- 09 Dec, 2016 6 commits
-
-
Steven Rostedt (Red Hat) authored
Instead of using get_user_pages_fast() and kmap_atomic() when writing to the trace_marker file, just allocate enough space on the ring buffer directly, and write into it via copy_from_user(). Writing into the trace_marker file use to allocate a temporary buffer to perform the copy_from_user(), as we didn't want to write into the ring buffer if the copy failed. But as a trace_marker write is suppose to be extremely fast, and allocating memory causes other tracepoints to trigger, Peter Zijlstra suggested using get_user_pages_fast() and kmap_atomic() to keep the user space pages in memory and reading it directly. But Henrik Austad had issues with this because it required taking the mm->mmap_sem and causing long delays with the write. Instead, just allocate the space in the ring buffer and use copy_from_user() directly. If it faults, return -EFAULT and write "<faulted>" into the ring buffer. Link: http://lkml.kernel.org/r/20161208124018.72dd0f86@gandalf.local.home Cc: Ingo Molnar <mingo@kernel.org> Cc: Henrik Austad <henrik@austad.us> Cc: Peter Zijlstra <peterz@infradead.org> Updates: d696b58c "tracing: Do not allocate buffer for trace_marker" Suggested-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
Steven Rostedt <rostedt@goodmis.org>
-
Steven Rostedt (Red Hat) authored
With new binutils, gcc may get smart with its optimization and change a jmp from a 5 byte jump to a 2 byte one even though it was jumping to a global function. But that global function existed within a 2 byte radius, and gcc was able to optimize it. Unfortunately, that jump was also being modified when function graph tracing begins. Since ftrace expected that jump to be 5 bytes, but it was only two, it overwrote code after the jump, causing a crash. This was fixed for x86_64 with commit 8329e818, with the same subject as this commit, but nothing was done for x86_32. Cc: stable@vger.kernel.org Fixes: d61f82d0 ("ftrace: use dynamic patching for updating mcount calls") Reported-by:
Colin Ian King <colin.king@canonical.com> Tested-by:
-
Steven Rostedt (Red Hat) authored
The trace event start up selftests fails when the trace benchmark is enabled, because it is disabled during boot. It really only needs to be disabled before scheduling is set up, as it creates a thread. Signed-off-by: -
Steven Rostedt (Red Hat) authored
If one of the events within a system fails to enable when "1" is written to the system "enable" file, it should return an error. Note, some events may still be enabled, but the user should know that something did go wrong. Signed-off-by: -
Steven Rostedt (Red Hat) authored
Trace events are enabled very early on boot up via the boot command line parameter. The benchmark tool creates a new thread to perform the trace event benchmarking. But at start up, it is called before scheduling is set up and because it creates a new thread before the init thread is created, this crashes the kernel. Have the benchmark fail to register when started via the kernel command line. Also, since the registering of a tracepoint now can handle failure cases, return -ENOMEM instead of warning if the thread cannot be created. Signed-off-by: -
Steven Rostedt (Red Hat) authored
Some tracepoints have a registration function that gets enabled when the tracepoint is enabled. There may be cases that the registraction function must fail (for example, can't allocate enough memory). In this case, the tracepoint should also fail to register, otherwise the user would not know why the tracepoint is not working. Cc: David Howells <dhowells@redhat.com> Cc: Seiji Aguchi <seiji.aguchi@hds.com> Cc: Anton Blanchard <anton@samba.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Signed-off-by:
-
- 24 Nov, 2016 4 commits
-
-
Steven Rostedt (Red Hat) authored
Both rb_end_commit() and rb_set_commit_to_write() are in the fast path of the ring buffer recording. Make sure they are always inlined. Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.orgReported-by:
Andi Kleen <andi@firstfloor.org> Signed-off-by:
-
Steven Rostedt (Red Hat) authored
The function rb_update_write_stamp() is in the hotpath of the ring buffer recording. Make sure that it is inlined as well. There's not many places that call it. Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.orgReported-by:
-
Steven Rostedt (Red Hat) authored
There's several small helper functions in ring_buffer.c that are used in the hot path. For some reason, even though they are marked inline, gcc tends not to enforce it. Make sure these functions are always inlined. Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.orgReported-by:
-
Steven Rostedt (Red Hat) authored
The function __buffer_unlock_commit() is called in a few places outside of trace.c. But for the most part, it should really be inlined, as it is in the hot path of the trace_events. For the callers outside of trace.c, create a new function trace_buffer_unlock_commit_nostack(), as the reason it was used was to avoid the stack tracing that trace_buffer_unlock_commit() could do. Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.orgReported-by:
-
- 23 Nov, 2016 4 commits
-
-
Steven Rostedt (Red Hat) authored
Currently, when tracepoint_printk is set (enabled by the "tp_printk" kernel command line), it causes trace events to print via printk(). This is a very dangerous operation, but is useful for debugging. The issue is, it's seldom used, but it is always checked even if it's not enabled by the kernel command line. Instead of having this feature called by a branch against a variable, turn that variable into a static key, and this will remove the test and jump. To simplify things, the functions output_printk() and trace_event_buffer_commit() were moved from trace_events.c to trace.c. Signed-off-by: -
Steven Rostedt (Red Hat) authored
The rb_event_data() is the fast path of getting the ring buffer data from an event. Externally, ring_buffer_event_data() is used to access this function. But unfortunately, rb_event_data() is not inlined, and calling ring_buffer_event_data() causes that function to be called again. Force rb_event_data() to be inlined to lower the number of operations needed when calling ring_buffer_event_data(). Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.orgReported-by:
-
Steven Rostedt (Red Hat) authored
The function rb_reserved_next_event() is called by two functions: ring_buffer_lock_reserve() and ring_buffer_write(). This is in a very hot path of the tracing code, and it is best that they are not functions. The two callers are basically wrapers for rb_reserver_next_event(). Removing the function calls can save execution time in the hotpath of tracing. Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.orgReported-by:
-
Steven Rostedt (Red Hat) authored
As Andi Kleen pointed out in the Link below, the trace events has quite a bit of code execution. A lot of that happens to be calling functions, where some of them should simply be inlined. One of these functions happens to be trace_buffer_lock_reserve() which is also a global, but it is used throughout the file it is defined in. Create a __trace_buffer_lock_reserve() that is always inlined that the file can benefit from. Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.orgReported-by:
-
- 22 Nov, 2016 15 commits
-
-
Steven Rostedt (Red Hat) authored
The creation of the set_event_pid file was assigned to a variable "entry" but that variable was never used. Ideally, it should be used to check if the file was created and warn if it was not. The files header_page, header_event should also be checked and a warning if they fail to be created. The "enable" file was moved up, as it is a more crucial file to have and a hard failure (return -ENOMEM) should be returned if it is not created. Reported-by:
David Binderman <dcb314@hotmail.com> Signed-off-by:
-
Chunyan Zhang authored
If CONFIG_STM_SOURCE_FTRACE is selected, Function trace data can be writen to sink via STM, all functions that related to writing data packets to STM should be marked 'notrace' to avoid being traced by Ftrace, otherwise the program would stall into an endless loop. Link: http://lkml.kernel.org/r/1479715043-6534-7-git-send-email-zhang.chunyan@linaro.orgSigned-off-by:
Chunyan Zhang <zhang.chunyan@linaro.org> Acked-by:
Alexander Shishkin <alexander.shishkin@linux.intel.com> Signed-off-by:
-
Chunyan Zhang authored
If CONFIG_STM_SOURCE_FTRACE is selected, Function trace data can be writen to sink via STM, all functions that related to writing data packets to STM should be marked 'notrace' to avoid being traced by Ftrace, otherwise the program would stall into an endless loop. Link: http://lkml.kernel.org/r/1479715043-6534-6-git-send-email-zhang.chunyan@linaro.orgSigned-off-by:
-
Chunyan Zhang authored
If CONFIG_STM_SOURCE_FTRACE is selected, Function trace data can be writen to sink via STM, all functions that related to writing data packets to STM should be marked 'notrace' to avoid being traced by Ftrace, otherwise the program would stall into an endless loop. Link: http://lkml.kernel.org/r/1479715043-6534-5-git-send-email-zhang.chunyan@linaro.orgSigned-off-by:
-
Chunyan Zhang authored
If CONFIG_STM_SOURCE_FTRACE is selected, Function trace data can be writen to sink via STM, all functions that related to writing data packets to STM should be marked 'notrace' to avoid being traced by Ftrace, otherwise the program would stall into an endless loop. Link: http://lkml.kernel.org/r/1479715043-6534-4-git-send-email-zhang.chunyan@linaro.orgSigned-off-by:
-
Chunyan Zhang authored
This patch adds a driver that models itself as an stm_source called stm_ftrace. Once the stm device and stm_ftrace have been linked via sysfs, the driver registers itself as a trace_export and everything passed to the interface from Ftrace subsystem will end up in the STM trace engine. Link: http://lkml.kernel.org/r/1479715043-6534-3-git-send-email-zhang.chunyan@linaro.orgSigned-off-by:
-
Chunyan Zhang authored
Currently Function traces can be only exported to the ring buffer. This adds a trace_export concept which can process traces and export them to a registered destination as an addition to the current one that outputs to Ftrace - i.e. ring buffer. In this way, if we want function traces to be sent to other destinations rather than only to the ring buffer, we just need to register a new trace_export and implement its own .write() function for writing traces to storage. With this patch, only function tracing (trace type is TRACE_FN) is supported. Link: http://lkml.kernel.org/r/1479715043-6534-2-git-send-email-zhang.chunyan@linaro.orgSigned-off-by:
-
Masami Hiramatsu authored
Add a testcase for types of kprobe event. This checks kprobe event can accept and correctly expressed the arguments typed as s32, u32, x32 and bitfield. Here is the test result. ----- # ./ftracetest test.d/kprobe/kprobe_args_type.tc === Ftrace unit tests === [1] Kprobes event arguments with types [PASS] # of passed: 1 # of failed: 0 # of unresolved: 0 # of untested: 0 # of unsupported: 0 # of xfailed: 0 # of undefined(test bug): 0 ----- Link: http://lkml.kernel.org/r/147928409063.22982.3499119203875115458.stgit@devboxSigned-off-by:
Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by:
-
Masami Hiramatsu authored
Add function filter glob matching test case. This checks whether the kernel supports glob matching (front match, end match, middle match, side match, character class and '?'). Here is the test result. ----- ./ftracetest test.d/ftrace/func-filter-glob.tc === Ftrace unit tests === [1] ftrace - function glob filters [PASS] # of passed: 1 # of failed: 0 # of unresolved: 0 # of untested: 0 # of unsupported: 0 # of xfailed: 0 # of undefined(test bug): 0 ----- Link: http://lkml.kernel.org/r/147928407589.22982.16364174511117104303.stgit@devboxSigned-off-by:
-
Masami Hiramatsu authored
Introduce TMPDIR variable which is removed after each test is done, so that the test script can put their temporary files in that. Link: http://lkml.kernel.org/r/147928406116.22982.8761924340108532378.stgit@devboxSigned-off-by:
-
Masami Hiramatsu authored
Hide ftracetest result log directory from git. Link: http://lkml.kernel.org/r/147928404640.22982.13173364949326289032.stgit@devboxSigned-off-by:
-
Masami Hiramatsu authored
Since histogram trigger id.syscall depends on CONFIG_FTRACE_SYSCALLS, a testcase in trigger-modifier test fails if that config is disabled. Fix this bug by using flexible pattern to check the histogram output. Link: http://lkml.kernel.org/r/147928402670.22982.15589445159052676877.stgit@devboxSigned-off-by:
-
Masami Hiramatsu authored
If "snapshot" special file doesn't exist, that kernel does not support snapshot and snapshot trigger too. In that case snapshot trigger test results to unsupported instead of fail. Link: http://lkml.kernel.org/r/147928401215.22982.10411665829041109794.stgit@devboxSigned-off-by:
-
Masami Hiramatsu authored
Since the verbose error logs scrolls out previous test results --quiet option suppress to show such message. e.g. # ./ftracetest -q === Ftrace unit tests === [1] Basic trace file check [PASS] [2] Basic test for tracers [PASS] [3] Basic trace clock test [PASS] [4] Basic event tracing check [PASS] [5] event tracing - enable/disable with event level files [PASS] [6] event tracing - restricts events based on pid [PASS] [7] event tracing - enable/disable with subsystem level files [PASS] [8] event tracing - enable/disable with top level files [PASS] [9] ftrace - function graph filters with stack tracer [UNSUPPORTED] [10] ftrace - function graph filters [UNSUPPORTED] [11] ftrace - function profiler with function tracing [UNSUPPORTED] [12] Test creation and deletion of trace instances while setting an event [PASS] [13] Test creation and deletion of trace instances [PASS] [14] Kprobe dynamic event - adding and removing [UNSUPPORTED] [15] Kprobe dynamic event - busy event check [UNSUPPORTED] [16] Kprobe dynamic event with arguments [UNSUPPORTED] [17] Kprobe dynamic event with function tracer [UNSUPPORTED] [18] Kretprobe dynamic event with arguments [UNSUPPORTED] [19] event trigger - test event enable/disable trigger [PASS] [20] event trigger - test trigger filter [PASS] [21] event trigger - test histogram modifiers [UNSUPPORTED] [22] event trigger - test histogram trigger [UNSUPPORTED] [23] event trigger - test multiple histogram triggers [UNSUPPORTED] [24] event trigger - test snapshot-trigger [FAIL] [25] event trigger - test stacktrace-trigger [PASS] [26] event trigger - test traceon/off trigger [PASS] # of passed: 14 # of failed: 1 # of unresolved: 0 # of untested: 0 # of unsupported: 11 # of xfailed: 0 # of undefined(test bug): 0 Link: http://lkml.kernel.org/r/147928399712.22982.8284640390982775052.stgit@devboxSigned-off-by:
-
Masami Hiramatsu authored
Reset ftrace to initial state before running each test. This fixes some test cases to enable tracing before starting trace test. This can avoid false-positive failure when previous testcase fails while disabling tracing. Link: http://lkml.kernel.org/r/147928398192.22982.7767460638302113002.stgit@devboxSigned-off-by:
-
- 15 Nov, 2016 1 commit
-
-
Steven Rostedt authored
A new file is created: /sys/kernel/debug/tracing/trace_marker_raw This allows for appications to create data structures and write the binary data directly into it, and then read the trace data out from trace_pipe_raw into the same type of data structure. This saves on converting numbers into ASCII that would be required by trace_marker. Suggested-by:
Olof Johansson <olof@lixom.net> Signed-off-by:
-
- 14 Nov, 2016 7 commits
-
-
Stephen Boyd authored
In similar spirit to x86 and arm64 support, add a make_nop_arm() to replace calls to mcount with a nop in sections that aren't traced. Link: http://lkml.kernel.org/r/20161018234200.5804-1-sboyd@codeaurora.orgAcked-by:
Russell King <rmk+kernel@armlinux.org.uk> Acked-by:
Rabin Vincent <rabin@rab.in> Signed-off-by:
Stephen Boyd <sboyd@codeaurora.org> Signed-off-by:
-
Zhou Chengming authored
Allow wakeup_dl tracer to be used by instances, like wakeup tracer and wakeup_rt tracer. Link: http://lkml.kernel.org/r/1479093553-31264-1-git-send-email-zhouchengming1@huawei.comSigned-off-by:
Zhou Chengming <zhouchengming1@huawei.com> Signed-off-by:
-
Steven Rostedt (Red Hat) authored
The trace_events_file.c filter logic can be a bit complex. I copy this into a userspace program where I can debug it a bit easier. One issue is the op is defined in most places as an int instead of as an enum, and gdb just gives the value when debugging. Having the actual op name shown in gdb is more useful. This has no functionality change, but helps in debugging when the file is debugged in user space. Signed-off-by: -
Steven Rostedt (Red Hat) authored
Currently the filter logic for comparisons (like greater-than and less-than) are used, they share the same function and a switch statement is used to jump to the comparison type to perform. This is done in the extreme hot path of the tracing code, and it does not take much more space to create a unique comparison function to perform each type of comparison and remove the switch statement. Also, a bug was found where the binary and operation for 64 bits could fail if the resulting bits were greater than 32 bits, because the result was passed into a 32 bit variable. This was fixed when adding the separate binary and function. Signed-off-by: -
Masami Hiramatsu authored
Use glob_match() to support flexible glob wildcards (*,?) and character classes ([) for ftrace. Since the full glob matching is slower than the current partial matching routines(*pat, pat*, *pat*), this leaves those routines and just add MATCH_GLOB for complex glob expression. e.g. ---- [root@localhost tracing]# echo 'sched*group' > set_ftrace_filter [root@localhost tracing]# cat set_ftrace_filter sched_free_group sched_change_group sched_create_group sched_online_group sched_destroy_group sched_offline_group [root@localhost tracing]# echo '[Ss]y[Ss]_*' > set_ftrace_filter [root@localhost tracing]# head set_ftrace_filter sys_arch_prctl sys_rt_sigreturn sys_ioperm SyS_iopl sys_modify_ldt SyS_mmap SyS_set_thread_area SyS_get_thread_area SyS_set_tid_address sys_fork ---- Link: http://lkml.kernel.org/r/147566869501.29136.6462645009894738056.stgit@devboxAcked-by:
Namhyung Kim <namhyung@kernel.org> Signed-off-by:
-
Steven Rostedt (Red Hat) authored
When a module is first loaded and its function ip records are added to the ftrace list of functions to modify, they are set to DISABLED, as their text is still in a read only state. When the module is fully loaded, and can be updated, the flag is cleared, and if their's any functions that should be tracing them, it is updated at that moment. But there's several locations that do record accounting and should ignore records that are marked as disabled, or they can cause issues. Alexei already fixed one location, but others need to be addressed. Cc: stable@vger.kernel.org Fixes: b7ffffbb "ftrace: Add infrastructure for delayed enabling of module functions" Reported-by:
Alexei Starovoitov <ast@kernel.org> Signed-off-by:
-
Alexei Starovoitov authored
ftrace_shutdown() checks for sanity of ftrace records and if dyn_ftrace->flags is not zero, it will warn. It can happen that 'flags' are set to FTRACE_FL_DISABLED at this point, since some module was loaded, but before ftrace_module_enable() cleared the flags for this module. In other words the module.c is doing: ftrace_module_init(mod); // calls ftrace_update_code() that sets flags=FTRACE_FL_DISABLED ... // here ftrace_shutdown() is called that warns, since err = prepare_coming_module(mod); // didn't have a chance to clear FTRACE_FL_DISABLED Fix it by ignoring disabled records. It's similar to what __ftrace_hash_rec_update() is already doing. Link: http://lkml.kernel.org/r/1478560460-3818619-1-git-send-email-ast@fb.com Cc: stable@vger.kernel.org Fixes: b7ffffbb "ftrace: Add infrastructure for delayed enabling of module functions" Signed-off-by:
-
- 13 Nov, 2016 3 commits
-
-
Linus Torvalds authored
-
git://git.kernel.org/pub/scm/virt/kvm/kvmLinus Torvalds authored
Pull KVM fixes from Paolo Bonzini: "ARM fixes. There are a couple pending x86 patches but they'll have to wait for next week" * tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: KVM: arm/arm64: vgic: Kick VCPUs when queueing already pending IRQs KVM: arm/arm64: vgic: Prevent access to invalid SPIs arm/arm64: KVM: Perform local TLB invalidation when multiplexing vcpus on a single CPU
-
Linus Torvalds authored
Merge media fixes from Mauro Carvalho Chehab: "This contains two patches fixing problems with my patch series meant to make USB drivers to work again after the DMA on stack changes. The last patch on this series is actually not related to DMA on stack. It solves a longstanding bug affecting module unload, causing module_put() to be called twice. It was reported by the user who reported and tested the issues with the gp8psk driver with the DMA fixup patches. As we're late at -rc cycle, maybe you prefer to not apply it right now. If this is the case, I'll add to the pile of patches for 4.10. Exceptionally this time, I'm sending the patches via e-mail, because I'm on another trip, and won't be able to use the usual procedure until Monday. Also, it is only three patches, and you followed already the discussions about the first one" * emailed patches from Mauro Carvalho Chehab <mchehab@osg.samsung.com>: gp8psk: Fix DVB frontend attach gp8psk: fix gp8psk_usb_in_op() logic dvb-usb: move data_mutex to struct dvb_usb_device
-
