[PATCH] ftrace: Handle commands when closing set_ftrace_filter file

[PATCH] ftrace: Handle commands when closing set_ftrace_filter file

[Steven Rostedt]

From: "Steven Rostedt (VMware)" <rostedt@goodmis.org>

 # echo switch_mm:traceoff > /sys/kernel/tracing/set_ftrace_filter

will cause switch_mm to stop tracing by the traceoff command.

 # echo -n switch_mm:traceoff > /sys/kernel/tracing/set_ftrace_filter

does nothing.

The reason is that the parsing in the write function only processes
commands if it finished parsing (there is white space written after the
command). That's to handle:

 write(fd, "switch_mm:", 10);
 write(fd, "traceoff", 8);

cases, where the command is broken over multiple writes.

The problem is if the file descriptor is closed, then the write call is
not processed, and the command needs to be processed in the release code.
The release code can handle matching of functions, but does not handle
commands.

Cc: stable@vger.kernel.org
Fixes: eda1e32855656 ("tracing: handle broken names in ftrace filter")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
---
 kernel/trace/ftrace.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index 057e962ca5ce..c57508445faa 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -5591,7 +5591,10 @@ int ftrace_regex_release(struct inode *inode, struct file *file)
 
 	parser = &iter->parser;
 	if (trace_parser_loaded(parser)) {
-		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
+		int enable = !(iter->flags & FTRACE_ITER_NOTRACE);
+
+		ftrace_process_regex(iter, parser->buffer,
+				     parser->idx, enable);
 	}
 
 	trace_parser_put(parser);
-- 
2.29.2

Re: [PATCH] ftrace: Handle commands when closing set_ftrace_filter file

[Jiri Olsa]

On Wed, May 05, 2021 at 10:48:18AM -0400, Steven Rostedt wrote:
> From: "Steven Rostedt (VMware)" <rostedt@goodmis.org>
> 
>  # echo switch_mm:traceoff > /sys/kernel/tracing/set_ftrace_filter
> 
> will cause switch_mm to stop tracing by the traceoff command.
> 
>  # echo -n switch_mm:traceoff > /sys/kernel/tracing/set_ftrace_filter
> 
> does nothing.
> 
> The reason is that the parsing in the write function only processes
> commands if it finished parsing (there is white space written after the
> command). That's to handle:
> 
>  write(fd, "switch_mm:", 10);
>  write(fd, "traceoff", 8);
> 
> cases, where the command is broken over multiple writes.
> 
> The problem is if the file descriptor is closed, then the write call is
> not processed, and the command needs to be processed in the release code.
> The release code can handle matching of functions, but does not handle
> commands.
> 
> Cc: stable@vger.kernel.org
> Fixes: eda1e32855656 ("tracing: handle broken names in ftrace filter")

nice, breaking kernel since 2009.. I'll put that on t-shirt ;-)

I saw the patch got already merged, FWIW:

Acked-by: Jiri Olsa <jolsa@kernel.org>

jirka

> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
> ---
>  kernel/trace/ftrace.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> index 057e962ca5ce..c57508445faa 100644
> --- a/kernel/trace/ftrace.c
> +++ b/kernel/trace/ftrace.c
> @@ -5591,7 +5591,10 @@ int ftrace_regex_release(struct inode *inode, struct file *file)
>  
>  	parser = &iter->parser;
>  	if (trace_parser_loaded(parser)) {
> -		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
> +		int enable = !(iter->flags & FTRACE_ITER_NOTRACE);
> +
> +		ftrace_process_regex(iter, parser->buffer,
> +				     parser->idx, enable);
>  	}
>  
>  	trace_parser_put(parser);
> -- 
> 2.29.2
> 

Re: [PATCH] ftrace: Handle commands when closing set_ftrace_filter file

[Steven Rostedt]

On Thu, 13 May 2021 22:24:19 +0200
Jiri Olsa <jolsa@redhat.com> wrote:

> > Fixes: eda1e32855656 ("tracing: handle broken names in ftrace filter")  
> 
> nice, breaking kernel since 2009.. I'll put that on t-shirt ;-)

Hmm, maybe I'll recommend that for the Linux Plumber's t-shirt ;-)

 "Breaking kernel's since 2009"!

> 
> I saw the patch got already merged, FWIW:
> 
> Acked-by: Jiri Olsa <jolsa@kernel.org>

I'm guessing you hit the bug too?

-- Steve

Re: [PATCH] ftrace: Handle commands when closing set_ftrace_filter file

[Jiri Olsa]

On Thu, May 13, 2021 at 04:45:04PM -0400, Steven Rostedt wrote:
> On Thu, 13 May 2021 22:24:19 +0200
> Jiri Olsa <jolsa@redhat.com> wrote:
> 
> > > Fixes: eda1e32855656 ("tracing: handle broken names in ftrace filter")  
> > 
> > nice, breaking kernel since 2009.. I'll put that on t-shirt ;-)
> 
> Hmm, maybe I'll recommend that for the Linux Plumber's t-shirt ;-)
> 
>  "Breaking kernel's since 2009"!
> 
> > 
> > I saw the patch got already merged, FWIW:
> > 
> > Acked-by: Jiri Olsa <jolsa@kernel.org>
> 
> I'm guessing you hit the bug too?

no, I did not notice before, just reproduced based on your changelog

jirka