Re: [linux-safety] [PATCH] kernel: trace: Remove deadstore in trace_uprobe.c

From: Milan Lakhani
Date: Thu Dec 10 2020 - 06:30:19 EST



On 10/12/2020 05:47, Lukas Bulwahn wrote:
On Wed, Dec 9, 2020 at 2:17 PM Milan Lakhani
<milan.lakhani@xxxxxxxxxxxxxxx> wrote:
In trace_uprobe.c, trace_uprobe_create assigns ret to 0 but then never
uses this value.

Milan, the patch makes sense, but I fear you did not run
./scripts/get_maintainers.pl because you did not include any specific
maintainer as recipient.

So I did actually run get_maintainers.pl but somehow in my text editor where I had pasted the e-mail addresses they had got onto two separate lines, and git send-email must have only taken the first line. I'll have to get the e-mails right first time next time, to move onto checking documentation!

The patch subject could be: remove unneeded initialization (instead of
the very generic "dead store" term).
Updated this, thanks

It is also interesting to see who added this initialization; was it
unneeded since the existence of this function, did it become obsolete
at some point in time due to refactoring?

The line was unneeded since the function was created in 2012 by Srikar Dronamraju in f3f096cfe, there have actually somehow been six commits to the lines between the initial assignment and reassignment of ret but they have all left the unnecessary initialization there!


Run ./scripts/get_maintainers.pl and please CC: me; then you will get
your Reviewed-by: tag.

Lukas

Signed-off-by: Milan Lakhani <milan.lakhani@xxxxxxxxxxxxxxx>
---
kernel/trace/trace_uprobe.c | 1 -
1 file changed, 1 deletion(-)

diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c
index 3cf7128..c7c7070 100644
--- a/kernel/trace/trace_uprobe.c
+++ b/kernel/trace/trace_uprobe.c
@@ -541,7 +541,6 @@ static int trace_uprobe_create(int argc, const char **argv)
bool is_return = false;
int i, ret;

- ret = 0;
ref_ctr_offset = 0;

switch (argv[0][0]) {
--
2.7.4