From: Colin Ian King
There is a spelling mistake in the -g help option, I believe
it should be "graph". There is also a spelling mistake in a
warning message. Fix both mistakes.
Signed-off-by: Colin Ian King
Signed-off-by: Viktor Rosendahl
---
tools/tracing/latency/latency-colle
Hi Steve,
I am not sure what is the correct approach to handle this latency-collector
fix. Can you take it into your queue?
best regards,
Viktor
Colin Ian King (1):
tracing/tools: fix a couple of spelling mistakes
tools/tracing/latency/latency-collector.c | 6 +++---
1 file changed, 3
the ftrace state when the program terminates
* Cleaned up the commit message a bit, removed example output etc.
Viktor Rosendahl (1):
Add the latency-collector to tools
tools/Makefile| 14 +-
tools/tracing/Makefile| 19 +
tools/tracing/latency
, preemptirqtest_1, preemptirqtest_2 in the
stack traces.
Signed-off-by: Viktor Rosendahl
---
tools/Makefile| 14 +-
tools/tracing/Makefile| 19 +
tools/tracing/latency/.gitignore |2 +
tools/tracing/latency/Makefile| 2
, preemptirqtest_1, preemptirqtest_2 in the
stack traces.
Signed-off-by: Viktor Rosendahl
---
tools/Makefile| 14 +-
tools/tracing/Makefile| 19 +
tools/tracing/latency/Makefile| 24 +
tools/tracing/latency/latency-
AR_CFLAGS)
+LDFLAGS = -lpthread $(VAR_LDLIBS)
+
+all: $(TARGETS)
+
+%: %.c
+ $(CC) $(CFLAGS) -o $@ $< $(LDFLAGS)
+
+clean:
+ $(RM) latency-collector
+
+sbindir ?= /usr/sbin
+
+install: all
+ install -d $(DESTDIR)$(sbindir)
+ install -m 755 -p $(TARGETS) $(DESTDIR)$(sbindir)
d
r even
how feasible it is without significant rework.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/trace_irqsoff.c | 4
1 file changed, 4 insertions(+)
diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c
index d06aab4dcbb8..6756379b661f 100644
--- a/kernel/trace/trace_irqso
lt;<<<<
In the example above, we randomly, happened to get the third latency
in a burst of 10. If the experiment is repeated enough times, we will
get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify loop
..or:
3
the next but when one is doing
validation with a fleet of devices being tested in a long and expensive test
campaign, then it is quite desirable to not lose any latencies.
best regards,
Viktor
Viktor Rosendahl (2):
Use pause-on-trace with the latency tracers
Add the latency-collector to tools
_on() could be used to queue the work on another CPU but it
seems difficult to select the right CPU.
Signed-off-by: Viktor Rosendahl (BMW)
---
kernel/trace/trace.c | 73 ++--
kernel/trace/trace.h | 7 +
2 files changed, 78 insertions(+), 2 deleti
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl (BMW)
---
.../admin-guide/kernel-parameters.txt | 4 +++
inc
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl (BMW)
Reviewed-by: Joel Fernandes (Google)
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 144
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify loop
..or
it does, so in principle it could be
applied separately from the others.
[PATCH 4/4] This adds the option trace_console_latency=1 to the kernel
parameters. This makes it possible to enable tracing of console latencies.
best regards,
Viktor
Viktor Rosendahl (BMW) (4):
ftrace: Implement fs
ute on the same CPU but if we are
unlucky that CPU could be too busy while there could be another CPU in
the system that would be able to process the work soon enough.
queue_work_on() could be used to queue the work on another CPU but it
seems difficult to select the right CPU.
Signed-off-by: Viktor
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify loop
..or
cleanly on its own. However, now it does, so in principle it could be
applied separately from the others.
[PATCH 4/4] This adds the option console-latency to the trace options. This
makes it possible to enable tracing of console latencies.
best regards,
Viktor
Viktor Rosendahl (BMW) (4):
ftrace
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl (BMW)
Reviewed-by: Joel Fernandes (Google)
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 144
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl (BMW)
---
.../admin-guide/kernel-parameters.txt | 4 +++
inc
_on() could be used to queue the work on another CPU but it
seems difficult to select the right CPU.
Signed-off-by: Viktor Rosendahl (BMW)
---
kernel/trace/trace.c | 75 ++--
kernel/trace/trace.h | 18 +++
2 files changed, 91 insertions(+), 2 deleti
Den ons 9 okt. 2019 kl 20:08 skrev Steven Rostedt :
>
> On Wed, 9 Oct 2019 16:51:07 +0200
> Viktor Rosendahl wrote:
>
> > Apologies, I should have replied there but I have been a bit busy the
> > last few days with other topics, and I felt a bit indecisive about
> >
Den ons 9 okt. 2019 kl 16:11 skrev Joel Fernandes :
>
> On Wed, Oct 09, 2019 at 12:08:24AM +0200, Viktor Rosendahl (BMW) wrote:
> > This new trace option "console-latency" will enable the latency
> > tracers to trace the console latencies. Previously this has always b
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl (BMW)
---
include/linux/irqflags.h | 22 ++
kernel/print
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl (BMW)
Reviewed-by: Joel Fernandes (Google)
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 144
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify l
it possible to enable tracing of console latencies.
best regards,
Viktor
Viktor Rosendahl (BMW) (4):
ftrace: Implement fs notification for tracing_max_latency
preemptirq_delay_test: Add the burst feature and a sysfs trigger
Add the latency-collector to tools
ftrace: Add an option for tracing
ugh.
queue_work_on() could be used to queue the work on another CPU but it
seems difficult to select the right CPU.
Signed-off-by: Viktor Rosendahl (BMW)
Reviewed-by: Joel Fernandes (Google)
---
kernel/trace/trace.c | 75 +-
kernel/trace/trace.h |
On 10/2/19 2:32 AM, Joel Fernandes wrote:
Some nits, but others looks good:
On Fri, Sep 20, 2019 at 05:22:16PM +0200, Viktor Rosendahl (BMW) wrote:
This patch implements the feature that the tracing_max_latency file,
e.g. /sys/kernel/debug/tracing/tracing_max_latency will receive
notifications
On 10/2/19 2:52 AM, Joel Fernandes wrote:
On Fri, Sep 20, 2019 at 05:22:19PM +0200, Viktor Rosendahl (BMW) wrote:
This new trace option "console-latency" will enable the latency
tracers to trace the console latencies. Previously this has always been
implicitely disabled. I guess this
On 10/2/19 5:13 PM, Steven Rostedt wrote:
On Fri, 20 Sep 2019 17:22:16 +0200
"Viktor Rosendahl (BMW)" wrote:
This patch implements the feature that the tracing_max_latency file,
e.g. /sys/kernel/debug/tracing/tracing_max_latency will receive
notifications through the fsnotify fram
ugh.
queue_work_on() could be used to queue the work on another CPU but it
seems difficult to select the right CPU.
Signed-off-by: Viktor Rosendahl (BMW)
---
kernel/trace/trace.c | 75 +-
kernel/trace/trace.h | 18 +
kernel/trace/trace_hwlat.c |
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify l
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl (BMW)
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 144 +++
2 files changed
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl (BMW)
---
include/linux/irqflags.h | 22 ++
kernel/print
it possible to enable tracing of console latencies.
best regards,
Viktor
Viktor Rosendahl (BMW) (4):
ftrace: Implement fs notification for tracing_max_latency
preemptirq_delay_test: Add the burst feature and a sysfs trigger
Add the latency-collector to tools
ftrace: Add an option for tracing
On 9/8/19 1:38 AM, Joel Fernandes wrote:> On Sat, Sep 07, 2019 at
11:12:59PM +0200, Viktor Rosendahl wrote:
>> On 9/6/19 4:17 PM, Joel Fernandes wrote:
>>> On Thu, Sep 05, 2019 at 03:25:45PM +0200, Viktor Rosendahl wrote:
>>
>>>> +
>>>&g
On 9/6/19 4:17 PM, Joel Fernandes wrote:
On Thu, Sep 05, 2019 at 03:25:45PM +0200, Viktor Rosendahl wrote:
+
+__init static int latency_fsnotify_init(void)
+{
+ fsnotify_wq = alloc_workqueue("tr_max_lat_wq",
+ WQ_UNBOUND | WQ_
On 9/5/19 6:52 PM, Steven Rostedt wrote:
On Wed, 4 Sep 2019 21:10:38 +0200
Viktor Rosendahl wrote:
On 9/4/19 1:42 PM, Steven Rostedt wrote:
I am not so keen to clutter source files with a new copyright message.
My problem is that git-send-email doesn't work well with my work email
address
happen with a softirq, kernel thread or tasklet. For this reason we
use the irq_work mechanism to call queue_work().
Signed-off-by: Viktor Rosendahl
---
kernel/trace/trace.c | 75 +-
kernel/trace/trace.h | 18 +
kernel/trace/trace_hwlat.c | 4
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl
---
include/linux/irqflags.h | 22 ++
kernel/printk/printk.c
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify loo
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 147 +++
2 files changed, 131
on its own. However, now it does, so in principle it could be
applied separately from the others.
[PATCH 4/4] This adds the option console-latency to the trace options. This
makes it possible to enable tracing of console latencies.
best regards,
Viktor
Viktor Rosendahl (4):
ftrace: Implement fs
On 9/4/19 1:42 PM, Steven Rostedt wrote:
On Tue, 3 Sep 2019 15:26:00 +0200
Viktor Rosendahl wrote:
diff --git a/kernel/trace/preemptirq_delay_test.c
b/kernel/trace/preemptirq_delay_test.c
index d8765c952fab..dc281fa75198 100644
--- a/kernel/trace/preemptirq_delay_test.c
+++ b/kernel/trace
On 9/4/19 1:39 PM, Steven Rostedt wrote:
On Tue, 3 Sep 2019 15:25:59 +0200
Viktor Rosendahl wrote:
+void latency_fsnotify_stop(void)
+{
+ /* Make sure all CPUs see caller's previous actions to stop tracer */
+ smp_wmb();
These memory barriers just look wrong. What exactly
On 9/4/19 10:19 AM, Peter Zijlstra wrote:
Adding Paul since RCU faces similar situations, i.e. raising softirq risks
scheduler deadlock in rcu_read_unlock_special() -- but RCU's solution is to
avoid raising the softirq and instead use irq_work.
Which is right.
Thanks Joel and Peter for
of __schedule(). On the other hand, in some places, like in idle and
the console we need start stop functions that prevents the
measurements from being make.
Signed-off-by: Viktor Rosendahl
---
include/linux/ftrace.h| 31 +
kernel/sched/core.c | 3 +
kernel/sched
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 145 +++
2 files changed, 129
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl
---
include/linux/irqflags.h | 21 +
kernel/printk/printk.c
so in principle it could be
applied separately from the others.
[PATCH 4/4] This adds the option console-latency to the trace options. This
makes it possible to enable tracing of console latencies.
best regards,
Viktor
Viktor Rosendahl (4):
ftrace: Implement fs notification for tracing_max_late
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify loo
On 5/21/19 6:01 PM, Steven Rostedt wrote:>
>
> Note, you need to add the scheduling and power management maintainers
> when adding trace events to their code.
>
My bad.
> As these trace events become visible to user space, and you only need a
> callback to disable fsnotify, it may be better to
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl
---
include/linux/irqflags.h | 21 +
kernel/printk/printk.c
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify loo
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 145 +++
2 files changed, 129
exited
do_idle()/__schedule().
Signed-off-by: Viktor Rosendahl
---
include/trace/events/power.h | 40 +
include/trace/events/sched.h | 40 +
kernel/sched/core.c | 2 +
kernel/sched/idle.c | 2 +
kernel/trace/trace.c | 168
own:
[PATCH 4/4] This adds the option console-latency to the trace options. This
makes it possible to enable tracing of console latencies.
best regards,
Viktor
Viktor Rosendahl (4):
ftrace: Implement fs notification for tracing_max_latency
preemptirq_delay_test: Add the burst feature
On 5/14/19 6:51 PM, Steven Rostedt wrote:
Hi Viktor,
Note, as the merge window is open and I'm also currently at a
conference, it may be a while before I can take a look at this. If you
don't hear something from me in 10 days, feel free to send me a ping.
-- Steve
Hi Steve,
Ok, good to
;>>>>>>> END <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is re
vertheless be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl
---
include/linux/irqflags.h| 21 +
kernel/print
exited
do_idle()/__schedule().
Signed-off-by: Viktor Rosendahl
---
include/trace/events/power.h | 40 +
include/trace/events/sched.h | 40 +
kernel/sched/core.c | 2 +
kernel/sched/idle.c | 2 +
kernel/trace/trace.c | 160
] This adds the option console-latency to the trace options.
best regards,
Viktor
Viktor Rosendahl (4):
ftrace: Implement fs notification for tracing_max_latency
preemptirq_delay_test: Add the burst feature and a sysfs trigger
Add the latency-collector to tools
ftrace: Add an option
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 145 +++
2 files changed, 129
On 5/6/19 1:01 AM, Steven Rostedt wrote:
> On Mon, 6 May 2019 00:39:15 +0200
> Viktor Rosendahl wrote:
>
>> Can you explain more precisely what you agree with?
>>
>> The general idea of being able to trace bursts of latencies?
>
> One thing I have an issue wit
On 5/4/19 6:47 PM, Joel Fernandes wrote:
> On Wed, May 01, 2019 at 10:36:47PM +0200, Viktor Rosendahl wrote:
>> immediately after each other in spite of the fact that the
>> preempt/irqsoff tracers operate in overwrite mode.
>>
>> Signed-off-by: Viktor Rosendahl
>
&
> Instead of this being turned into a nop, don't have a kconfig option
> but instead have this call into the trace_irqsoff.c code, and depending
> on what the options are, it should stop it. Of course, this would need
> to be smart enough to pair it. Perhaps return the result of
>
be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl
---
include/linux/irqflags.h | 13 +
kernel/printk/printk.c | 5 +++--
kernel
<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get all 10.
Sometimes, there will be lost latencies, so that we get:
3054.078294 Latency 44 printout skipped due to inotify loo
to
reload the module to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/Kconfig | 6 +-
kernel/trace/preemptirq_delay_test.c | 145 +++
2 files changed, 129
is not directly connected but doesn't apply cleanly on
its own:
[PATCH 4/4] This adds the option CONFIG_TRACE_CONSOLE_LATENCY to
decide whether we want to trace prints to the console or not.
best regards,
Viktor Rosendahl
Viktor Rosendahl (4):
ftrace: Implement fs notification for preempt/irqsoff
of latencies that occur
immediately after each other in spite of the fact that the
preempt/irqsoff tracers operate in overwrite mode.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/Kconfig | 10 ++
kernel/trace/trace.c | 31 +--
kernel/trace/trace.h
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
In the example above, we randomly, happened to get the third latency
in a burst burst of 10. If the experiment is repeated enough times,
we will get a
to repeat the test. The trigger will appear as
/sys/kernel/preemptirq_delay_test/trigger in sysfs.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/preemptirq_delay_test.c | 139 +++
1 file changed, 120 insertions(+), 19 deletions(-)
diff --git a/kernel/trace
be desirable to
trace them. Basically, we want to be able to tell that there are
latencies in the system under test because someone has incorrectly
enabled the serial console.
Signed-off-by: Viktor Rosendahl
---
include/linux/irqflags.h | 13 +
kernel/printk/printk.c | 5 +++--
kernel
of latencies that occur
immediately after each other in spite of the fact that the
preempt/irqsoff tracers operate in overwrite mode.
Signed-off-by: Viktor Rosendahl
---
kernel/trace/Kconfig | 10 ++
kernel/trace/trace.c | 31 +--
kernel/trace/trace.h
] This adds the option CONFIG_TRACE_CONSOLE_LATENCY to
decide whether we want to trace prints to the console or not.
best regards,
Viktor Rosendahl
Viktor Rosendahl (4):
ftrace: Implement fs notification for preempt/irqsoff tracers
preemptirq_delay_test: Add the burst feature and a sysfs trigger
76 matches
Mail list logo