/trace: Improve timer tracing (v5.2)

See upstream commit:

  commit f28d3d5346e97e60c81f933ac89ccf015430e5cf
  Author: Anna-Maria Gleixner <anna-maria@linutronix.de>
  Date:   Thu Mar 21 13:09:21 2019 +0100

    timer/trace: Improve timer tracing

    Timers are added to the timer wheel off by one. This is required in
    case a timer is queued directly before incrementing jiffies to prevent
    early timer expiry.

    When reading a timer trace and relying only on the expiry time of the timer
    in the timer_start trace point and on the now in the timer_expiry_entry
    trace point, it seems that the timer fires late. With the current
    timer_expiry_entry trace point information only now=jiffies is printed but
    not the value of base->clk. This makes it impossible to draw a conclusion
    to the index of base->clk and makes it impossible to examine timer problems
    without additional trace points.

    Therefore add the base->clk value to the timer_expire_entry trace
    point, to be able to calculate the index the timer base is located at
    during collecting expired timers.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoCleanup: bitfields: streamline use of underscores
Mathieu Desnoyers [Fri, 17 May 2019 14:01:36 +0000 (10:01 -0400)] 
Cleanup: bitfields: streamline use of underscores

Do not prefix macro arguments with underscores. Use one leading
underscore as prefix for local variables defined within macros.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoSilence compiler "always false comparison" warning
Mathieu Desnoyers [Tue, 14 May 2019 14:57:13 +0000 (10:57 -0400)] 
Silence compiler "always false comparison" warning

Compiling the bitfield test with gcc -Wextra generates those warnings:

 ../../include/babeltrace/bitfield-internal.h:38:45: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
 #define _bt_is_signed_type(type) ((type) -1 < (type) 0)

This is the intent of the macro. Disable compiler warnings around use of
that macro.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com
5 years agoFix: bitfield: shift undefined/implementation defined behaviors
Mathieu Desnoyers [Tue, 14 May 2019 14:56:14 +0000 (10:56 -0400)] 
Fix: bitfield: shift undefined/implementation defined behaviors

bitfield.h uses the left shift operator with a left operand which
may be negative. The C99 standard states that shifting a negative
value is undefined.

When building with -Wshift-negative-value, we get this gcc warning:

In file included from /home/smarchi/src/babeltrace/include/babeltrace/ctfser-internal.h:44:0,
                 from /home/smarchi/src/babeltrace/ctfser/ctfser.c:42:
/home/smarchi/src/babeltrace/include/babeltrace/ctfser-internal.h: In function ‘bt_ctfser_write_unsigned_int’:
/home/smarchi/src/babeltrace/include/babeltrace/bitfield-internal.h:116:24: error: left shift of negative value [-Werror=shift-negative-value]
   mask = ~((~(type) 0) << (__start % ts));  \
                        ^
/home/smarchi/src/babeltrace/include/babeltrace/bitfield-internal.h:222:2: note: in expansion of macro ‘_bt_bitfield_write_le’
  _bt_bitfield_write_le(ptr, type, _start, _length, _v)
  ^~~~~~~~~~~~~~~~~~~~~
/home/smarchi/src/babeltrace/include/babeltrace/ctfser-internal.h:418:3: note: in expansion of macro ‘bt_bitfield_write_le’
   bt_bitfield_write_le(mmap_align_addr(ctfser->base_mma) +
   ^~~~~~~~~~~~~~~~~~~~

This boils down to the fact that the expression ~((uint8_t)0) has type
"signed int", which is used as an operand of the left shift.  This is due
to the integer promotion rules of C99 (6.3.3.1):

    If an int can represent all values of the original type, the value is
    converted to an int; otherwise, it is converted to an unsigned int.
    These are called the integer promotions. All other types are unchanged
    by the integer promotions.

We also need to cast the result explicitly into the left hand
side type to deal with:

warning: large integer implicitly truncated to unsigned type [-Woverflow]

The C99 standard states that a right shift has implementation-defined
behavior when shifting a signed negative value. Add a preprocessor check
that the compiler provides the expected behavior, else provide an
alternative implementation which guarantees the intended behavior.

A preprocessor check is also added to ensure that the compiler
representation for signed values is two's complement, which is expected
by this header.

Document that this header strictly respects the C99 standard, with
the exception of its use of __typeof__.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: timestamp_end field should include all events within sub-buffer
Mathieu Desnoyers [Tue, 30 Apr 2019 03:44:30 +0000 (23:44 -0400)] 
Fix: timestamp_end field should include all events within sub-buffer

Fix for timestamp_end not including all events within sub-buffer. This
happens if a thread is preempted/interrupted for a long time between
reserve and commit (e.g. in the middle of a packet), which causes the
timestamp used for timestamp_end field of the packet header to be lower
than the timestamp of the last events in the buffer (those following the
event that was preempted/interrupted between reserve and commit).

The fix involves sampling the timestamp when doing the last space
reservation in a sub-buffer (which necessarily happens before doing the
delivery after its last commit). Save this timestamp temporarily in a
per-sub-buffer control area (we have exclusive access to that area until
we increment the commit counter).

Then, that timestamp value will be read when delivering the sub-buffer,
whichever event or switch happens to be the last to increment the commit
counter to perform delivery. The timestamp value can be read without
worrying about concurrent access, because at that point sub-buffer
delivery has exclusive access to the sub-buffer.

This ensures the timestamp_end value is always larger or equal to the
timestamp of the last event, always below or equal the timestamp_begin
of the following packet, and always below or equal the timestamp of the
first event in the following packet.

Fixes: #1183
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: Remove start and number from syscall_get_arguments() args (v5.1)
Michael Jeanson [Tue, 9 Apr 2019 18:12:41 +0000 (14:12 -0400)] 
Fix: Remove start and number from syscall_get_arguments() args (v5.1)

  commit b35f549df1d7520d37ba1e6d4a8d4df6bd52d136
  Author: Steven Rostedt (Red Hat) <rostedt@goodmis.org>
  Date:   Mon Nov 7 16:26:37 2016 -0500

    syscalls: Remove start and number from syscall_get_arguments() args

    At Linux Plumbers, Andy Lutomirski approached me and pointed out that the
    function call syscall_get_arguments() implemented in x86 was horribly
    written and not optimized for the standard case of passing in 0 and 6 for
    the starting index and the number of system calls to get. When looking at
    all the users of this function, I discovered that all instances pass in only
    0 and 6 for these arguments. Instead of having this function handle
    different cases that are never used, simply rewrite it to return the first 6
    arguments of a system call.

    This should help out the performance of tracing system calls by ptrace,
    ftrace and perf.

Link: http://lkml.kernel.org/r/20161107213233.754809394@goodmis.org
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agolttng abi documentation: clarify getter usage requirements
Mathieu Desnoyers [Wed, 10 Apr 2019 15:13:15 +0000 (11:13 -0400)] 
lttng abi documentation: clarify getter usage requirements

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: don't access packet header for stream_id and stream_instance_id getters
Mathieu Desnoyers [Wed, 10 Apr 2019 14:25:47 +0000 (10:25 -0400)] 
Fix: don't access packet header for stream_id and stream_instance_id getters

The stream ID and stream instance ID are invariant for a stream, so
there is no point reading them from the packet header currently owned by
the consumer (between get/put subbuf).

Actually, the consumer try to access the stream_id from the live timer
when sending a live beacon without getting the reader subbuffer first.
Doing so is racy against producers. In typical live scenarios
(non-overwrite channels), the producers will always write the same
stream id and stream instance id values at the same header offsets,
which will "work", except for the initial state of an empty buffer:
the value "0" will be returned (erroneously).

For the less frequently used scenario of a live session with "overwrite"
channels, this will trigger WARN_ON safety nets in libringbuffer. This
safety net triggers a kernel OOPS report and disables tracing for that
channel.

In the case where a ring buffer does not have any data ready, it makes
no sense to try to get a subbuffer for reading anyway, so the approach
was broken.

So return the stream id and stream instance id from the internal
data structures rather than reading it from the ring buffer.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: atomic_long_add_unless() returns a boolean
Michael Jeanson [Mon, 18 Mar 2019 20:20:36 +0000 (16:20 -0400)] 
Fix: atomic_long_add_unless() returns a boolean

Because of a documentation error in older kernels, it was assumed that
atomic_long_add_unless would return the old value, but the
implementation actually returns a boolean.

Also add missing error code int 'ret' and compare against the right type
max value.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: Revert "KVM: MMU: show mmu_valid_gen..." (v5.1)
Michael Jeanson [Mon, 18 Mar 2019 20:20:35 +0000 (16:20 -0400)] 
Fix: Revert "KVM: MMU: show mmu_valid_gen..." (v5.1)

See upstream commit :

  commit b59c4830ca185ba0e9f9e046fb1cd10a4a92627a
  Author: Sean Christopherson <sean.j.christopherson@intel.com>
  Date:   Tue Feb 5 13:01:30 2019 -0800

    Revert "KVM: MMU: show mmu_valid_gen in shadow page related tracepoints"

    ...as part of removing x86 KVM's fast invalidate mechanism, i.e. this
    is one part of a revert all patches from the series that introduced the
    mechanism[1].

    This reverts commit 2248b023219251908aedda0621251cffc548f258.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: pipe: stop using ->can_merge (v5.1)
Michael Jeanson [Mon, 18 Mar 2019 20:20:34 +0000 (16:20 -0400)] 
Fix: pipe: stop using ->can_merge (v5.1)

See upstream commit:

  commit 01e7187b41191376cee8bea8de9f907b001e87b4
  Author: Jann Horn <jannh@google.com>
  Date:   Wed Jan 23 15:19:18 2019 +0100

    pipe: stop using ->can_merge

    Al Viro pointed out that since there is only one pipe buffer type to which
    new data can be appended, it isn't necessary to have a ->can_merge field in
    struct pipe_buf_operations, we can just check for a magic type.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: rcu: Remove wrapper definitions for obsolete RCU... (v5.1)
Michael Jeanson [Mon, 18 Mar 2019 20:20:33 +0000 (16:20 -0400)] 
Fix: rcu: Remove wrapper definitions for obsolete RCU... (v5.1)

See upstream commit :

commit 6ba7d681aca22e53385bdb35b1d7662e61905760
Author: Paul E. McKenney <paulmck@linux.ibm.com>
Date:   Wed Jan 9 15:22:03 2019 -0800

    rcu: Remove wrapper definitions for obsolete RCU update functions

    None of synchronize_rcu_bh, synchronize_rcu_bh_expedited, call_rcu_bh,
    rcu_barrier_bh, synchronize_sched, synchronize_sched_expedited,
    call_rcu_sched, rcu_barrier_sched, get_state_synchronize_sched, and
    cond_synchronize_sched are actually used.  This commit therefore removes
    their trivial wrapper-function definitions.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: mm: create the new vm_fault_t type (v5.1)
Michael Jeanson [Mon, 18 Mar 2019 20:20:32 +0000 (16:20 -0400)] 
Fix: mm: create the new vm_fault_t type (v5.1)

See upstream commit:

  commit 3d3539018d2cbd12e5af4a132636ee7fd8d43ef0
  Author: Souptick Joarder <jrdr.linux@gmail.com>
  Date:   Thu Mar 7 16:31:14 2019 -0800

    mm: create the new vm_fault_t type

    Page fault handlers are supposed to return VM_FAULT codes, but some
    drivers/file systems mistakenly return error numbers.  Now that all
    drivers/file systems have been converted to use the vm_fault_t return
    type, change the type definition to no longer be compatible with 'int'.
    By making it an unsigned int, the function prototype becomes
    incompatible with a function which returns int.  Sparse will detect any
    attempts to return a value which is not a VM_FAULT code.

    VM_FAULT_SET_HINDEX and VM_FAULT_GET_HINDEX values are changed to avoid
    conflict with other VM_FAULT codes.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoFix: extra-version-git.sh redirect stderr to /dev/null
Mathieu Desnoyers [Fri, 15 Mar 2019 15:13:39 +0000 (11:13 -0400)] 
Fix: extra-version-git.sh redirect stderr to /dev/null

Running make in a git repo that does not contain any tag prints:

fatal: No names found, cannot describe anything.

in the make and make clean outputs.

It's fine to have no tag name available (extra-version-git.sh will
return the value 0), but we should not print an error in the make
output. Redirect this error to /dev/null.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Suggested-by: Michael Jeanson <mjeanson@efficios.com>
5 years agoMove timekeeping blacklisting to a header file
Jonathan Rajotte [Thu, 7 Mar 2019 19:58:00 +0000 (14:58 -0500)] 
Move timekeeping blacklisting to a header file

Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
5 years agoBlacklist: kprobe for arm
Jonathan Rajotte [Thu, 7 Mar 2019 19:57:59 +0000 (14:57 -0500)] 
Blacklist: kprobe for arm

This upstream kernel commit broke optimized kprobe.

commit e46daee53bb50bde38805f1823a182979724c229
Author: Kees Cook <keescook@chromium.org>
Date:   Tue Oct 30 22:12:56 2018 +0100

    ARM: 8806/1: kprobes: Fix false positive with FORTIFY_SOURCE

    The arm compiler internally interprets an inline assembly label
    as an unsigned long value, not a pointer. As a result, under
    CONFIG_FORTIFY_SOURCE, the address of a label has a size of 4 bytes,
    which was tripping the runtime checks. Instead, we can just cast the label
    (as done with the size calculations earlier).

    Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1639397

Reported-by: William Cohen <wcohen@redhat.com>
Fixes: 6974f0c4555e ("include/linux/string.h: add the option of fortified string.h functions")
Cc: stable@vger.kernel.org
Acked-by: Laura Abbott <labbott@redhat.com>
Acked-by: Masami Hiramatsu <mhiramat@kernel.org>
Tested-by: William Cohen <wcohen@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
It was introduced in the 4.20 cycle.
It was also backported to the 4.19 and 4.14 branch.

This issue is fixed upstream by [1] and is present in the 5.0 kernel
release.

[1] 0ac569bf6a7983c0c5747d6df8db9dc05bc92b6c

The fix was backported to 4.20, 4.19 and 4.14 branch.
It is included starting at:
    v5.0.0
    v4.20.13
    v4.19.26
    v4.14.104

Fixes #1174

Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
This page took 0.052569 seconds and 4 git commands to generate.