[v1] ftrace: Fix char print issue in print_ip_ins()

[PATCH] ftrace: Fix char print issue in print_ip_ins()

Posted by Zheng Yejian 3 years, 6 months ago

When ftrace bug happened, following log shows every byte data in
problem ip address:
  actual:   ffffffe8:6b:ffffffd9:01:21

However that seems a little confusing, and it is because format
'%x' being used to print signed chars in print_ip_ins().

After this patch, the log is like:
  actual:   e8:6b:d9:01:21

Fixes: 6c14133d2d3f ("ftrace: Do not blindly read the ip address in ftrace_bug()")
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
---
 kernel/trace/ftrace.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index 439e2ab6905e..251ba30c871d 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -2015,7 +2015,7 @@ static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
 
 static void print_ip_ins(const char *fmt, const unsigned char *p)
 {
-	char ins[MCOUNT_INSN_SIZE];
+	unsigned char ins[MCOUNT_INSN_SIZE];
 	int i;
 
 	if (copy_from_kernel_nofault(ins, p, MCOUNT_INSN_SIZE)) {
-- 
2.25.1

Re: [PATCH] ftrace: Fix char print issue in print_ip_ins()

Posted by Joe Perches 3 years, 6 months ago

On Mon, 2022-10-10 at 19:15 +0000, Zheng Yejian wrote:
> When ftrace bug happened, following log shows every byte data in
> problem ip address:
>   actual:   ffffffe8:6b:ffffffd9:01:21
> 
> However that seems a little confusing, and it is because format
> '%x' being used to print signed chars in print_ip_ins().
> 
> After this patch, the log is like:
>   actual:   e8:6b:d9:01:21
> 
> Fixes: 6c14133d2d3f ("ftrace: Do not blindly read the ip address in ftrace_bug()")
> Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
> ---
>  kernel/trace/ftrace.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
[]
> @@ -2015,7 +2015,7 @@ static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
>  
>  static void print_ip_ins(const char *fmt, const unsigned char *p)
>  {
> -	char ins[MCOUNT_INSN_SIZE];
> +	unsigned char ins[MCOUNT_INSN_SIZE];
>  	int i;
>  
>  	if (copy_from_kernel_nofault(ins, p, MCOUNT_INSN_SIZE)) {

Perhaps nicer would be to change the for loop below this from

	for (i = 0; i < MCOUNT_INSN_SIZE; i++)
		printk(KERN_CONT "%s%02x", i ? ":" : "", ins[i]);

to

	pr_cont("%*phC", MCOUNT_INSN_SIZE, ins);

Re: [PATCH] ftrace: Fix char print issue in print_ip_ins()

Posted by Zheng Yejian 3 years, 5 months ago

On Mon, 10 Oct 2022 19:35:38 -0700, Joe Perches <joe@perches.com> wrote:
> On Mon, 2022-10-10 at 19:15 +0000, Zheng Yejian wrote:
> > When ftrace bug happened, following log shows every byte data in
> > problem ip address:
> >   actual:   ffffffe8:6b:ffffffd9:01:21
> > 
> > However that seems a little confusing, and it is because format
> > '%x' being used to print signed chars in print_ip_ins().
> > 
> > After this patch, the log is like:
> >   actual:   e8:6b:d9:01:21
> > 
> > Fixes: 6c14133d2d3f ("ftrace: Do not blindly read the ip address in ftrace_bug()")
> > Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
> > ---
> >  kernel/trace/ftrace.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> []
> > @@ -2015,7 +2015,7 @@ static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
> >  
> >  static void print_ip_ins(const char *fmt, const unsigned char *p)
> >  {
> > -	char ins[MCOUNT_INSN_SIZE];
> > +	unsigned char ins[MCOUNT_INSN_SIZE];
> >  	int i;
> >  
> >  	if (copy_from_kernel_nofault(ins, p, MCOUNT_INSN_SIZE)) {
> 
> Perhaps nicer would be to change the for loop below this from
> 
> 	for (i = 0; i < MCOUNT_INSN_SIZE; i++)
> 		printk(KERN_CONT "%s%02x", i ? ":" : "", ins[i]);
> 
> to
> 
> 	pr_cont("%*phC", MCOUNT_INSN_SIZE, ins);

It works, thanks for your suggestion. I'll send a new version patch.

-- Zheng Yejian