2013-06-13 06:22:16

by zhangwei(Jovi)

Subject: [PATCH] uprobes: fix return value in error handling path

When I inject incorrect argument into uprobe_events,

[root@jovi tracing]# echo 'p:myprobe /bin/bash' > uprobe_events
[root@jovi tracing]#

it doesn't return any error value in there, this patch fix it.

Signed-off-by: zhangwei(Jovi) <[email protected]>
---
kernel/trace/trace_uprobe.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c
index 32494fb0..d5d0cd3 100644
--- a/kernel/trace/trace_uprobe.c
+++ b/kernel/trace/trace_uprobe.c
@@ -283,8 +283,10 @@ static int create_trace_uprobe(int argc, char **argv)
return -EINVAL;
}
arg = strchr(argv[1], ':');
- if (!arg)
+ if (!arg) {
+ ret = -EINVAL;
goto fail_address_parse;
+ }

*arg++ = '\0';
filename = argv[1];
--
1.7.9.7

2013-06-17 16:56:43

by Steven Rostedt

Subject: Re: [PATCH] uprobes: fix return value in error handling path

Srikar,

Can you give an Ack (or nack) for this?

Thanks,

-- Steve


On Thu, 2013-06-13 at 14:21 +0800, zhangwei(Jovi) wrote:
> When I inject incorrect argument into uprobe_events,
>
> [root@jovi tracing]# echo 'p:myprobe /bin/bash' > uprobe_events
> [root@jovi tracing]#
>
> it doesn't return any error value in there, this patch fix it.
>
> Signed-off-by: zhangwei(Jovi) <[email protected]>
> ---
> kernel/trace/trace_uprobe.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c
> index 32494fb0..d5d0cd3 100644
> --- a/kernel/trace/trace_uprobe.c
> +++ b/kernel/trace/trace_uprobe.c
> @@ -283,8 +283,10 @@ static int create_trace_uprobe(int argc, char **argv)
> return -EINVAL;
> }
> arg = strchr(argv[1], ':');
> - if (!arg)
> + if (!arg) {
> + ret = -EINVAL;
> goto fail_address_parse;
> + }
>
> *arg++ = '\0';
> filename = argv[1];

2013-07-02 23:06:09

by Steven Rostedt

Subject: Re: [PATCH] uprobes: fix return value in error handling path

On Mon, 2013-06-17 at 12:56 -0400, Steven Rostedt wrote:
> Srikar,
>
> Can you give an Ack (or nack) for this?
>

This looks rather obvious, I'll apply it and also tag it for stable.

-- Steve