Over the last years, I've been exploring the possibility of
verifying the Linux kernel behavior using Runtime Verification.
Runtime Verification (RV) is a lightweight (yet rigorous) method that
complements classical exhaustive verification techniques (such as model
checking and theorem proving) with a more practical approach for complex
systems.
Instead of relying on a fine-grained model of a system (e.g., a
re-implementation a instruction level), RV works by analyzing the trace of the
system's actual execution, comparing it against a formal specification of
the system behavior.
The usage of deterministic automaton for RV is a well-established
approach. In the specific case of the Linux kernel, you can check how
to model complex behavior of the Linux kernel with this paper:
De Oliveira, Daniel Bristot; Cucinotta, Tommaso; De Oliveira, Romulo Silva.
*Efficient formal verification for the Linux kernel.* In: International
Conference on Software Engineering and Formal Methods. Springer, Cham, 2019.
p. 315-332.
And how efficient is this approach here:
De Oliveira, Daniel B.; De Oliveira, Romulo S.; Cucinotta, Tommaso. *A thread
synchronization model for the PREEMPT_RT Linux kernel.* Journal of Systems
Architecture, 2020, 107: 101729.
tlrd: it is possible to model complex behaviors in a modular way, with
an acceptable overhead (even for production systems). See this
presentation at 2019's ELCE: https://www.youtube.com/watch?v=BfTuEHafNgg
Here I am proposing a more practical approach for the usage of deterministic
automata for runtime verification, and it includes:
- An interface for controlling the verification;
- A tool and set of headers that enables the automatic code
generation of the RV monitor (Monitor Synthesis);
- Sample monitors to evaluate the interface;
Given that RV is a tracing consumer, the code is being placed inside the
tracing subsystem (Steven and I have been talking about it for a while).
Features to be added after this patchset:
- safe_wtd monitor (requires further discussion with watchdog maintainers)
- Export symbols for external modules
- dot2bpf
- Add a reactor that enables the visualization of the visited
states via KCOV (Marco Elver & Dmitry Vyukov)
- Add a CRC method to check from user-space if the values
exported by the monitor were not corrupted by any other
kernel task (Gabriele Paoloni)
Changes from v5:
- Add task monitor slot checks (Daniel/Tao)
- Reset the monitors only after initializing the data (Daniel)
- Add static for static data (Daniel/0-day)
- Change start/stop *functions to enable/disable (like the user-
interface (Daniel)
- s/init/start/ for the functions starting the monitoring (Daniel)
- Access monitoring_on and reacting_on via functions (Daniel)
- Improved vector access checks (Tao)
- cleanups (Daniel/Tao)
Changes from v4:
- The watchdog monitor will be discussed on another thread (Daniel)
- s/safe/final/ in the tracepoint definition (Daniel)
- Improved error handling at __init functions (Daniel)
- Remove the hostname from example of commands in a shell (Bagas Sanjaya)
- Added documentation about automata representation in C/DOT/Formal
and this documentation is cited in a comment on all model.h
(Steven)
- Make wwnr a single patch (Daniel/Steven)
- Add the .dot file for each monitor (Daniel)
- Add a document for each monitor (Daniel)
- Add an order for documentation in the index.rst (Daniel)
- Add wip/wwnr/... long description (Steven/Randy)
- Add comments for helper functions (Steven)
- Improve checks in da_monitor.h (Tao Zhou)
- Change final states set to bool (Tao/Daniel)
- Adjust indentation on enabling monitor/reactor (Steven)
- Use strim on buffers from user-space (Steven)
- Remove ifdefs inside functions (Steven)
- Adjust depends on RV in Kconfig (Steven)
- Check empty enabled monitor list (Tao Zhou)
- Fixed Copyright (Steven)
- Adjusted structures' indentation (Steven)
- Fix rv/monitors/$monitor/enabled return value (Song Liu)
- Typos (Punit Agrawal/Randy)
- Improved python scripts w.r.t. consistency (Steve)
- Blamed myself for so many problems :-) (Daniel's mind)
Changes from v3:
- Rebased on 5.19
(rostedt's request were made on 1x1 meetings)
- Moved monitors to monitors/$name/ (Rostedt)
- Consolidate the tracepoints into a single include file in the default
directory (trace/events/rv.h) (Rostedt)
- The tracepoints now record the entire string to the buffer.
- Change the enable_monitors to disable monitors with ! (instead of -).
(Rostedt)
- Add a suffix to the state/events enums, to avoid conflict in the
vmlinux.h used by eBPF.
- The models are now placed in the $name.h (it used to store the
tracepoints, but they are now consolidated in a single file)
- dot2c and dot2k updated to the changes
- models re-generated with these new standards.
- user-space tools moved to an directory outside of tools/tracing as
other methods of verification/log sources are planned.
Changes from v2:
- Tons of checkpatch and kernel test robot
- Moved files to better places
- Adjusted watchdog tracepoints patch (Guenter Roeck)
- Added pretimeout watchdog events (Peter Enderborg)
- Used task struct to store per-task monitors (Peter Zijlstra)
- Changed the instrumentation to use internal definition of tracepoint
and check the callback signature (Steven Rostedt)
- Used printk_deferred() and removed the comment about deadlocks
(Shuah Khan/John Ogness)
- Some simplifications:
- Removed the safe watchdog nowayout for now (myself)
- Removed export symbols for now (myself)
Changes from V1:
- rebased to the latest kernel;
- code cleanup;
- the watchdog dev monitor;
- safety app;
Daniel Bristot de Oliveira (16):
rv: Add Runtime Verification (RV) interface
rv: Add runtime reactors interface
rv/include: Add helper functions for deterministic automata
rv/include: Add deterministic automata monitor definition via C macros
rv/include: Add instrumentation helper functions
Documentation/rv: Add a basic documentation
tools/rv: Add dot2c
Documentation/rv: Add deterministic automaton documentation
tools/rv: Add dot2k
Documentation/rv: Add deterministic automata monitor synthesis
documentation
Documentation/rv: Add deterministic automata instrumentation
documentation
rv/monitor: Add the wip monitor skeleton created by dot2k
rv/monitor: Add the wip monitor
rv/monitor: Add the wwnr monitor
rv/reactor: Add the printk reactor
rv/reactor: Add the panic reactor
Documentation/trace/index.rst | 1 +
.../trace/rv/da_monitor_instrumentation.rst | 169 ++++
.../trace/rv/da_monitor_synthesis.rst | 147 ++++
.../trace/rv/deterministic_automata.rst | 184 +++++
Documentation/trace/rv/index.rst | 14 +
Documentation/trace/rv/monitor_wip.rst | 55 ++
Documentation/trace/rv/monitor_wwnr.rst | 45 ++
.../trace/rv/runtime-verification.rst | 231 ++++++
include/linux/rv.h | 36 +
include/linux/sched.h | 11 +
include/rv/automata.h | 75 ++
include/rv/da_monitor.h | 510 ++++++++++++
include/rv/instrumentation.h | 29 +
include/rv/rv.h | 32 +
include/trace/events/rv.h | 142 ++++
kernel/fork.c | 14 +
kernel/trace/Kconfig | 2 +
kernel/trace/Makefile | 1 +
kernel/trace/rv/Kconfig | 81 ++
kernel/trace/rv/Makefile | 8 +
kernel/trace/rv/monitors/wip/wip.c | 91 +++
kernel/trace/rv/monitors/wip/wip.h | 46 ++
kernel/trace/rv/monitors/wwnr/wwnr.c | 90 +++
kernel/trace/rv/monitors/wwnr/wwnr.h | 46 ++
kernel/trace/rv/reactor_panic.c | 43 +
kernel/trace/rv/reactor_printk.c | 42 +
kernel/trace/rv/rv.c | 739 ++++++++++++++++++
kernel/trace/rv/rv.h | 62 ++
kernel/trace/rv/rv_reactors.c | 494 ++++++++++++
kernel/trace/trace.c | 2 +
kernel/trace/trace.h | 9 +
tools/verification/dot2/Makefile | 26 +
tools/verification/dot2/automata.py | 174 +++++
tools/verification/dot2/dot2c | 26 +
tools/verification/dot2/dot2c.py | 254 ++++++
tools/verification/dot2/dot2k | 47 ++
tools/verification/dot2/dot2k.py | 177 +++++
.../dot2/dot2k_templates/main_global.c | 94 +++
.../dot2/dot2k_templates/main_per_cpu.c | 94 +++
.../dot2/dot2k_templates/main_per_task.c | 94 +++
tools/verification/models/wip.dot | 16 +
tools/verification/models/wwnr.dot | 16 +
42 files changed, 4469 insertions(+)
create mode 100644 Documentation/trace/rv/da_monitor_instrumentation.rst
create mode 100644 Documentation/trace/rv/da_monitor_synthesis.rst
create mode 100644 Documentation/trace/rv/deterministic_automata.rst
create mode 100644 Documentation/trace/rv/index.rst
create mode 100644 Documentation/trace/rv/monitor_wip.rst
create mode 100644 Documentation/trace/rv/monitor_wwnr.rst
create mode 100644 Documentation/trace/rv/runtime-verification.rst
create mode 100644 include/linux/rv.h
create mode 100644 include/rv/automata.h
create mode 100644 include/rv/da_monitor.h
create mode 100644 include/rv/instrumentation.h
create mode 100644 include/rv/rv.h
create mode 100644 include/trace/events/rv.h
create mode 100644 kernel/trace/rv/Kconfig
create mode 100644 kernel/trace/rv/Makefile
create mode 100644 kernel/trace/rv/monitors/wip/wip.c
create mode 100644 kernel/trace/rv/monitors/wip/wip.h
create mode 100644 kernel/trace/rv/monitors/wwnr/wwnr.c
create mode 100644 kernel/trace/rv/monitors/wwnr/wwnr.h
create mode 100644 kernel/trace/rv/reactor_panic.c
create mode 100644 kernel/trace/rv/reactor_printk.c
create mode 100644 kernel/trace/rv/rv.c
create mode 100644 kernel/trace/rv/rv.h
create mode 100644 kernel/trace/rv/rv_reactors.c
create mode 100644 tools/verification/dot2/Makefile
create mode 100644 tools/verification/dot2/automata.py
create mode 100644 tools/verification/dot2/dot2c
create mode 100644 tools/verification/dot2/dot2c.py
create mode 100644 tools/verification/dot2/dot2k
create mode 100644 tools/verification/dot2/dot2k.py
create mode 100644 tools/verification/dot2/dot2k_templates/main_global.c
create mode 100644 tools/verification/dot2/dot2k_templates/main_per_cpu.c
create mode 100644 tools/verification/dot2/dot2k_templates/main_per_task.c
create mode 100644 tools/verification/models/wip.dot
create mode 100644 tools/verification/models/wwnr.dot
--
2.35.1
Add the da_monitor_instrumentation.rst. It describes the basics
of RV monitor instrumentation.
Cc: Wim Van Sebroeck <[email protected]>
Cc: Guenter Roeck <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Catalin Marinas <[email protected]>
Cc: Marco Elver <[email protected]>
Cc: Dmitry Vyukov <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Cc: Shuah Khan <[email protected]>
Cc: Gabriele Paoloni <[email protected]>
Cc: Juri Lelli <[email protected]>
Cc: Clark Williams <[email protected]>
Cc: Tao Zhou <[email protected]>
Cc: Randy Dunlap <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Daniel Bristot de Oliveira <[email protected]>
---
.../trace/rv/da_monitor_instrumentation.rst | 169 ++++++++++++++++++
Documentation/trace/rv/index.rst | 1 +
2 files changed, 170 insertions(+)
create mode 100644 Documentation/trace/rv/da_monitor_instrumentation.rst
diff --git a/Documentation/trace/rv/da_monitor_instrumentation.rst b/Documentation/trace/rv/da_monitor_instrumentation.rst
new file mode 100644
index 000000000000..f8c181d2f94f
--- /dev/null
+++ b/Documentation/trace/rv/da_monitor_instrumentation.rst
@@ -0,0 +1,169 @@
+Deterministic Automata Instrumentation
+========================================
+
+The RV monitor file created by dot2k, with the name "$MODEL_NAME.c"
+includes a section dedicated to instrumentation.
+
+In the example of the wip.dot monitor created on [1], it will look like::
+
+ /*
+ * This is the instrumentation part of the monitor.
+ *
+ * This is the section where manual work is required. Here the kernel events
+ * are translated into model's event.
+ *
+ */
+ static void handle_preempt_disable(void *data, /* XXX: fill header */)
+ {
+ da_handle_event_wip(preempt_disable_wip);
+ }
+
+ static void handle_preempt_enable(void *data, /* XXX: fill header */)
+ {
+ da_handle_event_wip(preempt_enable_wip);
+ }
+
+ static void handle_sched_waking(void *data, /* XXX: fill header */)
+ {
+ da_handle_event_wip(sched_waking_wip);
+ }
+
+ static int enable_wip(void)
+ {
+ int retval;
+
+ retval = da_monitor_init_wip();
+ if (retval)
+ return retval;
+
+ rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_disable);
+ rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_enable);
+ rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_sched_waking);
+
+ return 0;
+ }
+
+The comment at the top of the section explains the general idea: the
+instrumentation section translates *kernel events* into the *model's
+event*.
+
+Tracing callback functions
+-----------------------------
+
+The first three functions are the starting point of the callback *handler
+functions* for each of the three events from the wip model. The developer
+does not necessarily need to use them: they are just starting points.
+
+Using the example of::
+
+ void handle_preempt_disable(void *data, /* XXX: fill header */)
+ {
+ da_handle_event_wip(preempt_disable_wip);
+ }
+
+The preempt_disable event from the model connects directly to the
+preemptirq:preempt_disable. The preemptirq:preempt_disable event
+has the following signature, from include/trace/events/preemptirq.h::
+
+ TP_PROTO(unsigned long ip, unsigned long parent_ip)
+
+Hence, the handle_preempt_disable() function will look like::
+
+ void handle_preempt_disable(void *data, unsigned long ip, unsigned long parent_ip)
+
+In this case, the kernel event translates one to one with the automata
+event, and indeed, no other change is required for this function.
+
+The next handler function, handle_preempt_enable() has the same argument
+list from the handle_preempt_disable(). The difference is that the
+preempt_enable event will be used to synchronize the system to the model.
+
+Initially, the *model* is placed in the initial state. However, the *system*
+might or might not be in the initial state. The monitor cannot start
+processing events until it knows that the system has reached the initial state.
+Otherwise, the monitor and the system could be out-of-sync.
+
+Looking at the automata definition, it is possible to see that the system
+and the model are expected to return to the initial state after the
+preempt_enable execution. Hence, it can be used to synchronize the
+system and the model at the initialization of the monitoring section.
+
+The start is informed via a special handle function, the
+"da_handle_start_event_MONITOR_event)", in this case::
+
+ da_handle_start_event_wip(preempt_disable_wip);
+
+So, the callback function will look like::
+
+ void handle_preempt_enable(void *data, unsigned long ip, unsigned long parent_ip)
+ {
+ da_handle_start_event_wip(preempt_enable_wip);
+ }
+
+Finally, the "handle_sched_waking()" will look like::
+
+ void handle_sched_waking(void *data, struct task_struct *task)
+ {
+ da_handle_event_wip(sched_waking_wip);
+ }
+
+And the explanation is left for the reader as an exercise.
+
+Start and Stop functions
+------------------------
+
+dot2k automatically creates two special functions::
+
+ enable_MONITOR()
+ disable_MONITOR()
+
+These functions are called when the monitor is enabled and disabled,
+respectively.
+
+They should be used to *attach* and *detach* the instrumentation to the running
+system. The developer must add to the relative function all that is needed to
+*attach* and *detach* its monitor to the system.
+
+For the wip case, these functions were named::
+
+ enable_wip()
+ disable_wip()
+
+But no change was required because: by default, these functions *attach* and
+*detach* the tracepoints_to_attach, which was enough for this case.
+
+Instrumentation helpers
+--------------------------
+
+To complete the instrumentation, the *handler functions* need to be attached to a
+kernel event, at the monitoring enable phase.
+
+The RV interface also facilitates this step. For example, the macro "rv_attach_trace_probe()"
+is used to connect the wip model events to the relative kernel event. dot2k automatically
+adds "rv_attach_trace_probe()" function call for each model event in the enable phase, as
+a suggestion.
+
+For example, from the wip sample model::
+
+ static int enable_wip(void)
+ {
+ int retval;
+
+ retval = da_monitor_init_wip();
+ if (retval)
+ return retval;
+
+ rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_enable);
+ rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_sched_waking);
+ rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_disable);
+
+ return 0;
+ }
+
+The probes then need to be detached at the disable phase.
+
+[1] The wip model is presented in:
+ Documentation/trace/rv/deterministic_automata.rst
+
+ The wip monitor is presented in:
+ Documentation/trace/rv/da_monitor_synthesis.rst
diff --git a/Documentation/trace/rv/index.rst b/Documentation/trace/rv/index.rst
index 46d47f33052c..db2ae3f90b90 100644
--- a/Documentation/trace/rv/index.rst
+++ b/Documentation/trace/rv/index.rst
@@ -9,3 +9,4 @@ Runtime Verification
runtime-verification.rst
deterministic_automata.rst
da_monitor_synthesis.rst
+ da_monitor_instrumentation.rst
--
2.35.1
In Linux terms, the runtime verification monitors are encapsulated
inside the "RV monitor" abstraction. The "RV monitor" includes a set
of instances of the monitor (per-cpu monitor, per-task monitor, and
so on), the helper functions that glue the monitor to the system
reference model, and the trace output as a reaction for event parsing
and exceptions, as depicted below:
Linux +----- RV Monitor ----------------------------------+ Formal
Realm | | Realm
+-------------------+ +----------------+ +-----------------+
| Linux kernel | | Monitor | | Reference |
| Tracing | -> | Instance(s) | <- | Model |
| (instrumentation) | | (verification) | | (specification) |
+-------------------+ +----------------+ +-----------------+
| | |
| V |
| +----------+ |
| | Reaction | |
| +--+--+--+-+ |
| | | | |
| | | +-> trace output ? |
+------------------------|--|----------------------+
| +----> panic ?
+-------> <user-specified>
Add the rv/da_monitor.h, enabling automatic code generation for the
*Monitor Instance(s)* using C macros, and code to support it.
The benefits of the usage of macro for monitor synthesis are 3-fold as it:
- Reduces the code duplication;
- Facilitates the bug fix/improvement;
- Avoids the case of developers changing the core of the monitor code
to manipulate the model in a (let's say) non-standard way.
This initial implementation presents three different types of monitor
instances:
- DECLARE_DA_MON_GLOBAL(name, type)
- DECLARE_DA_MON_PER_CPU(name, type)
- DECLARE_DA_MON_PER_TASK(name, type)
The first declares the functions for a global deterministic automata monitor,
the second for monitors with per-cpu instances, and the third with per-task
instances.
Cc: Wim Van Sebroeck <[email protected]>
Cc: Guenter Roeck <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Catalin Marinas <[email protected]>
Cc: Marco Elver <[email protected]>
Cc: Dmitry Vyukov <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Cc: Shuah Khan <[email protected]>
Cc: Gabriele Paoloni <[email protected]>
Cc: Juri Lelli <[email protected]>
Cc: Clark Williams <[email protected]>
Cc: Tao Zhou <[email protected]>
Cc: Randy Dunlap <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Daniel Bristot de Oliveira <[email protected]>
---
include/linux/rv.h | 2 +
include/rv/da_monitor.h | 507 ++++++++++++++++++++++++++++++++++++++
include/rv/rv.h | 9 +
include/trace/events/rv.h | 120 +++++++++
kernel/fork.c | 14 ++
kernel/trace/rv/Kconfig | 14 ++
kernel/trace/rv/rv.c | 5 +
7 files changed, 671 insertions(+)
create mode 100644 include/rv/da_monitor.h
create mode 100644 include/trace/events/rv.h
diff --git a/include/linux/rv.h b/include/linux/rv.h
index 4f5b70eee557..31d8b2614eae 100644
--- a/include/linux/rv.h
+++ b/include/linux/rv.h
@@ -7,6 +7,8 @@
#ifndef _LINUX_RV_H
#define _LINUX_RV_H
+#define MAX_DA_NAME_LEN 24
+
struct rv_reactor {
char *name;
char *description;
diff --git a/include/rv/da_monitor.h b/include/rv/da_monitor.h
new file mode 100644
index 000000000000..ef7ee3ffcad6
--- /dev/null
+++ b/include/rv/da_monitor.h
@@ -0,0 +1,507 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2019-2022 Red Hat, Inc. Daniel Bristot de Oliveira <[email protected]>
+ *
+ * Deterministic automata (DA) monitor functions, to be used together
+ * with automata models in C generated by the dot2k tool.
+ *
+ * The dot2k tool is available at tools/verification/dot2k/
+ */
+
+#include <rv/automata.h>
+#include <linux/rv.h>
+#include <linux/bug.h>
+
+/*
+ * Generic helpers for all types of deterministic automata monitors.
+ */
+#define DECLARE_DA_MON_GENERIC_HELPERS(name, type) \
+ \
+static char REACT_MSG[1024]; \
+ \
+static inline char *format_react_msg(type curr_state, type event) \
+{ \
+ snprintf(REACT_MSG, 1024, \
+ "rv: monitor %s does not allow event %s on state %s\n", \
+ #name, \
+ model_get_event_name_##name(event), \
+ model_get_state_name_##name(curr_state)); \
+ return REACT_MSG; \
+} \
+ \
+static void cond_react(char *msg) \
+{ \
+ if (rv_##name.react) \
+ rv_##name.react(msg); \
+} \
+ \
+/* \
+ * da_monitor_reset_##name - reset a monitor and setting it to init state \
+ */ \
+static inline void da_monitor_reset_##name(struct da_monitor *da_mon) \
+{ \
+ da_mon->monitoring = 0; \
+ da_mon->curr_state = model_get_initial_state_##name(); \
+} \
+ \
+/* \
+ * da_monitor_curr_state_##name - return the current state \
+ */ \
+static inline type da_monitor_curr_state_##name(struct da_monitor *da_mon) \
+{ \
+ return da_mon->curr_state; \
+} \
+ \
+/* \
+ * da_monitor_set_state_##name - set the new current state \
+ */ \
+static inline void \
+da_monitor_set_state_##name(struct da_monitor *da_mon, enum states_##name state) \
+{ \
+ da_mon->curr_state = state; \
+} \
+ \
+/* \
+ * da_monitor_start_##name - start monitoring \
+ * \
+ * The monitor will ignore all events until monitoring is set to true. This \
+ * function needs to be called to tell the monitor to start monitoring. \
+ */ \
+static inline void da_monitor_start_##name(struct da_monitor *da_mon) \
+{ \
+ da_mon->monitoring = 1; \
+} \
+ \
+/* \
+ * da_monitoring_##name - returns true if the monitor is processing events \
+ */ \
+static inline bool da_monitoring_##name(struct da_monitor *da_mon) \
+{ \
+ return da_mon->monitoring; \
+} \
+ \
+/* \
+ * da_monitor_enabled_##name - checks if the monitor is enabled \
+ */ \
+static inline bool da_monitor_enabled_##name(void) \
+{ \
+ /* global switch */ \
+ if (unlikely(!rv_monitoring_on())) \
+ return 0; \
+ \
+ /* monitor enabled */ \
+ if (unlikely(!rv_##name.enabled)) \
+ return 0; \
+ \
+ return 1; \
+} \
+ \
+/* \
+ * da_monitor_handling_event_##name - checks if the monitor is ready to handle events \
+ */ \
+static inline bool da_monitor_handling_event_##name(struct da_monitor *da_mon) \
+{ \
+ \
+ if (!da_monitor_enabled_##name()) \
+ return 0; \
+ \
+ /* monitor is actually monitoring */ \
+ if (unlikely(!da_monitoring_##name(da_mon))) \
+ return 0; \
+ \
+ return 1; \
+}
+
+/*
+ * Event handler for implicit monitors. Implicit monitor is the one which the
+ * handler does not need to specify which da_monitor to manipulate. Examples
+ * of implicit monitor are the per_cpu or the global ones.
+ */
+#define DECLARE_DA_MON_MODEL_HANDLER_IMPLICIT(name, type) \
+ \
+static inline bool \
+da_event_##name(struct da_monitor *da_mon, enum events_##name event) \
+{ \
+ type curr_state = da_monitor_curr_state_##name(da_mon); \
+ type next_state = model_get_next_state_##name(curr_state, event); \
+ \
+ if (next_state != INVALID_STATE) { \
+ da_monitor_set_state_##name(da_mon, next_state); \
+ \
+ trace_event_##name(model_get_state_name_##name(curr_state), \
+ model_get_event_name_##name(event), \
+ model_get_state_name_##name(next_state), \
+ model_is_final_state_##name(next_state)); \
+ \
+ return true; \
+ } \
+ \
+ if (rv_reacting_on()) \
+ cond_react(format_react_msg(curr_state, event)); \
+ \
+ trace_error_##name(model_get_state_name_##name(curr_state), \
+ model_get_event_name_##name(event)); \
+ \
+ return false; \
+} \
+
+/*
+ * Event handler for per_task monitors.
+ */
+#define DECLARE_DA_MON_MODEL_HANDLER_PER_TASK(name, type) \
+ \
+static inline type da_event_##name(struct da_monitor *da_mon, struct task_struct *tsk, \
+ enum events_##name event) \
+{ \
+ type curr_state = da_monitor_curr_state_##name(da_mon); \
+ type next_state = model_get_next_state_##name(curr_state, event); \
+ \
+ if (next_state != INVALID_STATE) { \
+ da_monitor_set_state_##name(da_mon, next_state); \
+ \
+ trace_event_##name(tsk->pid, \
+ model_get_state_name_##name(curr_state), \
+ model_get_event_name_##name(event), \
+ model_get_state_name_##name(next_state), \
+ model_is_final_state_##name(next_state)); \
+ \
+ return true; \
+ } \
+ \
+ if (rv_reacting_on()) \
+ cond_react(format_react_msg(curr_state, event)); \
+ \
+ trace_error_##name(tsk->pid, \
+ model_get_state_name_##name(curr_state), \
+ model_get_event_name_##name(event)); \
+ \
+ return false; \
+}
+
+/*
+ * Functions to define, init and get a global monitor.
+ */
+#define DECLARE_DA_MON_INIT_GLOBAL(name, type) \
+ \
+/* \
+ * global monitor (a single variable) \
+ */ \
+static struct da_monitor da_mon_##name; \
+ \
+/* \
+ * da_get_monitor_##name - return the global monitor address \
+ */ \
+static struct da_monitor *da_get_monitor_##name(void) \
+{ \
+ return &da_mon_##name; \
+} \
+ \
+/* \
+ * da_monitor_reset_all_##name - reset the single monitor \
+ */ \
+static void da_monitor_reset_all_##name(void) \
+{ \
+ da_monitor_reset_##name(da_get_monitor_##name()); \
+} \
+ \
+/* \
+ * da_monitor_init_##name - initialize a monitor \
+ */ \
+static inline int da_monitor_init_##name(void) \
+{ \
+ da_monitor_reset_all_##name(); \
+ return 0; \
+} \
+ \
+/* \
+ * da_monitor_destroy_##name - destroy the monitor \
+ */ \
+static inline void da_monitor_destroy_##name(void) \
+{ \
+ return; \
+}
+
+/*
+ * Functions to define, init and get a per-cpu monitor.
+ */
+#define DECLARE_DA_MON_INIT_PER_CPU(name, type) \
+ \
+/* \
+ * per-cpu monitor variables \
+ */ \
+DEFINE_PER_CPU(struct da_monitor, da_mon_##name); \
+ \
+/* \
+ * da_get_monitor_##name - return current CPU monitor address \
+ */ \
+static struct da_monitor *da_get_monitor_##name(void) \
+{ \
+ return this_cpu_ptr(&da_mon_##name); \
+} \
+ \
+/* \
+ * da_monitor_reset_all_##name - reset all CPUs' monitor \
+ */ \
+static void da_monitor_reset_all_##name(void) \
+{ \
+ struct da_monitor *da_mon; \
+ int cpu; \
+ for_each_cpu(cpu, cpu_online_mask) { \
+ da_mon = per_cpu_ptr(&da_mon_##name, cpu); \
+ da_monitor_reset_##name(da_mon); \
+ } \
+} \
+ \
+/* \
+ * da_monitor_init_##name - initialize all CPUs' monitor \
+ */ \
+static inline int da_monitor_init_##name(void) \
+{ \
+ da_monitor_reset_all_##name(); \
+ return 0; \
+} \
+ \
+/* \
+ * da_monitor_destroy_##name - destroy the monitor \
+ */ \
+static inline void da_monitor_destroy_##name(void) \
+{ \
+ return; \
+}
+
+/*
+ * Functions to define, init and get a per-task monitor.
+ */
+#define DECLARE_DA_MON_INIT_PER_TASK(name, type) \
+ \
+/* \
+ * The per-task monitor is stored a vector in the task struct. This variable \
+ * stores the position on the vector reserved for this monitor. \
+ */ \
+static int task_mon_slot_##name = RV_PER_TASK_MONITOR_INIT; \
+ \
+/* \
+ * da_get_monitor_##name - return the monitor in the allocated slot for tsk \
+ */ \
+static inline struct da_monitor *da_get_monitor_##name(struct task_struct *tsk) \
+{ \
+ return &tsk->rv[task_mon_slot_##name].da_mon; \
+} \
+ \
+static void da_monitor_reset_all_##name(void) \
+{ \
+ struct task_struct *g, *p; \
+ \
+ read_lock(&tasklist_lock); \
+ for_each_process_thread(g, p) \
+ da_monitor_reset_##name(da_get_monitor_##name(p)); \
+ read_unlock(&tasklist_lock); \
+} \
+ \
+/* \
+ * da_monitor_init_##name - initialize the per-task monitor \
+ * \
+ * Try to allocate a slot in the task's vector of monitors. If there \
+ * is an available slot, use it and reset all task's monitor. \
+ */ \
+static int da_monitor_init_##name(void) \
+{ \
+ int slot; \
+ \
+ slot = get_task_monitor_slot(); \
+ if (slot < 0 || slot >= RV_PER_TASK_MONITOR_INIT) \
+ return slot; \
+ \
+ task_mon_slot_##name = slot; \
+ \
+ da_monitor_reset_all_##name(); \
+ return 0; \
+} \
+ \
+/* \
+ * da_monitor_destroy_##name - return the allocated slot \
+ */ \
+static inline void da_monitor_destroy_##name(void) \
+{ \
+ if (task_mon_slot_##name == RV_PER_TASK_MONITOR_INIT) { \
+ WARN_ONCE(1, "Disabling a disabled monitor: " #name); \
+ return; \
+ } \
+ put_task_monitor_slot(task_mon_slot_##name); \
+ task_mon_slot_##name = RV_PER_TASK_MONITOR_INIT; \
+ return; \
+}
+
+/*
+ * Handle event for implicit monitor: da_get_monitor_##name() will figure out
+ * the monitor.
+ */
+#define DECLARE_DA_MON_MONITOR_HANDLER_IMPLICIT(name, type) \
+ \
+static inline void __da_handle_event_##name(struct da_monitor *da_mon, \
+ enum events_##name event) \
+{ \
+ int retval; \
+ \
+ retval = da_monitor_handling_event_##name(da_mon); \
+ if (!retval) \
+ return; \
+ \
+ retval = da_event_##name(da_mon, event); \
+ if (!retval) \
+ da_monitor_reset_##name(da_mon); \
+} \
+ \
+/* \
+ * da_handle_event_##name - handle an event \
+ */ \
+static inline void da_handle_event_##name(enum events_##name event) \
+{ \
+ struct da_monitor *da_mon = da_get_monitor_##name(); \
+ __da_handle_event_##name(da_mon, event); \
+} \
+ \
+/* \
+ * da_handle_start_event_##name - start monitoring or handle event \
+ * \
+ * This function is used notify the monitor that the system is returning \
+ * to the initial state, so the monitor can start monitoring in the next event. \
+ * Thus: \
+ * \
+ * If the monitor already started, handle the event. \
+ * If the monitor did not start yet, start the monitor but skip the event. \
+ */ \
+static inline bool da_handle_start_event_##name(enum events_##name event) \
+{ \
+ struct da_monitor *da_mon; \
+ \
+ if (!da_monitor_enabled_##name()) \
+ return 0; \
+ \
+ da_mon = da_get_monitor_##name(); \
+ \
+ if (unlikely(!da_monitoring_##name(da_mon))) { \
+ da_monitor_start_##name(da_mon); \
+ return 0; \
+ } \
+ \
+ __da_handle_event_##name(da_mon, event); \
+ \
+ return 1; \
+} \
+ \
+/* \
+ * da_handle_start_run_event_##name - start monitoring and handle event \
+ * \
+ * This function is used notify the monitor that the system is in the \
+ * initial state, so the monitor can start monitoring and handling event. \
+ */ \
+static inline bool da_handle_start_run_event_##name(enum events_##name event) \
+{ \
+ struct da_monitor *da_mon; \
+ \
+ if (!da_monitor_enabled_##name()) \
+ return 0; \
+ \
+ da_mon = da_get_monitor_##name(); \
+ \
+ if (unlikely(!da_monitoring_##name(da_mon))) \
+ da_monitor_start_##name(da_mon); \
+ \
+ __da_handle_event_##name(da_mon, event); \
+ \
+ return 1; \
+}
+
+/*
+ * Handle event for per task.
+ */
+#define DECLARE_DA_MON_MONITOR_HANDLER_PER_TASK(name, type) \
+ \
+static inline void \
+__da_handle_event_##name(struct da_monitor *da_mon, struct task_struct *tsk, \
+ enum events_##name event) \
+{ \
+ int retval; \
+ \
+ retval = da_monitor_handling_event_##name(da_mon); \
+ if (!retval) \
+ return; \
+ \
+ retval = da_event_##name(da_mon, tsk, event); \
+ if (!retval) \
+ da_monitor_reset_##name(da_mon); \
+} \
+ \
+/* \
+ * da_handle_event_##name - handle an event \
+ */ \
+static inline void \
+da_handle_event_##name(struct task_struct *tsk, enum events_##name event) \
+{ \
+ struct da_monitor *da_mon = da_get_monitor_##name(tsk); \
+ __da_handle_event_##name(da_mon, tsk, event); \
+} \
+ \
+/* \
+ * da_handle_start_event_##name - start monitoring or handle event \
+ * \
+ * This function is used notify the monitor that the system is returning \
+ * to the initial state, so the monitor can start monitoring in the next event. \
+ * Thus: \
+ * \
+ * If the monitor already started, handle the event. \
+ * If the monitor did not start yet, start the monitor but skip the event. \
+ */ \
+static inline bool \
+da_handle_start_event_##name(struct task_struct *tsk, enum events_##name event) \
+{ \
+ struct da_monitor *da_mon; \
+ \
+ if (!da_monitor_enabled_##name()) \
+ return 0; \
+ \
+ da_mon = da_get_monitor_##name(tsk); \
+ \
+ if (unlikely(!da_monitoring_##name(da_mon))) { \
+ da_monitor_start_##name(da_mon); \
+ return 0; \
+ } \
+ \
+ __da_handle_event_##name(da_mon, tsk, event); \
+ \
+ return 1; \
+}
+
+/*
+ * Entry point for the global monitor.
+ */
+#define DECLARE_DA_MON_GLOBAL(name, type) \
+ \
+DECLARE_AUTOMATA_HELPERS(name, type) \
+DECLARE_DA_MON_GENERIC_HELPERS(name, type) \
+DECLARE_DA_MON_MODEL_HANDLER_IMPLICIT(name, type) \
+DECLARE_DA_MON_INIT_GLOBAL(name, type) \
+DECLARE_DA_MON_MONITOR_HANDLER_IMPLICIT(name, type)
+
+/*
+ * Entry point for the per-cpu monitor.
+ */
+#define DECLARE_DA_MON_PER_CPU(name, type) \
+ \
+DECLARE_AUTOMATA_HELPERS(name, type) \
+DECLARE_DA_MON_GENERIC_HELPERS(name, type) \
+DECLARE_DA_MON_MODEL_HANDLER_IMPLICIT(name, type) \
+DECLARE_DA_MON_INIT_PER_CPU(name, type) \
+DECLARE_DA_MON_MONITOR_HANDLER_IMPLICIT(name, type)
+
+/*
+ * Entry point for the per-task monitor.
+ */
+#define DECLARE_DA_MON_PER_TASK(name, type) \
+ \
+DECLARE_AUTOMATA_HELPERS(name, type) \
+DECLARE_DA_MON_GENERIC_HELPERS(name, type) \
+DECLARE_DA_MON_MODEL_HANDLER_PER_TASK(name, type) \
+DECLARE_DA_MON_INIT_PER_TASK(name, type) \
+DECLARE_DA_MON_MONITOR_HANDLER_PER_TASK(name, type)
diff --git a/include/rv/rv.h b/include/rv/rv.h
index 27a108881d35..b0658cdc53d9 100644
--- a/include/rv/rv.h
+++ b/include/rv/rv.h
@@ -3,6 +3,14 @@
#ifndef _RV_RV_H
#define _RV_RV_H
+/*
+ * Deterministic automaton per-object variables.
+ */
+struct da_monitor {
+ bool monitoring;
+ int curr_state;
+};
+
/*
* Per-task RV monitors count. Nowadays fixed in RV_PER_TASK_MONITORS.
* If we find justification for more monitors, we can think about
@@ -16,6 +24,7 @@
* Futher monitor types are expected, so make this a union.
*/
union rv_task_monitor {
+ struct da_monitor da_mon;
};
int get_task_monitor_slot(void);
diff --git a/include/trace/events/rv.h b/include/trace/events/rv.h
new file mode 100644
index 000000000000..20a2e09c6416
--- /dev/null
+++ b/include/trace/events/rv.h
@@ -0,0 +1,120 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM rv
+
+#if !defined(_TRACE_RV_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_RV_H
+
+#include <linux/rv.h>
+#include <linux/tracepoint.h>
+
+#ifdef CONFIG_DA_MON_EVENTS_IMPLICIT
+DECLARE_EVENT_CLASS(event_da_monitor,
+
+ TP_PROTO(char *state, char *event, char *next_state, bool final_state),
+
+ TP_ARGS(state, event, next_state, final_state),
+
+ TP_STRUCT__entry(
+ __array( char, state, MAX_DA_NAME_LEN )
+ __array( char, event, MAX_DA_NAME_LEN )
+ __array( char, next_state, MAX_DA_NAME_LEN )
+ __field( bool, final_state )
+ ),
+
+ TP_fast_assign(
+ memcpy(__entry->state, state, MAX_DA_NAME_LEN);
+ memcpy(__entry->event, event, MAX_DA_NAME_LEN);
+ memcpy(__entry->next_state, next_state, MAX_DA_NAME_LEN);
+ __entry->final_state = final_state;
+ ),
+
+ TP_printk("%s x %s -> %s %s",
+ __entry->state,
+ __entry->event,
+ __entry->next_state,
+ __entry->final_state ? "(final)" : "")
+);
+
+DECLARE_EVENT_CLASS(error_da_monitor,
+
+ TP_PROTO(char *state, char *event),
+
+ TP_ARGS(state, event),
+
+ TP_STRUCT__entry(
+ __array( char, state, MAX_DA_NAME_LEN )
+ __array( char, event, MAX_DA_NAME_LEN )
+ ),
+
+ TP_fast_assign(
+ memcpy(__entry->state, state, MAX_DA_NAME_LEN);
+ memcpy(__entry->event, event, MAX_DA_NAME_LEN);
+ ),
+
+ TP_printk("event %s not expected in the state %s",
+ __entry->event,
+ __entry->state)
+);
+#endif /* CONFIG_DA_MON_EVENTS_IMPLICIT */
+
+#ifdef CONFIG_DA_MON_EVENTS_ID
+DECLARE_EVENT_CLASS(event_da_monitor_id,
+
+ TP_PROTO(int id, char *state, char *event, char *next_state, bool final_state),
+
+ TP_ARGS(id, state, event, next_state, final_state),
+
+ TP_STRUCT__entry(
+ __field( int, id )
+ __array( char, state, MAX_DA_NAME_LEN )
+ __array( char, event, MAX_DA_NAME_LEN )
+ __array( char, next_state, MAX_DA_NAME_LEN )
+ __field( bool, final_state )
+ ),
+
+ TP_fast_assign(
+ memcpy(__entry->state, state, MAX_DA_NAME_LEN);
+ memcpy(__entry->event, event, MAX_DA_NAME_LEN);
+ memcpy(__entry->next_state, next_state, MAX_DA_NAME_LEN);
+ __entry->id = id;
+ __entry->final_state = final_state;
+ ),
+
+ TP_printk("%d: %s x %s -> %s %s",
+ __entry->id,
+ __entry->state,
+ __entry->event,
+ __entry->next_state,
+ __entry->final_state ? "(final)" : "")
+);
+
+DECLARE_EVENT_CLASS(error_da_monitor_id,
+
+ TP_PROTO(int id, char *state, char *event),
+
+ TP_ARGS(id, state, event),
+
+ TP_STRUCT__entry(
+ __field( int, id )
+ __array( char, state, MAX_DA_NAME_LEN )
+ __array( char, event, MAX_DA_NAME_LEN )
+ ),
+
+ TP_fast_assign(
+ memcpy(__entry->state, state, MAX_DA_NAME_LEN);
+ memcpy(__entry->event, event, MAX_DA_NAME_LEN);
+ __entry->id = id;
+ ),
+
+ TP_printk("%d: event %s not expected in the state %s",
+ __entry->id,
+ __entry->event,
+ __entry->state)
+);
+#endif /* CONFIG_DA_MON_EVENTS_ID */
+#endif /* _TRACE_RV_H */
+
+/* This part ust be outside protection */
+#undef TRACE_INCLUDE_PATH
+#include <trace/define_trace.h>
diff --git a/kernel/fork.c b/kernel/fork.c
index 9d44f2d46c69..6f1f82ccd5f2 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1964,6 +1964,18 @@ static void copy_oom_score_adj(u64 clone_flags, struct task_struct *tsk)
mutex_unlock(&oom_adj_mutex);
}
+#ifdef CONFIG_RV
+static void rv_task_fork(struct task_struct *p)
+{
+ int i;
+
+ for (i = 0; i < RV_PER_TASK_MONITORS; i++)
+ p->rv[i].da_mon.monitoring = false;
+}
+#else
+#define rv_task_fork(p) do {} while (0)
+#endif
+
/*
* This creates a new process as a copy of the old one,
* but does not actually start it yet.
@@ -2399,6 +2411,8 @@ static __latent_entropy struct task_struct *copy_process(
*/
copy_seccomp(p);
+ rv_task_fork(p);
+
rseq_fork(p, clone_flags);
/* Don't start children in a dying pid namespace */
diff --git a/kernel/trace/rv/Kconfig b/kernel/trace/rv/Kconfig
index 3eb5d48ab4f6..0123bdf7052a 100644
--- a/kernel/trace/rv/Kconfig
+++ b/kernel/trace/rv/Kconfig
@@ -1,5 +1,19 @@
# SPDX-License-Identifier: GPL-2.0-only
#
+config DA_MON_EVENTS
+ default n
+ bool
+
+config DA_MON_EVENTS_IMPLICIT
+ select DA_MON_EVENTS
+ default n
+ bool
+
+config DA_MON_EVENTS_ID
+ select DA_MON_EVENTS
+ default n
+ bool
+
menuconfig RV
bool "Runtime Verification"
depends on TRACING
diff --git a/kernel/trace/rv/rv.c b/kernel/trace/rv/rv.c
index eb835777a59b..00183e056dfd 100644
--- a/kernel/trace/rv/rv.c
+++ b/kernel/trace/rv/rv.c
@@ -141,6 +141,11 @@
#include <linux/slab.h>
#include <rv/rv.h>
+#ifdef CONFIG_DA_MON_EVENTS
+#define CREATE_TRACE_POINTS
+#include <trace/events/rv.h>
+#endif
+
#include "rv.h"
DEFINE_MUTEX(rv_interface_lock);
--
2.35.1
The wakeup in preemptive (wip) monitor verifies if the
wakeup events always take place with preemption disabled:
|
|
v
#==================#
H preemptive H <+
#==================# |
| |
| preempt_disable | preempt_enable
v |
sched_waking +------------------+ |
+--------------- | | |
| | non_preemptive | |
+--------------> | | -+
+------------------+
The wakeup event always takes place with preemption disabled because
of the scheduler synchronization. However, because the preempt_count
and its trace event are not atomic with regard to interrupts, some
inconsistencies might happen.
The documentation illustrates one of these cases.
Cc: Wim Van Sebroeck <[email protected]>
Cc: Guenter Roeck <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Catalin Marinas <[email protected]>
Cc: Marco Elver <[email protected]>
Cc: Dmitry Vyukov <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Cc: Shuah Khan <[email protected]>
Cc: Gabriele Paoloni <[email protected]>
Cc: Juri Lelli <[email protected]>
Cc: Clark Williams <[email protected]>
Cc: Tao Zhou <[email protected]>
Cc: Randy Dunlap <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Daniel Bristot de Oliveira <[email protected]>
---
Documentation/trace/rv/index.rst | 1 +
Documentation/trace/rv/monitor_wip.rst | 55 ++++++++++++++++++++++++++
include/trace/events/rv.h | 10 +++++
kernel/trace/rv/Kconfig | 13 ++++++
kernel/trace/rv/Makefile | 1 +
kernel/trace/rv/monitors/wip/wip.c | 51 +++++++-----------------
tools/verification/models/wip.dot | 16 ++++++++
7 files changed, 111 insertions(+), 36 deletions(-)
create mode 100644 Documentation/trace/rv/monitor_wip.rst
create mode 100644 tools/verification/models/wip.dot
diff --git a/Documentation/trace/rv/index.rst b/Documentation/trace/rv/index.rst
index db2ae3f90b90..4cb71ed628b8 100644
--- a/Documentation/trace/rv/index.rst
+++ b/Documentation/trace/rv/index.rst
@@ -10,3 +10,4 @@ Runtime Verification
deterministic_automata.rst
da_monitor_synthesis.rst
da_monitor_instrumentation.rst
+ monitor_wip.rst
diff --git a/Documentation/trace/rv/monitor_wip.rst b/Documentation/trace/rv/monitor_wip.rst
new file mode 100644
index 000000000000..a95763438c48
--- /dev/null
+++ b/Documentation/trace/rv/monitor_wip.rst
@@ -0,0 +1,55 @@
+Monitor wip
+===========
+
+- Name: wip - wakeup in preemptive
+- Type: per-cpu deterministic automaton
+- Author: Daniel Bristot de Oliveira <[email protected]>
+
+Description
+-----------
+
+The wakeup in preemptive (wip) monitor is a sample per-cpu monitor
+that verifies if the wakeup events always take place with
+preemption disabled::
+
+ |
+ |
+ v
+ #==================#
+ H preemptive H <+
+ #==================# |
+ | |
+ | preempt_disable | preempt_enable
+ v |
+ sched_waking +------------------+ |
+ +--------------- | | |
+ | | non_preemptive | |
+ +--------------> | | -+
+ +------------------+
+
+The wakeup event always takes place with preemption disabled because
+of the scheduler synchronization. However, because the preempt_count
+and its trace event are not atomic with regard to interrupts, some
+inconsistencies might happen. For example::
+
+ preempt_disable() {
+ __preempt_count_add(1)
+ -------> smp_apic_timer_interrupt() {
+ preempt_disable()
+ do not trace (preempt count >= 1)
+
+ wake up a thread
+
+ preempt_enable()
+ do not trace (preempt count >= 1)
+ }
+ <------
+ trace_preempt_disable();
+ }
+
+This problem was reported and discussed here:
+ https://lore.kernel.org/r/[email protected]/
+
+Specification
+-------------
+Grapviz Dot file in tools/verification/models/wip.dot
diff --git a/include/trace/events/rv.h b/include/trace/events/rv.h
index 20a2e09c6416..e972f27d8df3 100644
--- a/include/trace/events/rv.h
+++ b/include/trace/events/rv.h
@@ -56,6 +56,16 @@ DECLARE_EVENT_CLASS(error_da_monitor,
__entry->event,
__entry->state)
);
+
+#ifdef CONFIG_RV_MON_WIP
+DEFINE_EVENT(event_da_monitor, event_wip,
+ TP_PROTO(char *state, char *event, char *next_state, bool final_state),
+ TP_ARGS(state, event, next_state, final_state));
+
+DEFINE_EVENT(error_da_monitor, error_wip,
+ TP_PROTO(char *state, char *event),
+ TP_ARGS(state, event));
+#endif /* CONFIG_RV_MON_WIP */
#endif /* CONFIG_DA_MON_EVENTS_IMPLICIT */
#ifdef CONFIG_DA_MON_EVENTS_ID
diff --git a/kernel/trace/rv/Kconfig b/kernel/trace/rv/Kconfig
index 7ad6c93cda64..8755ad74ec22 100644
--- a/kernel/trace/rv/Kconfig
+++ b/kernel/trace/rv/Kconfig
@@ -28,6 +28,19 @@ menuconfig RV
For further information, see:
Documentation/trace/rv/runtime-verification.rst
+config RV_MON_WIP
+ depends on RV
+ depends on PREEMPT_TRACER
+ select DA_MON_EVENTS_IMPLICIT
+ bool "wip monitor"
+ help
+ Enable wip (wakeup in preemptive) sample monitor that illustrates
+ the usage of per-cpu monitors, and one limitation of the
+ preempt_disable/enable events.
+
+ For further information, see:
+ Documentation/trace/rv/monitor_wip.rst
+
config RV_REACTORS
bool "Runtime verification reactors"
default y
diff --git a/kernel/trace/rv/Makefile b/kernel/trace/rv/Makefile
index 8944274d9b41..b41109d2750a 100644
--- a/kernel/trace/rv/Makefile
+++ b/kernel/trace/rv/Makefile
@@ -2,3 +2,4 @@
obj-$(CONFIG_RV) += rv.o
obj-$(CONFIG_RV_REACTORS) += rv_reactors.o
+obj-$(CONFIG_RV_MON_WIP) += monitors/wip/wip.o
diff --git a/kernel/trace/rv/monitors/wip/wip.c b/kernel/trace/rv/monitors/wip/wip.c
index 7a06b841db0f..2fab02f1467d 100644
--- a/kernel/trace/rv/monitors/wip/wip.c
+++ b/kernel/trace/rv/monitors/wip/wip.c
@@ -10,44 +10,26 @@
#define MODULE_NAME "wip"
-/*
- * XXX: include required tracepoint headers, e.g.,
- * #include <linux/trace/events/sched.h>
- */
#include <trace/events/rv.h>
+#include <trace/events/sched.h>
+#include <trace/events/preemptirq.h>
-/*
- * This is the self-generated part of the monitor. Generally, there is no need
- * to touch this section.
- */
#include "wip.h"
-/*
- * Declare the deterministic automata monitor.
- *
- * The rv monitor reference is needed for the monitor declaration.
- */
struct rv_monitor rv_wip;
DECLARE_DA_MON_PER_CPU(wip, unsigned char);
-/*
- * This is the instrumentation part of the monitor.
- *
- * This is the section where manual work is required. Here the kernel events
- * are translated into model's event.
- *
- */
-static void handle_preempt_disable(void *data, /* XXX: fill header */)
+static void handle_preempt_disable(void *data, unsigned long ip, unsigned long parent_ip)
{
da_handle_event_wip(preempt_disable_wip);
}
-static void handle_preempt_enable(void *data, /* XXX: fill header */)
+static void handle_preempt_enable(void *data, unsigned long ip, unsigned long parent_ip)
{
- da_handle_event_wip(preempt_enable_wip);
+ da_handle_start_event_wip(preempt_enable_wip);
}
-static void handle_sched_waking(void *data, /* XXX: fill header */)
+static void handle_sched_waking(void *data, struct task_struct *task)
{
da_handle_event_wip(sched_waking_wip);
}
@@ -60,9 +42,9 @@ static int enable_wip(void)
if (retval)
return retval;
- rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_disable);
- rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_enable);
- rv_attach_trace_probe("wip", /* XXX: tracepoint */, handle_sched_waking);
+ rv_attach_trace_probe("wip", preempt_enable, handle_preempt_enable);
+ rv_attach_trace_probe("wip", sched_waking, handle_sched_waking);
+ rv_attach_trace_probe("wip", preempt_disable, handle_preempt_disable);
return 0;
}
@@ -71,19 +53,16 @@ static void disable_wip(void)
{
rv_wip.enabled = 0;
- rv_detach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_disable);
- rv_detach_trace_probe("wip", /* XXX: tracepoint */, handle_preempt_enable);
- rv_detach_trace_probe("wip", /* XXX: tracepoint */, handle_sched_waking);
+ rv_detach_trace_probe("wip", preempt_disable, handle_preempt_disable);
+ rv_detach_trace_probe("wip", preempt_enable, handle_preempt_enable);
+ rv_detach_trace_probe("wip", sched_waking, handle_sched_waking);
da_monitor_destroy_wip();
}
-/*
- * This is the monitor register section.
- */
struct rv_monitor rv_wip = {
.name = "wip",
- .description = "auto-generated wip",
+ .description = "wakeup in preemptive per-cpu testing monitor.",
.enable = enable_wip,
.disable = disable_wip,
.reset = da_monitor_reset_all_wip,
@@ -108,5 +87,5 @@ module_init(register_wip);
module_exit(unregister_wip);
MODULE_LICENSE("GPL");
-MODULE_AUTHOR("dot2k: auto-generated");
-MODULE_DESCRIPTION("wip");
+MODULE_AUTHOR("Daniel Bristot de Oliveira <[email protected]>");
+MODULE_DESCRIPTION("wip: wakeup in preemptive - per-cpu sample monitor.");
diff --git a/tools/verification/models/wip.dot b/tools/verification/models/wip.dot
new file mode 100644
index 000000000000..2a53a9700a89
--- /dev/null
+++ b/tools/verification/models/wip.dot
@@ -0,0 +1,16 @@
+digraph state_automaton {
+ {node [shape = circle] "non_preemptive"};
+ {node [shape = plaintext, style=invis, label=""] "__init_preemptive"};
+ {node [shape = doublecircle] "preemptive"};
+ {node [shape = circle] "preemptive"};
+ "__init_preemptive" -> "preemptive";
+ "non_preemptive" [label = "non_preemptive"];
+ "non_preemptive" -> "non_preemptive" [ label = "sched_waking" ];
+ "non_preemptive" -> "preemptive" [ label = "preempt_enable" ];
+ "preemptive" [label = "preemptive"];
+ "preemptive" -> "non_preemptive" [ label = "preempt_disable" ];
+ { rank = min ;
+ "__init_preemptive";
+ "preemptive";
+ }
+}
--
2.35.1
RV is a lightweight (yet rigorous) method that complements classical
exhaustive verification techniques (such as model checking and
theorem proving) with a more practical approach to complex systems.
RV works by analyzing the trace of the system's actual execution,
comparing it against a formal specification of the system behavior.
RV can give precise information on the runtime behavior of the
monitored system while enabling the reaction for unexpected
events, avoiding, for example, the propagation of a failure on
safety-critical systems.
The development of this interface roots in the development of the
paper:
De Oliveira, Daniel Bristot; Cucinotta, Tommaso; De Oliveira, Romulo
Silva. Efficient formal verification for the Linux kernel. In:
International Conference on Software Engineering and Formal Methods.
Springer, Cham, 2019. p. 315-332.
And:
De Oliveira, Daniel Bristot. Automata-based formal analysis
and verification of the real-time Linux kernel. PhD Thesis, 2020.
The RV interface resembles the tracing/ interface on purpose. The current
path for the RV interface is /sys/kernel/tracing/rv/.
It presents these files:
"available_monitors"
- List the available monitors, one per line.
For example:
# cat available_monitors
wip
wwnr
"enabled_monitors"
- Lists the enabled monitors, one per line;
- Writing to it enables a given monitor;
- Writing a monitor name with a '!' prefix disables it;
- Truncating the file disables all enabled monitors.
For example:
# cat enabled_monitors
# echo wip > enabled_monitors
# echo wwnr >> enabled_monitors
# cat enabled_monitors
wip
wwnr
# echo '!wip' >> enabled_monitors
# cat enabled_monitors
wwnr
# echo > enabled_monitors
# cat enabled_monitors
#
Note that more than one monitor can be enabled concurrently.
"monitoring_on"
- It is an on/off general switcher for monitoring. Note
that it does not disable enabled monitors or detach events,
but stop the per-entity monitors of monitoring the events
received from the system. It resembles the "tracing_on" switcher.
"monitors/"
Each monitor will have its one directory inside "monitors/". There
the monitor specific files will be presented.
The "monitors/" directory resembles the "events" directory on
tracefs.
For example:
# cd monitors/wip/
# ls
desc enable
# cat desc
wakeup in preemptive per-cpu testing monitor.
# cat enable
0
For further information, see the comments in the header of
kernel/trace/rv/rv.c from this patch.
Cc: Wim Van Sebroeck <[email protected]>
Cc: Guenter Roeck <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Catalin Marinas <[email protected]>
Cc: Marco Elver <[email protected]>
Cc: Dmitry Vyukov <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Cc: Shuah Khan <[email protected]>
Cc: Gabriele Paoloni <[email protected]>
Cc: Juri Lelli <[email protected]>
Cc: Clark Williams <[email protected]>
Cc: Tao Zhou <[email protected]>
Cc: Randy Dunlap <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Daniel Bristot de Oliveira <[email protected]>
---
include/linux/rv.h | 22 ++
include/linux/sched.h | 11 +
include/rv/rv.h | 23 ++
kernel/trace/Kconfig | 2 +
kernel/trace/Makefile | 1 +
kernel/trace/rv/Kconfig | 12 +
kernel/trace/rv/Makefile | 3 +
kernel/trace/rv/rv.c | 722 +++++++++++++++++++++++++++++++++++++++
kernel/trace/rv/rv.h | 34 ++
kernel/trace/trace.c | 2 +
kernel/trace/trace.h | 9 +
11 files changed, 841 insertions(+)
create mode 100644 include/linux/rv.h
create mode 100644 include/rv/rv.h
create mode 100644 kernel/trace/rv/Kconfig
create mode 100644 kernel/trace/rv/Makefile
create mode 100644 kernel/trace/rv/rv.c
create mode 100644 kernel/trace/rv/rv.h
diff --git a/include/linux/rv.h b/include/linux/rv.h
new file mode 100644
index 000000000000..b6bcdb4888e6
--- /dev/null
+++ b/include/linux/rv.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Runtime Verification.
+ *
+ * For futher information, see: kernel/trace/rv/rv.c.
+ */
+#ifndef _LINUX_RV_H
+#define _LINUX_RV_H
+
+struct rv_monitor {
+ const char *name;
+ const char *description;
+ bool enabled;
+ int (*enable)(void);
+ void (*disable)(void);
+ void (*reset)(void);
+};
+
+bool rv_monitoring_on(void);
+int rv_unregister_monitor(struct rv_monitor *monitor);
+int rv_register_monitor(struct rv_monitor *monitor);
+#endif /* _LINUX_RV_H */
diff --git a/include/linux/sched.h b/include/linux/sched.h
index c46f3a63b758..b037f364efdc 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -35,6 +35,7 @@
#include <linux/seqlock.h>
#include <linux/kcsan.h>
#include <asm/kmap_size.h>
+#include <rv/rv.h>
/* task_struct member predeclarations (sorted alphabetically): */
struct audit_context;
@@ -1500,6 +1501,16 @@ struct task_struct {
struct callback_head l1d_flush_kill;
#endif
+#ifdef CONFIG_RV
+ /*
+ * Per-task RV monitor. Nowadays fixed in RV_PER_TASK_MONITORS.
+ * If we find justification for more monitors, we can think
+ * about adding more or developing a dynamic method. So far,
+ * none of these are justified.
+ */
+ union rv_task_monitor rv[RV_PER_TASK_MONITORS];
+#endif
+
/*
* New fields for task_struct should be added above here, so that
* they are included in the randomized portion of task_struct.
diff --git a/include/rv/rv.h b/include/rv/rv.h
new file mode 100644
index 000000000000..27a108881d35
--- /dev/null
+++ b/include/rv/rv.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef _RV_RV_H
+#define _RV_RV_H
+
+/*
+ * Per-task RV monitors count. Nowadays fixed in RV_PER_TASK_MONITORS.
+ * If we find justification for more monitors, we can think about
+ * adding more or developing a dynamic method. So far, none of
+ * these are justified.
+ */
+#define RV_PER_TASK_MONITORS 1
+#define RV_PER_TASK_MONITOR_INIT (RV_PER_TASK_MONITORS)
+
+/*
+ * Futher monitor types are expected, so make this a union.
+ */
+union rv_task_monitor {
+};
+
+int get_task_monitor_slot(void);
+void put_task_monitor_slot(int slot);
+#endif /* _RV_RV_H */
diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index ccd6a5ade3e9..1052126bdca2 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig
@@ -1106,4 +1106,6 @@ config HIST_TRIGGERS_DEBUG
If unsure, say N.
+source "kernel/trace/rv/Kconfig"
+
endif # FTRACE
diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile
index 0d261774d6f3..c6651e16b557 100644
--- a/kernel/trace/Makefile
+++ b/kernel/trace/Makefile
@@ -106,5 +106,6 @@ obj-$(CONFIG_FPROBE) += fprobe.o
obj-$(CONFIG_RETHOOK) += rethook.o
obj-$(CONFIG_TRACEPOINT_BENCHMARK) += trace_benchmark.o
+obj-$(CONFIG_RV) += rv/
libftrace-y := ftrace.o
diff --git a/kernel/trace/rv/Kconfig b/kernel/trace/rv/Kconfig
new file mode 100644
index 000000000000..6d127cdb00dd
--- /dev/null
+++ b/kernel/trace/rv/Kconfig
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+menuconfig RV
+ bool "Runtime Verification"
+ depends on TRACING
+ help
+ Enable the kernel runtime verification infrastructure. RV is a
+ lightweight (yet rigorous) method that complements classical
+ exhaustive verification techniques (such as model checking and
+ theorem proving). RV works by analyzing the trace of the system's
+ actual execution, comparing it against a formal specification of
+ the system behavior.
diff --git a/kernel/trace/rv/Makefile b/kernel/trace/rv/Makefile
new file mode 100644
index 000000000000..fd995379df67
--- /dev/null
+++ b/kernel/trace/rv/Makefile
@@ -0,0 +1,3 @@
+# SPDX-License-Identifier: GPL-2.0
+
+obj-$(CONFIG_RV) += rv.o
diff --git a/kernel/trace/rv/rv.c b/kernel/trace/rv/rv.c
new file mode 100644
index 000000000000..89cd766cfacd
--- /dev/null
+++ b/kernel/trace/rv/rv.c
@@ -0,0 +1,722 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2019-2022 Red Hat, Inc. Daniel Bristot de Oliveira <[email protected]>
+ *
+ * This is the online Runtime Verification (RV) interface.
+ *
+ * RV is a lightweight (yet rigorous) method that complements classical
+ * exhaustive verification techniques (such as model checking and
+ * theorem proving) with a more practical approach to complex systems.
+ *
+ * RV works by analyzing the trace of the system's actual execution,
+ * comparing it against a formal specification of the system behavior.
+ * RV can give precise information on the runtime behavior of the
+ * monitored system while enabling the reaction for unexpected
+ * events, avoiding, for example, the propagation of a failure on
+ * safety-critical systems.
+ *
+ * The development of this interface roots in the development of the
+ * paper:
+ *
+ * De Oliveira, Daniel Bristot; Cucinotta, Tommaso; De Oliveira, Romulo
+ * Silva. Efficient formal verification for the Linux kernel. In:
+ * International Conference on Software Engineering and Formal Methods.
+ * Springer, Cham, 2019. p. 315-332.
+ *
+ * And:
+ *
+ * De Oliveira, Daniel Bristot, et al. Automata-based formal analysis
+ * and verification of the real-time Linux kernel. PhD Thesis, 2020.
+ *
+ * == Runtime monitor interface ==
+ *
+ * A monitor is the central part of the runtime verification of a system.
+ *
+ * The monitor stands in between the formal specification of the desired
+ * (or undesired) behavior, and the trace of the actual system.
+ *
+ * In Linux terms, the runtime verification monitors are encapsulated
+ * inside the "RV monitor" abstraction. A RV monitor includes a reference
+ * model of the system, a set of instances of the monitor (per-cpu monitor,
+ * per-task monitor, and so on), and the helper functions that glue the
+ * monitor to the system via trace. Generally, a monitor includes some form
+ * of trace output as a reaction for event parsing and exceptions,
+ * as depicted bellow:
+ *
+ * Linux +----- RV Monitor ----------------------------------+ Formal
+ * Realm | | Realm
+ * +-------------------+ +----------------+ +-----------------+
+ * | Linux kernel | | Monitor | | Reference |
+ * | Tracing | -> | Instance(s) | <- | Model |
+ * | (instrumentation) | | (verification) | | (specification) |
+ * +-------------------+ +----------------+ +-----------------+
+ * | | |
+ * | V |
+ * | +----------+ |
+ * | | Reaction | |
+ * | +--+--+--+-+ |
+ * | | | | |
+ * | | | +-> trace output ? |
+ * +------------------------|--|----------------------+
+ * | +----> panic ?
+ * +-------> <user-specified>
+ *
+ * This file implements the interface for loading RV monitors, and
+ * to control the verification session.
+ *
+ * == Registering monitors ==
+ *
+ * The struct rv_monitor defines a set of callback functions to control
+ * a verification session. For instance, when a given monitor is enabled,
+ * the "enable" callback function is called to hook the instrumentation
+ * functions to the kernel trace events. The "disable" function is called
+ * when disabling the verification session.
+ *
+ * A RV monitor is registered via:
+ * int rv_register_monitor(struct rv_monitor *monitor);
+ * And unregistered via:
+ * int rv_unregister_monitor(struct rv_monitor *monitor);
+ *
+ * == User interface ==
+ *
+ * The user interface resembles kernel tracing interface. It presents
+ * these files:
+ *
+ * "available_monitors"
+ * - List the available monitors, one per line.
+ *
+ * For example:
+ * # cat available_monitors
+ * wip
+ * wwnr
+ *
+ * "enabled_monitors"
+ * - Lists the enabled monitors, one per line;
+ * - Writing to it enables a given monitor;
+ * - Writing a monitor name with a '!' prefix disables it;
+ * - Truncating the file disables all enabled monitors.
+ *
+ * For example:
+ * # cat enabled_monitors
+ * # echo wip > enabled_monitors
+ * # echo wwnr >> enabled_monitors
+ * # cat enabled_monitors
+ * wip
+ * wwnr
+ * # echo '!wip' >> enabled_monitors
+ * # cat enabled_monitors
+ * wwnr
+ * # echo > enabled_monitors
+ * # cat enabled_monitors
+ * #
+ *
+ * Note that more than one monitor can be enabled concurrently.
+ *
+ * "monitoring_on"
+ * - It is an on/off general switcher for monitoring. Note
+ * that it does not disable enabled monitors or detach events,
+ * but stops the per-entity monitors from monitoring the events
+ * received from the instrumentation. It resembles the "tracing_on"
+ * switcher.
+ *
+ * "monitors/"
+ * Each monitor will have its own directory inside "monitors/". There
+ * the monitor specific files will be presented.
+ * The "monitors/" directory resembles the "events" directory on
+ * tracefs.
+ *
+ * For example:
+ * # cd monitors/wip/
+ * # ls
+ * desc enable
+ * # cat desc
+ * auto-generated wakeup in preemptive monitor.
+ * # cat enable
+ * 0
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <rv/rv.h>
+
+#include "rv.h"
+
+DEFINE_MUTEX(rv_interface_lock);
+
+static struct rv_interface rv_root;
+
+struct dentry *get_monitors_root(void)
+{
+ return rv_root.monitors_dir;
+}
+
+/*
+ * Monitoring on global switcher!
+ */
+static bool __read_mostly monitoring_on;
+
+/**
+ * rv_monitoring_on - checks if monitoring is on
+ *
+ * Returns 1 if on, 0 otherwise.
+ */
+bool rv_monitoring_on(void)
+{
+ return monitoring_on;
+}
+
+/*
+ * Interface for the monitor register.
+ */
+static LIST_HEAD(rv_monitors_list);
+
+static int task_monitor_count;
+static bool task_monitor_slots[RV_PER_TASK_MONITORS];
+
+int get_task_monitor_slot(void)
+{
+ int i;
+
+ lockdep_assert_held(&rv_interface_lock);
+
+ if (task_monitor_count == RV_PER_TASK_MONITORS)
+ return -EBUSY;
+
+ task_monitor_count++;
+
+ for (i = 0; i < RV_PER_TASK_MONITORS; i++) {
+ if (task_monitor_slots[i] == false) {
+ task_monitor_slots[i] = true;
+ return i;
+ }
+ }
+
+ WARN_ONCE(1, "RV task_monitor_count and slots are out of sync\n");
+
+ return -EINVAL;
+}
+
+void put_task_monitor_slot(int slot)
+{
+ lockdep_assert_held(&rv_interface_lock);
+
+ if (slot < 0 || slot >= RV_PER_TASK_MONITORS) {
+ WARN_ONCE(1, "RV releasing an invalid slot!: %d\n", slot);
+ return;
+ }
+
+ WARN_ONCE(!task_monitor_slots[slot], "RV releasing unused task_monitor_slots: %d\n",
+ slot);
+
+ task_monitor_count--;
+ task_monitor_slots[slot] = false;
+}
+
+/*
+ * This section collects the monitor/ files and folders.
+ */
+static ssize_t monitor_enable_read_data(struct file *filp, char __user *user_buf, size_t count,
+ loff_t *ppos)
+{
+ struct rv_monitor_def *mdef = filp->private_data;
+ const char *buff;
+
+ mutex_lock(&rv_interface_lock);
+ buff = mdef->monitor->enabled ? "1\n" : "0\n";
+ mutex_unlock(&rv_interface_lock);
+
+ return simple_read_from_buffer(user_buf, count, ppos, buff, strlen(buff)+1);
+}
+
+/*
+ * Disable a given runtime monitor.
+ */
+static int disable_monitor(struct rv_monitor_def *mdef)
+{
+ if (mdef->monitor->enabled) {
+ mdef->monitor->enabled = 0;
+ mdef->monitor->disable();
+ }
+
+ return 0;
+}
+
+/*
+ * Enable a given monitor.
+ */
+static int enable_monitor(struct rv_monitor_def *mdef)
+{
+ int retval;
+
+ if (!mdef->monitor->enabled) {
+ retval = mdef->monitor->enable();
+ if (retval)
+ return retval;
+ }
+
+ mdef->monitor->enabled = 1;
+
+ return 0;
+}
+
+/*
+ * interface for enabling/disabling a monitor.
+ */
+static ssize_t monitor_enable_write_data(struct file *filp, const char __user *user_buf,
+ size_t count, loff_t *ppos)
+{
+ struct rv_monitor_def *mdef = filp->private_data;
+ int retval;
+ bool val;
+
+ retval = kstrtobool_from_user(user_buf, count, &val);
+ if (retval)
+ return retval;
+
+ retval = count;
+
+ mutex_lock(&rv_interface_lock);
+
+ if (val)
+ retval = enable_monitor(mdef);
+ else
+ retval = disable_monitor(mdef);
+
+ mutex_unlock(&rv_interface_lock);
+
+ return retval ? : count;
+}
+
+static const struct file_operations interface_enable_fops = {
+ .open = simple_open,
+ .llseek = no_llseek,
+ .write = monitor_enable_write_data,
+ .read = monitor_enable_read_data,
+};
+
+/*
+ * Interface to read the enable/disable status of a monitor.
+ */
+static ssize_t monitor_desc_read_data(struct file *filp, char __user *user_buf, size_t count,
+ loff_t *ppos)
+{
+ struct rv_monitor_def *mdef = filp->private_data;
+ char buff[256];
+
+ memset(buff, 0, sizeof(buff));
+
+ mutex_lock(&rv_interface_lock);
+ snprintf(buff, sizeof(buff), "%s\n", mdef->monitor->description);
+ mutex_unlock(&rv_interface_lock);
+
+ return simple_read_from_buffer(user_buf, count, ppos, buff, strlen(buff) + 1);
+}
+
+static const struct file_operations interface_desc_fops = {
+ .open = simple_open,
+ .llseek = no_llseek,
+ .read = monitor_desc_read_data,
+};
+
+/*
+ * During the registration of a monitor, this function creates
+ * the monitor dir, where the specific options of the monitor
+ * are exposed.
+ */
+static int create_monitor_dir(struct rv_monitor_def *mdef)
+{
+ struct dentry *root = get_monitors_root();
+ const char *name = mdef->monitor->name;
+ struct dentry *tmp;
+ int retval;
+
+ mdef->root_d = rv_create_dir(name, root);
+ if (!mdef->root_d)
+ return -ENOMEM;
+
+ tmp = rv_create_file("enable", RV_MODE_WRITE, mdef->root_d, mdef, &interface_enable_fops);
+ if (!tmp) {
+ retval = -ENOMEM;
+ goto out_remove_root;
+ }
+
+ tmp = rv_create_file("desc", RV_MODE_READ, mdef->root_d, mdef, &interface_desc_fops);
+ if (!tmp) {
+ retval = -ENOMEM;
+ goto out_remove_root;
+ }
+
+ return 0;
+
+out_remove_root:
+ rv_remove(mdef->root_d);
+ return retval;
+}
+
+/*
+ * Available/Enable monitor shared seq functions.
+ */
+static int monitors_show(struct seq_file *m, void *p)
+{
+ struct rv_monitor_def *mon_def = p;
+
+ seq_printf(m, "%s\n", mon_def->monitor->name);
+ return 0;
+}
+
+/*
+ * Used by the seq file operations at the end of a read
+ * operation.
+ */
+static void monitors_stop(struct seq_file *m, void *p)
+{
+ mutex_unlock(&rv_interface_lock);
+}
+
+/*
+ * Available monitor seq functions.
+ */
+static void *available_monitors_start(struct seq_file *m, loff_t *pos)
+{
+ mutex_lock(&rv_interface_lock);
+ return seq_list_start(&rv_monitors_list, *pos);
+}
+
+static void *available_monitors_next(struct seq_file *m, void *p, loff_t *pos)
+{
+ return seq_list_next(p, &rv_monitors_list, pos);
+}
+
+/*
+ * Enable monitor seq functions.
+ */
+static void *enabled_monitors_next(struct seq_file *m, void *p, loff_t *pos)
+{
+ struct rv_monitor_def *m_def = p;
+
+ (*pos)++;
+
+ list_for_each_entry_continue(m_def, &rv_monitors_list, list) {
+ if (m_def->monitor->enabled)
+ return m_def;
+ }
+
+ return NULL;
+}
+
+static void *enabled_monitors_start(struct seq_file *m, loff_t *pos)
+{
+ struct rv_monitor_def *m_def;
+ loff_t l;
+
+ mutex_lock(&rv_interface_lock);
+
+ if (list_empty(&rv_monitors_list))
+ return NULL;
+
+ m_def = list_entry(&rv_monitors_list, struct rv_monitor_def, list);
+
+ for (l = 0; l <= *pos; ) {
+ m_def = enabled_monitors_next(m, m_def, &l);
+ if (!m_def)
+ break;
+ }
+
+ return m_def;
+}
+
+/*
+ * available/enabled monitors seq definition.
+ */
+static const struct seq_operations available_monitors_seq_ops = {
+ .start = available_monitors_start,
+ .next = available_monitors_next,
+ .stop = monitors_stop,
+ .show = monitors_show
+};
+
+static const struct seq_operations enabled_monitors_seq_ops = {
+ .start = enabled_monitors_start,
+ .next = enabled_monitors_next,
+ .stop = monitors_stop,
+ .show = monitors_show
+};
+
+/*
+ * available_monitors interface.
+ */
+static int available_monitors_open(struct inode *inode, struct file *file)
+{
+ return seq_open(file, &available_monitors_seq_ops);
+};
+
+static const struct file_operations available_monitors_ops = {
+ .open = available_monitors_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release
+};
+
+/*
+ * enabled_monitors interface.
+ */
+static void disable_all_monitors(void)
+{
+ struct rv_monitor_def *mdef;
+
+ list_for_each_entry(mdef, &rv_monitors_list, list)
+ disable_monitor(mdef);
+}
+
+static int enabled_monitors_open(struct inode *inode, struct file *file)
+{
+ if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC))
+ disable_all_monitors();
+
+ return seq_open(file, &enabled_monitors_seq_ops);
+};
+
+static ssize_t enabled_monitors_write(struct file *filp, const char __user *user_buf,
+ size_t count, loff_t *ppos)
+{
+ char buff[MAX_RV_MONITOR_NAME_SIZE + 2];
+ struct rv_monitor_def *mdef;
+ int retval = -EINVAL;
+ bool enable = true;
+ char *ptr = buff;
+ int len;
+
+ if (count < 1 || count > MAX_RV_MONITOR_NAME_SIZE + 1)
+ return -EINVAL;
+
+ memset(buff, 0, sizeof(buff));
+
+ retval = simple_write_to_buffer(buff, sizeof(buff) - 1, ppos, user_buf, count);
+ if (retval < 0)
+ return -EFAULT;
+
+ ptr = strim(buff);
+
+ if (ptr[0] == '!') {
+ enable = false;
+ ptr++;
+ }
+
+ len = strlen(ptr);
+ if (!len)
+ return count;
+
+ mutex_lock(&rv_interface_lock);
+
+ retval = -EINVAL;
+
+ list_for_each_entry(mdef, &rv_monitors_list, list) {
+ if (strcmp(ptr, mdef->monitor->name) != 0)
+ continue;
+
+ /*
+ * Monitor found!
+ */
+ if (enable)
+ retval = enable_monitor(mdef);
+ else
+ retval = disable_monitor(mdef);
+
+ if (!retval)
+ retval = count;
+
+ break;
+ }
+
+ mutex_unlock(&rv_interface_lock);
+ return retval;
+}
+
+static const struct file_operations enabled_monitors_ops = {
+ .open = enabled_monitors_open,
+ .read = seq_read,
+ .write = enabled_monitors_write,
+ .llseek = seq_lseek,
+ .release = seq_release,
+};
+
+/*
+ * monitoring_on general switcher.
+ */
+static ssize_t monitoring_on_read_data(struct file *filp, char __user *user_buf,
+ size_t count, loff_t *ppos)
+{
+ const char *buff;
+
+ mutex_lock(&rv_interface_lock);
+ buff = monitoring_on ? "1\n" : "0\n";
+ mutex_unlock(&rv_interface_lock);
+
+ return simple_read_from_buffer(user_buf, count, ppos, buff, strlen(buff) + 1);
+}
+
+static void turn_monitoring_off(void)
+{
+ monitoring_on = false;
+}
+
+static void reset_all_monitors(void)
+{
+ struct rv_monitor_def *mdef;
+
+ list_for_each_entry(mdef, &rv_monitors_list, list) {
+ if (mdef->monitor->enabled)
+ mdef->monitor->reset();
+ }
+}
+
+
+static void turn_monitoring_on(void)
+{
+ reset_all_monitors();
+ monitoring_on = true;
+}
+
+static ssize_t monitoring_on_write_data(struct file *filp, const char __user *user_buf,
+ size_t count, loff_t *ppos)
+{
+ int retval;
+ bool val;
+
+ retval = kstrtobool_from_user(user_buf, count, &val);
+ if (retval)
+ return retval;
+
+ mutex_lock(&rv_interface_lock);
+
+ if (val)
+ turn_monitoring_on();
+ else
+ turn_monitoring_off();
+
+ mutex_unlock(&rv_interface_lock);
+
+ return count;
+}
+
+static const struct file_operations monitoring_on_fops = {
+ .open = simple_open,
+ .llseek = no_llseek,
+ .write = monitoring_on_write_data,
+ .read = monitoring_on_read_data,
+};
+
+static void destroy_monitor_dir(struct rv_monitor_def *mdef)
+{
+ rv_remove(mdef->root_d);
+}
+
+/**
+ * rv_register_monitor - register a rv monitor.
+ * @monitor: The rv_monitor to be registered.
+ *
+ * Returns 0 if successful, error otherwise.
+ */
+int rv_register_monitor(struct rv_monitor *monitor)
+{
+ struct rv_monitor_def *r;
+ int retval = 0;
+
+ if (strlen(monitor->name) >= MAX_RV_MONITOR_NAME_SIZE) {
+ pr_info("Monitor %s has a name longer than %d\n", monitor->name,
+ MAX_RV_MONITOR_NAME_SIZE);
+ return -1;
+ }
+
+ mutex_lock(&rv_interface_lock);
+
+ list_for_each_entry(r, &rv_monitors_list, list) {
+ if (strcmp(monitor->name, r->monitor->name) == 0) {
+ pr_info("Monitor %s is already registered\n", monitor->name);
+ retval = -1;
+ goto out_unlock;
+ }
+ }
+
+ r = kzalloc(sizeof(struct rv_monitor_def), GFP_KERNEL);
+ if (!r) {
+ retval = -ENOMEM;
+ goto out_unlock;
+ }
+
+ r->monitor = monitor;
+
+ retval = create_monitor_dir(r);
+ if (retval) {
+ kfree(r);
+ goto out_unlock;
+ }
+
+ list_add_tail(&r->list, &rv_monitors_list);
+
+out_unlock:
+ mutex_unlock(&rv_interface_lock);
+ return retval;
+}
+
+/**
+ * rv_unregister_monitor - unregister a rv monitor.
+ * @monitor: The rv_monitor to be unregistered.
+ *
+ * Returns 0 if successful, error otherwise.
+ */
+int rv_unregister_monitor(struct rv_monitor *monitor)
+{
+ struct rv_monitor_def *ptr, *next;
+
+ mutex_lock(&rv_interface_lock);
+
+ list_for_each_entry_safe(ptr, next, &rv_monitors_list, list) {
+ if (strcmp(monitor->name, ptr->monitor->name) == 0) {
+ list_del(&ptr->list);
+ destroy_monitor_dir(ptr);
+ }
+ }
+
+ mutex_unlock(&rv_interface_lock);
+ return 0;
+}
+
+int __init rv_init_interface(void)
+{
+ struct dentry *tmp;
+
+ rv_root.root_dir = rv_create_dir("rv", NULL);
+ if (!rv_root.root_dir)
+ goto out_err;
+
+ rv_root.monitors_dir = rv_create_dir("monitors", rv_root.root_dir);
+ if (!rv_root.monitors_dir)
+ goto out_err;
+
+ tmp = rv_create_file("available_monitors", RV_MODE_READ, rv_root.root_dir, NULL,
+ &available_monitors_ops);
+ if (!tmp)
+ goto out_err;
+
+ tmp = rv_create_file("enabled_monitors", RV_MODE_WRITE, rv_root.root_dir, NULL,
+ &enabled_monitors_ops);
+ if (!tmp)
+ goto out_err;
+
+ tmp = rv_create_file("monitoring_on", RV_MODE_WRITE, rv_root.root_dir, NULL,
+ &monitoring_on_fops);
+ if (!tmp)
+ goto out_err;
+
+ monitoring_on = true;
+
+ return 0;
+
+out_err:
+ rv_remove(rv_root.root_dir);
+ printk(KERN_ERR "RV: Error while creating the RV interface\n");
+ return 1;
+}
diff --git a/kernel/trace/rv/rv.h b/kernel/trace/rv/rv.h
new file mode 100644
index 000000000000..2631ada61ebc
--- /dev/null
+++ b/kernel/trace/rv/rv.h
@@ -0,0 +1,34 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#include <linux/mutex.h>
+
+struct rv_interface {
+ struct dentry *root_dir;
+ struct dentry *monitors_dir;
+};
+
+#include "../trace.h"
+#include <linux/tracefs.h>
+#include <linux/rv.h>
+
+#define RV_MODE_WRITE TRACE_MODE_WRITE
+#define RV_MODE_READ TRACE_MODE_READ
+
+#define rv_create_dir tracefs_create_dir
+#define rv_create_file tracefs_create_file
+#define rv_remove tracefs_remove
+
+#define MAX_RV_MONITOR_NAME_SIZE 32
+
+extern struct mutex rv_interface_lock;
+
+struct rv_monitor_def {
+ struct list_head list;
+ struct rv_monitor *monitor;
+ struct dentry *root_d;
+ bool task_monitor;
+};
+
+struct dentry *get_monitors_root(void);
+int init_rv_monitors(struct dentry *root_dir);
+int get_task_monitor_slot(void);
+void put_task_monitor_slot(int slot);
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index b8dd54627075..062423371741 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -9772,6 +9772,8 @@ static __init int tracer_init_tracefs(void)
tracer_init_tracefs_work_func(NULL);
}
+ rv_init_interface();
+
return 0;
}
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index ff816fb41e48..900e75d96c84 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -2005,4 +2005,13 @@ struct trace_min_max_param {
extern const struct file_operations trace_min_max_fops;
+#ifdef CONFIG_RV
+extern int rv_init_interface(void);
+#else
+static inline int rv_init_interface(void)
+{
+ return 0;
+}
+#endif
+
#endif /* _LINUX_KERNEL_TRACE_H */
--
2.35.1
dot2c is a tool that transforms an automata in the graphiviz .dot file
into an C representation of the automata.
usage: dot2c [-h] dot_file
dot2c: converts a .dot file into a C structure
positional arguments:
dot_file The dot file to be converted
optional arguments:
-h, --help show this help message and exit
Cc: Wim Van Sebroeck <[email protected]>
Cc: Guenter Roeck <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Catalin Marinas <[email protected]>
Cc: Marco Elver <[email protected]>
Cc: Dmitry Vyukov <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Cc: Shuah Khan <[email protected]>
Cc: Gabriele Paoloni <[email protected]>
Cc: Juri Lelli <[email protected]>
Cc: Clark Williams <[email protected]>
Cc: Tao Zhou <[email protected]>
Cc: Randy Dunlap <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Daniel Bristot de Oliveira <[email protected]>
---
tools/verification/dot2/Makefile | 21 +++
tools/verification/dot2/automata.py | 171 +++++++++++++++++++
tools/verification/dot2/dot2c | 23 +++
tools/verification/dot2/dot2c.py | 251 ++++++++++++++++++++++++++++
4 files changed, 466 insertions(+)
create mode 100644 tools/verification/dot2/Makefile
create mode 100644 tools/verification/dot2/automata.py
create mode 100644 tools/verification/dot2/dot2c
create mode 100644 tools/verification/dot2/dot2c.py
diff --git a/tools/verification/dot2/Makefile b/tools/verification/dot2/Makefile
new file mode 100644
index 000000000000..235d182f6b2c
--- /dev/null
+++ b/tools/verification/dot2/Makefile
@@ -0,0 +1,21 @@
+INSTALL=install
+
+prefix ?= /usr
+bindir ?= $(prefix)/bin
+mandir ?= $(prefix)/share/man
+miscdir ?= $(prefix)/share/dot2
+srcdir ?= $(prefix)/src
+
+PYLIB ?= $(shell python3 -c 'import sysconfig; print (sysconfig.get_path("purelib"))')
+
+.PHONY: all
+all:
+
+.PHONY: clean
+clean:
+
+.PHONY: install
+install:
+ $(INSTALL) automata.py -D -m 644 $(DESTDIR)$(PYLIB)/dot2/automata.py
+ $(INSTALL) dot2c.py -D -m 644 $(DESTDIR)$(PYLIB)/dot2/dot2c.py
+ $(INSTALL) dot2c -D -m 755 $(DESTDIR)$(bindir)/
diff --git a/tools/verification/dot2/automata.py b/tools/verification/dot2/automata.py
new file mode 100644
index 000000000000..f22e1dff19ce
--- /dev/null
+++ b/tools/verification/dot2/automata.py
@@ -0,0 +1,171 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Copyright (C) 2019-2022 Red Hat, Inc. Daniel Bristot de Oliveira <[email protected]>
+#
+# Automata object: parse an automata in dot file digraph format into a python object
+
+import ntpath
+
+class Automata:
+ """Automata class: Reads a dot file and part it as an automata.
+
+ Attributes:
+ dot_file: A dot file with an state_automaton definition.
+ """
+
+ invalid_state_str = "INVALID_STATE"
+
+ def __init__(self, file_path):
+ self.__dot_path = file_path
+ self.name = self.__get_model_name()
+ self.__dot_lines = self.__open_dot()
+ self.states, self.initial_state, self.final_states = self.__get_state_variables()
+ self.events = self.__get_event_variables()
+ self.function = self.__create_matrix()
+
+ def __get_model_name(self):
+ basename = ntpath.basename(self.__dot_path)
+ if basename.endswith(".dot") == False:
+ print("not a dot file")
+ raise Exception("not a dot file: %s" % self.__dot_path)
+
+ model_name = basename[0:-4]
+ if model_name.__len__() == 0:
+ raise Exception("not a dot file: %s" % self.__dot_path)
+
+ return model_name
+
+ def __open_dot(self):
+ cursor = 0
+ dot_lines = []
+ try:
+ dot_file = open(self.__dot_path)
+ except:
+ raise Exception("Cannot open the file: %s" % self.__dot_path)
+
+ dot_lines = dot_file.read().splitlines()
+ dot_file.close()
+
+ # checking the first line:
+ line = dot_lines[cursor].split()
+
+ if (line[0] != "digraph") and (line[1] != "state_automaton"):
+ raise Exception("Not a valid .dot format: %s" % self.__dot_path)
+ else:
+ cursor += 1
+ return dot_lines
+
+ def __get_cursor_begin_states(self):
+ cursor = 0
+ while self.__dot_lines[cursor].split()[0] != "{node":
+ cursor += 1
+ return cursor
+
+ def __get_cursor_begin_events(self):
+ cursor = 0
+ while self.__dot_lines[cursor].split()[0] != "{node":
+ cursor += 1
+ while self.__dot_lines[cursor].split()[0] == "{node":
+ cursor += 1
+ # skip initial state transition
+ cursor += 1
+ return cursor
+
+ def __get_state_variables(self):
+ # wait for node declaration
+ states = []
+ final_states = []
+
+ has_final_states = False
+ cursor = self.__get_cursor_begin_states()
+
+ # process nodes
+ while self.__dot_lines[cursor].split()[0] == "{node":
+ line = self.__dot_lines[cursor].split()
+ raw_state = line[-1]
+
+ # "enabled_fired"}; -> enabled_fired
+ state = raw_state.replace('"', '').replace('};', '').replace(',','_')
+ if state[0:7] == "__init_":
+ initial_state = state[7:]
+ else:
+ states.append(state)
+ if self.__dot_lines[cursor].__contains__("doublecircle") == True:
+ final_states.append(state)
+ has_final_states = True
+
+ if self.__dot_lines[cursor].__contains__("ellipse") == True:
+ final_states.append(state)
+ has_final_states = True
+
+ cursor += 1
+
+ states = sorted(set(states))
+ states.remove(initial_state)
+
+ # Insert the initial state at the bein og the states
+ states.insert(0, initial_state)
+
+ if has_final_states == False:
+ final_states.append(initial_state)
+
+ return states, initial_state, final_states
+
+ def __get_event_variables(self):
+ # here we are at the begin of transitions, take a note, we will return later.
+ cursor = self.__get_cursor_begin_events()
+
+ events = []
+ while self.__dot_lines[cursor][1] == '"':
+ # transitions have the format:
+ # "all_fired" -> "both_fired" [ label = "disable_irq" ];
+ # ------------ event is here ------------^^^^^
+ if self.__dot_lines[cursor].split()[1] == "->":
+ line = self.__dot_lines[cursor].split()
+ event = line[-2].replace('"','')
+
+ # when a transition has more than one lables, they are like this
+ # "local_irq_enable\nhw_local_irq_enable_n"
+ # so split them.
+
+ event = event.replace("\\n", " ")
+ for i in event.split():
+ events.append(i)
+ cursor += 1
+
+ return sorted(set(events))
+
+ def __create_matrix(self):
+ # transform the array into a dictionary
+ events = self.events
+ states = self.states
+ events_dict = {}
+ states_dict = {}
+ nr_event = 0
+ for event in events:
+ events_dict[event] = nr_event
+ nr_event += 1
+
+ nr_state = 0
+ for state in states:
+ states_dict[state] = nr_state
+ nr_state += 1
+
+ # declare the matrix....
+ matrix = [[ self.invalid_state_str for x in range(nr_event)] for y in range(nr_state)]
+
+ # and we are back! Let's fill the matrix
+ cursor = self.__get_cursor_begin_events()
+
+ while self.__dot_lines[cursor][1] == '"':
+ if self.__dot_lines[cursor].split()[1] == "->":
+ line = self.__dot_lines[cursor].split()
+ origin_state = line[0].replace('"','').replace(',','_')
+ dest_state = line[2].replace('"','').replace(',','_')
+ possible_events = line[-2].replace('"','').replace("\\n", " ")
+ for event in possible_events.split():
+ matrix[states_dict[origin_state]][events_dict[event]] = dest_state
+ cursor += 1
+
+ return matrix
diff --git a/tools/verification/dot2/dot2c b/tools/verification/dot2/dot2c
new file mode 100644
index 000000000000..8a8cd84bdfcf
--- /dev/null
+++ b/tools/verification/dot2/dot2c
@@ -0,0 +1,23 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Copyright (C) 2019-2022 Red Hat, Inc. Daniel Bristot de Oliveira <[email protected]>
+#
+# dot2c: parse an automata in dot file digraph format into a C
+#
+# This program was written in the development of this paper:
+# de Oliveira, D. B. and Cucinotta, T. and de Oliveira, R. S.
+# "Efficient Formal Verification for the Linux Kernel." International
+# Conference on Software Engineering and Formal Methods. Springer, Cham, 2019.
+
+if __name__ == '__main__':
+ from dot2 import dot2c
+ import argparse
+ import sys
+
+ parser = argparse.ArgumentParser(description='dot2c: converts a .dot file into a C structure')
+ parser.add_argument('dot_file', help='The dot file to be converted')
+
+ args = parser.parse_args()
+ d = dot2c.Dot2c(args.dot_file)
+ d.print_model_classic()
diff --git a/tools/verification/dot2/dot2c.py b/tools/verification/dot2/dot2c.py
new file mode 100644
index 000000000000..bca902eec483
--- /dev/null
+++ b/tools/verification/dot2/dot2c.py
@@ -0,0 +1,251 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Copyright (C) 2019-2022 Red Hat, Inc. Daniel Bristot de Oliveira <[email protected]>
+#
+# dot2c: parse an automata in dot file digraph format into a C
+#
+# This program was written in the development of this paper:
+# de Oliveira, D. B. and Cucinotta, T. and de Oliveira, R. S.
+# "Efficient Formal Verification for the Linux Kernel." International
+# Conference on Software Engineering and Formal Methods. Springer, Cham, 2019.
+
+from dot2.automata import Automata
+
+class Dot2c(Automata):
+ enum_suffix = ""
+ enum_states_def = "states"
+ enum_events_def = "events"
+ struct_automaton_def = "automaton"
+ var_automaton_def = "aut"
+
+ def __init__(self, file_path):
+ super().__init__(file_path)
+ self.line_length = 100
+
+ def __buff_to_string(self, buff):
+ string = ""
+
+ for line in buff:
+ string = string + line + "\n"
+
+ # cut off the last \n
+ return string[:-1]
+
+ def __get_enum_states_content(self):
+ buff = []
+ buff.append("\t%s%s = 0," % (self.initial_state, self.enum_suffix))
+ for state in self.states:
+ if state != self.initial_state:
+ buff.append("\t%s%s," % (state, self.enum_suffix))
+ buff.append("\tstate_max%s" % (self.enum_suffix))
+
+ return buff
+
+ def get_enum_states_string(self):
+ buff = self.__get_enum_states_content()
+ return self.__buff_to_string(buff)
+
+ def format_states_enum(self):
+ buff = []
+ buff.append("enum %s {" % self.enum_states_def)
+ buff.append(self.get_enum_states_string())
+ buff.append("};\n")
+
+ return buff
+
+ def __get_enum_events_content(self):
+ buff = []
+ first = True
+ for event in self.events:
+ if first:
+ buff.append("\t%s%s = 0," % (event, self.enum_suffix))
+ first = False
+ else:
+ buff.append("\t%s%s," % (event, self.enum_suffix))
+
+ buff.append("\tevent_max%s" % self.enum_suffix)
+
+ return buff
+
+ def get_enum_events_string(self):
+ buff = self.__get_enum_events_content()
+ return self.__buff_to_string(buff)
+
+ def format_events_enum(self):
+ buff = []
+ buff.append("enum %s {" % self.enum_events_def)
+ buff.append(self.get_enum_events_string())
+ buff.append("};\n")
+
+ return buff
+
+ def get_minimun_type(self):
+ min_type = "unsigned char"
+
+ if self.states.__len__() > 255:
+ min_type = "unsigned short"
+
+ if self.states.__len__() > 65535:
+ min_type = "unsigned int"
+
+ if self.states.__len__() > 1000000:
+ raise Exception("Too many states: %d" % self.states.__len__())
+
+ return min_type
+
+ def format_automaton_definition(self):
+ min_type = self.get_minimun_type()
+ buff = []
+ buff.append("struct %s {" % self.struct_automaton_def)
+ buff.append("\tchar *state_names[state_max%s];" % (self.enum_suffix))
+ buff.append("\tchar *event_names[event_max%s];" % (self.enum_suffix))
+ buff.append("\t%s function[state_max%s][event_max%s];" % (min_type, self.enum_suffix, self.enum_suffix))
+ buff.append("\t%s initial_state;" % min_type)
+ buff.append("\tbool final_states[state_max%s];" % (self.enum_suffix))
+ buff.append("};\n")
+ return buff
+
+ def format_aut_init_header(self):
+ buff = []
+ buff.append("struct %s %s = {" % (self.struct_automaton_def, self.var_automaton_def))
+ return buff
+
+ def __get_string_vector_per_line_content(self, buff):
+ first = True
+ string = ""
+ for entry in buff:
+ if first:
+ string = string + "\t\t\"" + entry
+ first = False;
+ else:
+ string = string + "\",\n\t\t\"" + entry
+ string = string + "\""
+
+ return string
+
+ def get_aut_init_events_string(self):
+ return self.__get_string_vector_per_line_content(self.events)
+
+ def get_aut_init_states_string(self):
+ return self.__get_string_vector_per_line_content(self.states)
+
+ def format_aut_init_events_string(self):
+ buff = []
+ buff.append("\t.event_names = {")
+ buff.append(self.get_aut_init_events_string())
+ buff.append("\t},")
+ return buff
+
+ def format_aut_init_states_string(self):
+ buff = []
+ buff.append("\t.state_names = {")
+ buff.append(self.get_aut_init_states_string())
+ buff.append("\t},")
+
+ return buff
+
+ def __get_max_strlen_of_states(self):
+ max_state_name = max(self.states, key = len).__len__()
+ return max(max_state_name, self.invalid_state_str.__len__())
+
+ def __get_state_string_length(self):
+ maxlen = self.__get_max_strlen_of_states() + self.enum_suffix.__len__()
+ return "%" + str(maxlen) + "s"
+
+ def get_aut_init_function(self):
+ nr_states = self.states.__len__()
+ nr_events = self.events.__len__()
+ buff = []
+
+ strformat = self.__get_state_string_length()
+
+ for x in range(nr_states):
+ line = "\t\t{ "
+ for y in range(nr_events):
+ next_state = self.function[x][y]
+ if next_state != self.invalid_state_str:
+ next_state = self.function[x][y] + self.enum_suffix
+
+ if y != nr_events-1:
+ line = line + strformat % next_state + ", "
+ else:
+ line = line + strformat % next_state + " },"
+ buff.append(line)
+
+ return self.__buff_to_string(buff)
+
+ def format_aut_init_function(self):
+ buff = []
+ buff.append("\t.function = {")
+ buff.append(self.get_aut_init_function())
+ buff.append("\t},")
+
+ return buff
+
+ def get_aut_init_initial_state(self):
+ return self.initial_state
+
+ def format_aut_init_initial_state(self):
+ buff = []
+ initial_state = self.get_aut_init_initial_state()
+ buff.append("\t.initial_state = " + initial_state + self.enum_suffix + ",")
+
+ return buff
+
+ def get_aut_init_final_states(self):
+ line = ""
+ first = True
+ for state in self.states:
+ if first == False:
+ line = line + ', '
+ else:
+ first = False
+
+ if self.final_states.__contains__(state):
+ line = line + '1'
+ else:
+ line = line + '0'
+ return line
+
+ def format_aut_init_final_states(self):
+ buff = []
+ buff.append("\t.final_states = { %s }," % self.get_aut_init_final_states())
+
+ return buff
+
+ def __get_automaton_initialization_footer_string(self):
+ footer = "};\n"
+ return footer
+
+ def format_aut_init_footer(self):
+ buff = []
+ buff.append(self.__get_automaton_initialization_footer_string())
+
+ return buff
+
+ def format_invalid_state(self):
+ buff = []
+ buff.append("#define %s state_max%s\n" % (self.invalid_state_str, self.enum_suffix))
+
+ return buff
+
+ def format_model(self):
+ buff = []
+ buff += self.format_states_enum()
+ buff += self.format_invalid_state()
+ buff += self.format_events_enum()
+ buff += self.format_automaton_definition()
+ buff += self.format_aut_init_header()
+ buff += self.format_aut_init_states_string()
+ buff += self.format_aut_init_events_string()
+ buff += self.format_aut_init_function()
+ buff += self.format_aut_init_initial_state()
+ buff += self.format_aut_init_final_states()
+ buff += self.format_aut_init_footer()
+
+ return buff
+
+ def print_model_classic(self):
+ buff = self.format_model()
+ print(self.__buff_to_string(buff))
--
2.35.1
On Tue, 19 Jul 2022 19:27:06 +0200
Daniel Bristot de Oliveira <[email protected]> wrote:
> +/*
> + * Interface to read the enable/disable status of a monitor.
> + */
> +static ssize_t monitor_desc_read_data(struct file *filp, char __user *user_buf, size_t count,
> + loff_t *ppos)
> +{
> + struct rv_monitor_def *mdef = filp->private_data;
> + char buff[256];
> +
> + memset(buff, 0, sizeof(buff));
> +
> + mutex_lock(&rv_interface_lock);
> + snprintf(buff, sizeof(buff), "%s\n", mdef->monitor->description);
> + mutex_unlock(&rv_interface_lock);
What exactly is the mutex protecting here?
-- Steve
> +
> + return simple_read_from_buffer(user_buf, count, ppos, buff, strlen(buff) + 1);
> +}
> +
On 7/20/22 16:47, Steven Rostedt wrote:
> On Tue, 19 Jul 2022 19:27:06 +0200
> Daniel Bristot de Oliveira <[email protected]> wrote:
>
>> +/*
>> + * Interface to read the enable/disable status of a monitor.
>> + */
>> +static ssize_t monitor_desc_read_data(struct file *filp, char __user *user_buf, size_t count,
>> + loff_t *ppos)
>> +{
>> + struct rv_monitor_def *mdef = filp->private_data;
>> + char buff[256];
>> +
>> + memset(buff, 0, sizeof(buff));
>> +
>> + mutex_lock(&rv_interface_lock);
>> + snprintf(buff, sizeof(buff), "%s\n", mdef->monitor->description);
>> + mutex_unlock(&rv_interface_lock);
>
> What exactly is the mutex protecting here?
I placed it around all access on read/write in the interface,
but indeed, it is not needed here (static data). It is in a file inside
the monitor's directory... so the directory is there, and the
module is there as well.
-- Daniel
On 7/20/22 17:21, Steven Rostedt wrote:
>> +static void disable_all_monitors(void)
>> +{
>> + struct rv_monitor_def *mdef;
>> +
> I think you need:
>
> mutex_lock(&rv_interface_lock);
>
>> + list_for_each_entry(mdef, &rv_monitors_list, list)
>> + disable_monitor(mdef);
> mutex_unlock(&rv_interface_lock);
>
>> +}
yes, we need.
-- Daniel
On Tue, 19 Jul 2022 19:27:06 +0200
Daniel Bristot de Oliveira <[email protected]> wrote:
> +/*
> + * enabled_monitors interface.
> + */
Doesn't the rv_monitors_list need to be protected by the rv_interface_lock
mutex?
> +static void disable_all_monitors(void)
> +{
> + struct rv_monitor_def *mdef;
> +
I think you need:
mutex_lock(&rv_interface_lock);
> + list_for_each_entry(mdef, &rv_monitors_list, list)
> + disable_monitor(mdef);
mutex_unlock(&rv_interface_lock);
> +}
> +
> +static int enabled_monitors_open(struct inode *inode, struct file *file)
> +{
As this gets called directly from the open system call.
> + if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC))
> + disable_all_monitors();
> +
> + return seq_open(file, &enabled_monitors_seq_ops);
> +};
> +
[..]
> +
> +static const struct file_operations enabled_monitors_ops = {
> + .open = enabled_monitors_open,
Here.
-- Steve
> + .read = seq_read,
> + .write = enabled_monitors_write,
> + .llseek = seq_lseek,
> + .release = seq_release,
> +};
On Tue, 19 Jul 2022 19:27:09 +0200
Daniel Bristot de Oliveira <[email protected]> wrote:
> diff --git a/include/linux/rv.h b/include/linux/rv.h
> index 4f5b70eee557..31d8b2614eae 100644
> --- a/include/linux/rv.h
> +++ b/include/linux/rv.h
> @@ -7,6 +7,8 @@
> #ifndef _LINUX_RV_H
> #define _LINUX_RV_H
>
> +#define MAX_DA_NAME_LEN 24
> +
> struct rv_reactor {
> char *name;
> char *description;
> diff --git a/include/rv/da_monitor.h b/include/rv/da_monitor.h
> new file mode 100644
> index 000000000000..ef7ee3ffcad6
> --- /dev/null
> +++ b/include/rv/da_monitor.h
> @@ -0,0 +1,507 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + * Copyright (C) 2019-2022 Red Hat, Inc. Daniel Bristot de Oliveira <[email protected]>
> + *
> + * Deterministic automata (DA) monitor functions, to be used together
> + * with automata models in C generated by the dot2k tool.
> + *
> + * The dot2k tool is available at tools/verification/dot2k/
> + */
> +
> +#include <rv/automata.h>
> +#include <linux/rv.h>
> +#include <linux/bug.h>
> +
> +/*
> + * Generic helpers for all types of deterministic automata monitors.
> + */
> +#define DECLARE_DA_MON_GENERIC_HELPERS(name, type) \
> + \
> +static char REACT_MSG[1024]; \
> + \
> +static inline char *format_react_msg(type curr_state, type event) \
You probably want to call this format_react_msg_##name() too.
> +{ \
> + snprintf(REACT_MSG, 1024, \
> + "rv: monitor %s does not allow event %s on state %s\n", \
> + #name, \
> + model_get_event_name_##name(event), \
> + model_get_state_name_##name(curr_state)); \
> + return REACT_MSG; \
> +} \
> + \
> +static void cond_react(char *msg) \
And this cond_react_##name() as well. Otherwise you can have issues with
the same function being used by multiple monitors. What if two are declared
in the same file? This will fail to build.
> +{ \
> + if (rv_##name.react) \
> + rv_##name.react(msg); \
> +} \
> + \
> +/* \
> + * da_monitor_reset_##name - reset a monitor and setting it to init state \
> + */ \
> +static inline void da_monitor_reset_##name(struct da_monitor *da_mon) \
> +{ \
> + da_mon->monitoring = 0; \
> + da_mon->curr_state = model_get_initial_state_##name(); \
> +} \
> + \
> +/* \
> + * da_monitor_curr_state_##name - return the current state \
> + */ \
> +static inline type da_monitor_curr_state_##name(struct da_monitor *da_mon) \
> +{ \
> + return da_mon->curr_state; \
> +} \
> + \
> +/* \
> + * da_monitor_set_state_##name - set the new current state \
> + */ \
> +static inline void \
> +da_monitor_set_state_##name(struct da_monitor *da_mon, enum states_##name state) \
> +{ \
> + da_mon->curr_state = state; \
> +} \
> + \
> +/* \
> + * da_monitor_start_##name - start monitoring \
> + * \
> + * The monitor will ignore all events until monitoring is set to true. This \
> + * function needs to be called to tell the monitor to start monitoring. \
> + */ \
> +static inline void da_monitor_start_##name(struct da_monitor *da_mon) \
> +{ \
> + da_mon->monitoring = 1; \
> +} \
> + \
> +/* \
> + * da_monitoring_##name - returns true if the monitor is processing events \
> + */ \
> +static inline bool da_monitoring_##name(struct da_monitor *da_mon) \
> +{ \
> + return da_mon->monitoring; \
> +} \
> + \
> +/* \
> + * da_monitor_enabled_##name - checks if the monitor is enabled \
> + */ \
> +static inline bool da_monitor_enabled_##name(void) \
> +{ \
Should we add a:
smp_rmb();
here? And then a smp_wmb() where these switches get updated?
I guess how critical is it that these turn off immediately after the switch
is flipped?
> + /* global switch */ \
> + if (unlikely(!rv_monitoring_on())) \
> + return 0; \
> + \
> + /* monitor enabled */ \
> + if (unlikely(!rv_##name.enabled)) \
> + return 0; \
> + \
> + return 1; \
> +} \
> + \
> +/* \
> + * da_monitor_handling_event_##name - checks if the monitor is ready to handle events \
> + */ \
> +static inline bool da_monitor_handling_event_##name(struct da_monitor *da_mon) \
> +{ \
> + \
> + if (!da_monitor_enabled_##name()) \
> + return 0; \
> + \
> + /* monitor is actually monitoring */ \
> + if (unlikely(!da_monitoring_##name(da_mon))) \
> + return 0; \
> + \
> + return 1; \
> +}
> diff --git a/kernel/trace/rv/Kconfig b/kernel/trace/rv/Kconfig
> index 3eb5d48ab4f6..0123bdf7052a 100644
> --- a/kernel/trace/rv/Kconfig
> +++ b/kernel/trace/rv/Kconfig
> @@ -1,5 +1,19 @@
> # SPDX-License-Identifier: GPL-2.0-only
> #
> +config DA_MON_EVENTS
> + default n
> + bool
> +
> +config DA_MON_EVENTS_IMPLICIT
> + select DA_MON_EVENTS
> + default n
> + bool
> +
> +config DA_MON_EVENTS_ID
> + select DA_MON_EVENTS
> + default n
> + bool
The "default n" are not needed. The default is 'n' without it.
-- Steve
> +
> menuconfig RV
> bool "Runtime Verification"
> depends on TRACING
> diff --git a/kernel/trace/rv/rv.c b/kernel/trace/rv/rv.c
> index eb835777a59b..00183e056dfd 100644
> --- a/kernel/trace/rv/rv.c
> +++ b/kernel/trace/rv/rv.c
> @@ -141,6 +141,11 @@
> #include <linux/slab.h>
> #include <rv/rv.h>
>
> +#ifdef CONFIG_DA_MON_EVENTS
> +#define CREATE_TRACE_POINTS
> +#include <trace/events/rv.h>
> +#endif
> +
> #include "rv.h"
>
> DEFINE_MUTEX(rv_interface_lock);
On 7/20/22 22:06, Steven Rostedt wrote:
>> +/* \
>> + * da_monitor_enabled_##name - checks if the monitor is enabled \
>> + */ \
>> +static inline bool da_monitor_enabled_##name(void) \
>> +{ \
> Should we add a:
>
> smp_rmb();
>
> here? And then a smp_wmb() where these switches get updated?
>
Makes sense.
Should I also add the READ_ONCE/WRITE_ONCE? like
smp_rmb()
READ_ONCE(var)
WRITE_ONCE(var, value)
smp_wmb()
for all these on/off knobs, or just the barriers?
> I guess how critical is it that these turn off immediately after the switch
> is flipped?
It is not critical to continue the execution of those that have already crossed by
the variable. Still, waiting for the tracepoints to finish their execution before
returning to the user-space task that disabled the variable might be a good thing.
IIRC, we can do that via RCU... like, synchronize_rcu()?
>> + /* global switch */ \
>> + if (unlikely(!rv_monitoring_on())) \
>> + return 0; \
>> + \
>> + /* monitor enabled */ \
>> + if (unlikely(!rv_##name.enabled)) \
>> + return 0; \
>> + \
>> + return 1; \
>> +} \
>> + \
On Thu, 21 Jul 2022 14:08:38 +0200
Daniel Bristot de Oliveira <[email protected]> wrote:
> On 7/20/22 22:06, Steven Rostedt wrote:
> >> +/* \
> >> + * da_monitor_enabled_##name - checks if the monitor is enabled \
> >> + */ \
> >> +static inline bool da_monitor_enabled_##name(void) \
> >> +{ \
> > Should we add a:
> >
> > smp_rmb();
> >
> > here? And then a smp_wmb() where these switches get updated?
> >
>
> Makes sense.
>
> Should I also add the READ_ONCE/WRITE_ONCE? like
>
> smp_rmb()
> READ_ONCE(var)
>
> WRITE_ONCE(var, value)
> smp_wmb()
I'm not sure the WRITE_ONCE() is necessary with the memory barriers.
Because they should also prevent gcc from doing anything after that
barrier. As Linus once stated, most cases WRITE_ONCE() is useless, but it's
fine to keep more for annotation (as to pair with the READ_ONCE()) than for
anything that is critical.
>
> for all these on/off knobs, or just the barriers?
>
> > I guess how critical is it that these turn off immediately after the switch
> > is flipped?
>
> It is not critical to continue the execution of those that have already crossed by
> the variable. Still, waiting for the tracepoints to finish their execution before
> returning to the user-space task that disabled the variable might be a good thing.
You mean after disabling, to wait for the tracepoints that are currently
running to end?
>
> IIRC, we can do that via RCU... like, synchronize_rcu()?
We have tracepoint_synchronize_unregister() that does that, as some
traceponits use SRCU and not RCU.
-- Steve
>
> >> + /* global switch */ \
> >> + if (unlikely(!rv_monitoring_on())) \
> >> + return 0; \
> >> + \
> >> + /* monitor enabled */ \
> >> + if (unlikely(!rv_##name.enabled)) \
> >> + return 0; \
> >> + \
> >> + return 1; \
> >> +} \
> >> + \
On Tue, Jul 19, 2022 at 07:27:16PM +0200, Daniel Bristot de Oliveira wrote:
> +Looking at the automata definition, it is possible to see that the system
> +and the model are expected to return to the initial state after the
> +preempt_enable execution. Hence, it can be used to synchronize the
> +system and the model at the initialization of the monitoring section.
> +
> +The start is informed via a special handle function, the
> +"da_handle_start_event_MONITOR_event)", in this case::
da_handle_start_event_$MONITOR(event) the same as the previous version
should be right.
> +
> + da_handle_start_event_wip(preempt_disable_wip);
I didn't see this in last version that the event should be preempt_enable_wip.
> +
> +So, the callback function will look like::
> +
> + void handle_preempt_enable(void *data, unsigned long ip, unsigned long parent_ip)
> + {
> + da_handle_start_event_wip(preempt_enable_wip);
> + }
> +
> +Finally, the "handle_sched_waking()" will look like::
> +
> + void handle_sched_waking(void *data, struct task_struct *task)
> + {
> + da_handle_event_wip(sched_waking_wip);
> + }
> +
> +And the explanation is left for the reader as an exercise.
> +
> +Start and Stop functions
Start and Stop functions are changed to enable and disable functions.
> +------------------------
> +
> +dot2k automatically creates two special functions::
> +
> + enable_MONITOR()
> + disable_MONITOR()
> +
> +These functions are called when the monitor is enabled and disabled,
> +respectively.
> +
> +They should be used to *attach* and *detach* the instrumentation to the running
> +system. The developer must add to the relative function all that is needed to
> +*attach* and *detach* its monitor to the system.
> +
> +For the wip case, these functions were named::
> +
> + enable_wip()
> + disable_wip()
> +
> +But no change was required because: by default, these functions *attach* and
> +*detach* the tracepoints_to_attach, which was enough for this case.
On Thu, 21 Jul 2022 22:38:25 +0800
Tao Zhou <[email protected]> wrote:
> > +static int enable_monitor(struct rv_monitor_def *mdef)
> > +{
> > + int retval;
> > +
> > + if (!mdef->monitor->enabled) {
> > + retval = mdef->monitor->enable();
> > + if (retval)
> > + return retval;
> > + }
> > +
> > + mdef->monitor->enabled = 1;
>
> This should be placed at the end of the last if block. Otherwise
> another assignment may be duplicated because it is already 1 now.
> no?(not sure how compiler treat this..)
It really doesn't matter, it will just sent enabled to one even though it's
already one.
You could simplify this to be:
static int enable_monitor(struct rv_monitor_def *mdef)
{
int retval;
if (mdef->monitor->enabled)
return 0;
retval = mdef->monitor->enable();
if (!retval)
mdef->monitor->enabled = 1;
return retval;
}
-- Steve
On Tue, Jul 19, 2022 at 07:27:06PM +0200, Daniel Bristot de Oliveira wrote:
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/slab.h>
> +#include <rv/rv.h>
> +
> +#include "rv.h"
> +
> +DEFINE_MUTEX(rv_interface_lock);
> +/*
> + * Enable a given monitor.
> + */
> +static int enable_monitor(struct rv_monitor_def *mdef)
> +{
> + int retval;
> +
> + if (!mdef->monitor->enabled) {
> + retval = mdef->monitor->enable();
> + if (retval)
> + return retval;
> + }
> +
> + mdef->monitor->enabled = 1;
This should be placed at the end of the last if block. Otherwise
another assignment may be duplicated because it is already 1 now.
no?(not sure how compiler treat this..)
> +
> +/*
> + * Interface to read the enable/disable status of a monitor.
I see the 'status' but look a little more this function is about the
description of monitor, so:
/* Interface to read the description message of a monitor */
or others.
> + */
> +static ssize_t monitor_desc_read_data(struct file *filp, char __user *user_buf, size_t count,
> + loff_t *ppos)
> +{
> + struct rv_monitor_def *mdef = filp->private_data;
> + char buff[256];
> +
> + memset(buff, 0, sizeof(buff));
> +
> + mutex_lock(&rv_interface_lock);
> + snprintf(buff, sizeof(buff), "%s\n", mdef->monitor->description);
> + mutex_unlock(&rv_interface_lock);
On 7/21/22 15:59, Steven Rostedt wrote:
> On Thu, 21 Jul 2022 14:08:38 +0200
> Daniel Bristot de Oliveira <[email protected]> wrote:
>
>> On 7/20/22 22:06, Steven Rostedt wrote:
>>>> +/* \
>>>> + * da_monitor_enabled_##name - checks if the monitor is enabled \
>>>> + */ \
>>>> +static inline bool da_monitor_enabled_##name(void) \
>>>> +{ \
>>> Should we add a:
>>>
>>> smp_rmb();
>>>
>>> here? And then a smp_wmb() where these switches get updated?
>>>
>>
>> Makes sense.
>>
>> Should I also add the READ_ONCE/WRITE_ONCE? like
>>
>> smp_rmb()
>> READ_ONCE(var)
>>
>> WRITE_ONCE(var, value)
>> smp_wmb()
>
> I'm not sure the WRITE_ONCE() is necessary with the memory barriers.
> Because they should also prevent gcc from doing anything after that
> barrier. As Linus once stated, most cases WRITE_ONCE() is useless, but it's
> fine to keep more for annotation (as to pair with the READ_ONCE()) than for
> anything that is critical.
Ack, I can keep for annotation.
>>
>> for all these on/off knobs, or just the barriers?
>>
>>> I guess how critical is it that these turn off immediately after the switch
>>> is flipped?
>>
>> It is not critical to continue the execution of those that have already crossed by
>> the variable. Still, waiting for the tracepoints to finish their execution before
>> returning to the user-space task that disabled the variable might be a good thing.
>
> You mean after disabling, to wait for the tracepoints that are currently
> running to end?
yes, after disabling tracepoints.
>>
>> IIRC, we can do that via RCU... like, synchronize_rcu()?
>
> We have tracepoint_synchronize_unregister() that does that, as some
> traceponits use SRCU and not RCU.
yep, that is it!
-- Daniel
> -- Steve
>
>
>>
>>>> + /* global switch */ \
>>>> + if (unlikely(!rv_monitoring_on())) \
>>>> + return 0; \
>>>> + \
>>>> + /* monitor enabled */ \
>>>> + if (unlikely(!rv_##name.enabled)) \
>>>> + return 0; \
>>>> + \
>>>> + return 1; \
>>>> +} \
>>>> + \
>