On Wed, 8 Jan 2025 10:30:08 +0800
Donglin Peng wrote:
> Steven Rostedt 于2024年12月24日 周二04:14写道：
>
> > From: Sven Schnelle
> >
> > Add a function to decode argument types with the help of BTF. Will
> > be used to display arguments in the function and function graph
> > tracer.
> >
> > It can only
From: Steven Rostedt
Instead of having the main code live in a header file and included twice
with MACROs that define the Elf structures for 64 bit or 32 bit, move the
code in the C file now that the Elf structures are defined in a union that
has both. All accesses to the Elf structure fields are
From: Masami Hiramatsu (Google)
Adopt __free() and guard() for trace_fprobe.c to remove gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
kernel/trace/trace_fprobe.c | 129 ---
1 file changed, 60 insertions(+), 69 deletions(-)
diff --git a/kernel/tra
Hi,
Here is the 5th version of the series to fix and cleanup probe events in
ftrace with __free(). The previous series is here;
https://lore.kernel.org/all/173629943548.1450013.12412761424554510119.stgit@devnote2/
In this version, I updated [3/5] according to the comment from Steve.
Thanks,
--
From: Masami Hiramatsu (Google)
Simplify __trace_kprobe_create() by removing gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
kernel/trace/trace_kprobe.c | 82 ++-
1 file changed, 41 insertions(+), 41 deletions(-)
diff --git a/kernel/trace/trace_kp
From: Masami Hiramatsu (Google)
Use __free() in trace_kprobe.c to cleanup code.
Signed-off-by: Masami Hiramatsu (Google)
---
Changes in v4:
- Use no_free_ptr(tk)->tp instead of assiging NULL to tk.
Changes in v3:
- Rename to __free(free_trace_kprobe) to clarify what function will be calle
From: Masami Hiramatsu (Google)
Use __free() in trace_probe to cleanup some gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
Changes in v5:
- Fix to remove kfree() in for loop. Since I confirmed __free(kfree) is
called in each iteration.
---
kernel/trace/trace_probe.c | 51
From: Masami Hiramatsu (Google)
In __trace_kprobe_create(), if something fails it must goto error block
to free objects. But when strdup() a symbol, it returns without that.
Fix it to goto the error block to free objects correctly.
Fixes: 6212dd29683e ("tracing/kprobes: Use dyn_event framework f
On Tue, 7 Jan 2025 20:34:32 -0500
Steven Rostedt wrote:
> On Wed, 8 Jan 2025 09:38:43 +0900
> Masami Hiramatsu (Google) wrote:
>
> > > I don't get this? You are telling the compiler not to free tmp, because
> > > you
> > > decided to free it yourself? Why not just remove the kfree() here
> >
在 2025/1/7 20:53, Lukas Wunner 写道:
On Tue, Jan 07, 2025 at 07:30:28PM +0800, Shuai Xue wrote:
2024/11/23 19:31, Shuai Xue:
To this end, define a new TRACING_SYSTEM named pci, add a generic RAS
tracepoint for hotplug event to help healthy check, and generate
tracepoints for pcie hotplug event
On Wed, 8 Jan 2025 09:38:43 +0900
Masami Hiramatsu (Google) wrote:
> > I don't get this? You are telling the compiler not to free tmp, because you
> > decided to free it yourself? Why not just remove the kfree() here
> > altogether?
>
> In the for-loop block, the __free() work only when we ex
From: Masami Hiramatsu (Google)
Adopt __free() and guard() for trace_fprobe.c to remove gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
kernel/trace/trace_fprobe.c | 129 ---
1 file changed, 60 insertions(+), 69 deletions(-)
diff --git a/kernel/tra
From: Masami Hiramatsu (Google)
Simplify __trace_kprobe_create() by removing gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
kernel/trace/trace_kprobe.c | 82 ++-
1 file changed, 41 insertions(+), 41 deletions(-)
diff --git a/kernel/trace/trace_kp
From: Masami Hiramatsu (Google)
Use __free() in trace_kprobe.c to cleanup code.
Signed-off-by: Masami Hiramatsu (Google)
---
Changes in v4:
- Use no_free_ptr(tk)->tp instead of assiging NULL to tk.
Changes in v3:
- Rename to __free(free_trace_kprobe) to clarify what function will be calle
From: Masami Hiramatsu (Google)
Use __free() in trace_probe to cleanup some gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
Changes in v4:
- Run kfree() right before kstrdup() in a loop since __free(kfree) works only
when exiting the loop, not each iteration.
---
kernel/trace/trace_
From: Masami Hiramatsu (Google)
In __trace_kprobe_create(), if something fails it must goto error block
to free objects. But when strdup() a symbol, it returns without that.
Fix it to goto the error block to free objects correctly.
Fixes: 6212dd29683e ("tracing/kprobes: Use dyn_event framework f
Hi,
Here is the 4th version of the series to fix and cleanup probe events in
ftrace with __free(). The previous series is here;
https://lore.kernel.org/all/173625060316.1375434.11048027439794595989.stgit@devnote2/
In this version, I updated [3/5] and [5/5] according to the comment from Steve.
T
On Tue, 7 Jan 2025 10:42:31 -0500
Steven Rostedt wrote:
> On Tue, 7 Jan 2025 20:50:35 +0900
> "Masami Hiramatsu (Google)" wrote:
>
> > @@ -1898,7 +1899,8 @@ create_local_trace_kprobe(char *func, void *addr,
> > unsigned long offs,
> > bool is_return)
> > {
> > enum
On Tue, 7 Jan 2025 10:36:43 -0500
Steven Rostedt wrote:
> On Tue, 7 Jan 2025 20:50:25 +0900
> "Masami Hiramatsu (Google)" wrote:
>
> > @@ -1790,18 +1777,15 @@ int traceprobe_expand_dentry_args(int argc, const
> > char *argv[], char **buf)
> >offsetof(struct
On Sat, Nov 23, 2024 at 07:31:08PM +0800, Shuai Xue wrote:
> Hotplug events are critical indicators for analyzing hardware health,
> particularly in AI supercomputers where surprise link downs can
> significantly impact system performance and reliability. The failure
> characterization analysis ill
On Tue, 7 Jan 2025 18:02:24 -0500
Steven Rostedt wrote:
> diff --git a/src/event-parse.c b/src/event-parse.c
> index 33ed7fb47fff..f2e50b0e8992 100644
> --- a/src/event-parse.c
> +++ b/src/event-parse.c
> @@ -7841,7 +7841,7 @@ static enum tep_errno parse_format(struct tep_event
> **eventp,
>
On Tue, 7 Jan 2025 11:04:57 +
Shiju Jose wrote:
> Please find attached format file, which might have failed to parse because of
> the
> unsupported formats in the libtraceevent you mentioned in the following
> thread.
> https://lore.kernel.org/lkml/20241127104132.6c172...@gandalf.local.home
On Tue, Jan 07, 2025 at 02:54:39PM -0500, Steven Rostedt wrote:
>
> Hi Vincent,
>
> Are you still working on this series? I was expecting an update.
Yes I am! I had actually a new version ready just before christmas but I thought
I'd better wait for the new year before sending it so it does not
On Tue, 7 Jan 2025 15:48:20 +0100
Tomas Glozar wrote:
> The patchset makes both rtla-timerlat-hist and rtla-timerlat-top set
> OSNOISE_WORKLOAD to the proper value ("on" when running with -k, "off"
> when running with -u) every time the option is available instead of setting it
> only when runni
Hi Vincent,
Are you still working on this series? I was expecting an update.
-- Steve
On Wed, 11 Sep 2024 10:30:16 +0100
Vincent Donnefort wrote:
> The growing set of features supported by the hypervisor in protected
> mode necessitates debugging and profiling tools. Tracefs is the
> ideal
On 2024-12-27 12:19:46 [-0500], Steven Rostedt wrote:
…
> Is easier to understand than:
>
> if (!mod) {
> guard(rcu)();
> mod = __module_text_address(pc);
> }
>
> Because it makes me wonder, why use a guard() for a one liner?
Why not? The context ends imme
On Tue, 7 Jan 2025 20:50:35 +0900
"Masami Hiramatsu (Google)" wrote:
> @@ -1898,7 +1899,8 @@ create_local_trace_kprobe(char *func, void *addr,
> unsigned long offs,
> bool is_return)
> {
> enum probe_print_type ptype;
> - struct trace_kprobe *tk;
> + struc
On Tue, 7 Jan 2025 20:50:25 +0900
"Masami Hiramatsu (Google)" wrote:
> @@ -1790,18 +1777,15 @@ int traceprobe_expand_dentry_args(int argc, const
> char *argv[], char **buf)
> offsetof(struct file, f_path.dentry),
> equal
Apply fix from previous commit also to timerlat-top.
Cc: sta...@vger.kernel.org
Fixes: cdca4f4e5e8e ("rtla/timerlat_top: Add timerlat user-space support")
Signed-off-by: Tomas Glozar
---
tools/tracing/rtla/src/timerlat_top.c | 15 +--
1 file changed, 9 insertions(+), 6 deletions(-)
osnoise_set_workload returns -1 for both missing OSNOISE_WORKLOAD option
and failure in setting the option.
Return -1 for missing and -2 for failure to distinguish them.
Cc: sta...@vger.kernel.org
Signed-off-by: Tomas Glozar
---
tools/tracing/rtla/src/osnoise.c | 2 +-
1 file changed, 1 inserti
When using rtla timerlat with userspace threads (-u or -U), rtla
disables the OSNOISE_WORKLOAD option in
/sys/kernel/tracing/osnoise/options. This option is not re-enabled in a
subsequent run with kernel-space threads, leading to rtla collecting no
results if the previous run exited abnormally:
$
The patchset makes both rtla-timerlat-hist and rtla-timerlat-top set
OSNOISE_WORKLOAD to the proper value ("on" when running with -k, "off"
when running with -u) every time the option is available instead of setting it
only when running with -u.
This prevents rtla timerlat -k from giving no result
On Tue, 7 Jan 2025 18:29:20 +0530
Dheeraj Reddy Jonnalagadda wrote:
> Fix an issue in the tracing_set_tracer() where the variable ret could
> be used uninitialized under certain conditions. Specifically, when
> t->use_max_tr is true but tr->cond_snapshot is false, the code would
> check the valu
From: Steven Rostedt
The mcount_loc section holds the addresses of the functions that get
patched by ftrace when enabling function callbacks. It can contain tens of
thousands of entries. These addresses must be sorted. If they are not
sorted at compile time, they are sorted at boot. Sorting at bo
Fix an issue in the tracing_set_tracer() where the variable ret could
be used uninitialized under certain conditions. Specifically, when
t->use_max_tr is true but tr->cond_snapshot is false, the code would
check the value of ret without first initializing it.
Fixes: a35873a0993b ("tracing: Add con
On Tue, Jan 07, 2025 at 07:30:28PM +0800, Shuai Xue wrote:
> 2024/11/23 19:31, Shuai Xue:
> > To this end, define a new TRACING_SYSTEM named pci, add a generic RAS
> > tracepoint for hotplug event to help healthy check, and generate
> > tracepoints for pcie hotplug event. To monitor these tracepoin
From: Masami Hiramatsu (Google)
Make add_remove_uprobe test case more robust against various real
binary name.
Current add_remove_uprobe.tc test expects the real binary of /bin/sh
is '*/bin/*sh', but it does not work on busybox environment.
Instead of using fixed pattern, use readlink to identify
From: Masami Hiramatsu (Google)
Fix mount_options.tc to use remount option to mount the tracefs.
Since the current implementation does not umount the tracefs,
this test always fails because of -EBUSY error.
Using remount option will allow us to change the mount option.
Fixes: 8b55572e5180 ("trac
Here is a couple of patches to fix issues. I think mount_options.tc's one
is a real bug(I'm not sure how it worked), but another one is an enhancement
for (my) execution environment.
Anyway, those should go through kselftests tree.
Thank you,
---
Masami Hiramatsu (Google) (2):
selftests/ftr
On Tue, 7 Jan 2025 20:50:03 +0900
"Masami Hiramatsu (Google)" wrote:
> Hi,
>
> Here is the 3rd version of the series to fix and cleanup probe events in
> ftrace with __free(). I resend this without dynevent and argv_free
> parts because it has been sent by Steve[1]. And I updated the version
>
From: Masami Hiramatsu (Google)
Adopt __free() and guard() for trace_fprobe.c to remove gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
kernel/trace/trace_fprobe.c | 129 ---
1 file changed, 60 insertions(+), 69 deletions(-)
diff --git a/kernel/tra
From: Masami Hiramatsu (Google)
Simplify __trace_kprobe_create() by removing gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
kernel/trace/trace_kprobe.c | 82 ++-
1 file changed, 41 insertions(+), 41 deletions(-)
diff --git a/kernel/trace/trace_kp
From: Masami Hiramatsu (Google)
Use __free() in trace_kprobe.c to cleanup code.
Signed-off-by: Masami Hiramatsu (Google)
---
Changes in v3:
- Rename to __free(free_trace_kprobe) to clarify what function will be called.
- Add !IS_ERR_OR_NULL() check because alloc_trace_kprobe() returns an e
From: Masami Hiramatsu (Google)
Use __free() in trace_probe to cleanup some gotos.
Signed-off-by: Masami Hiramatsu (Google)
---
kernel/trace/trace_probe.c | 52 +++-
1 file changed, 18 insertions(+), 34 deletions(-)
diff --git a/kernel/trace/trace_pro
Hi,
Here is the 3rd version of the series to fix and cleanup probe events in
ftrace with __free(). I resend this without dynevent and argv_free
parts because it has been sent by Steve[1]. And I updated the version
tag.
In this version, I fixed some issues[5/7] and update DEFINE_FREE() tag
name to
From: Masami Hiramatsu (Google)
In __trace_kprobe_create(), if something fails it must goto error block
to free objects. But when strdup() a symbol, it returns without that.
Fix it to goto the error block to free objects correctly.
Fixes: 6212dd29683e ("tracing/kprobes: Use dyn_event framework f
On Tue, 7 Jan 2025 01:03:27 +0900
"Masami Hiramatsu (Google)" wrote:
> From: Masami Hiramatsu (Google)
>
> Use __free() for the args allocated by argv_split() in dynevent.
Lol, the same one has already done by Steve.
https://lore.kernel.org/linux-trace-kernel/20241220103313.4a74e...@gandalf.
在 2024/11/23 19:31, Shuai Xue 写道:
Hotplug events are critical indicators for analyzing hardware health,
particularly in AI supercomputers where surprise link downs can
significantly impact system performance and reliability. The failure
characterization analysis illustrates the significance of
>-Original Message-
>From: Steven Rostedt
>Sent: 06 January 2025 22:12
>To: Shiju Jose
>Cc: mhira...@kernel.org; mathieu.desnoy...@efficios.com; linux-trace-
>ker...@vger.kernel.org; linux-ker...@vger.kernel.org; Linuxarm
>; Jonathan Cameron
>; tanxiaofei ;
>Zengtao (B)
>Subject: Re: [PA
49 matches
Mail list logo
