[PATCH 1/2] tracing: fprobe: Remove unused local variable

[PATCH 1/2] tracing: fprobe: Remove unused local variable

[Masami Hiramatsu (Google)]

From: Masami Hiramatsu (Google) <mhiramat@kernel.org>

The 'ret' local variable in fprobe_remove_node_in_module() was used
for checking the error state in the loop, but commit dfe0d675df82
("tracing: fprobe: use rhltable for fprobe_ip_table") removed the loop.
So we don't need it anymore.

Fixes: dfe0d675df82 ("tracing: fprobe: use rhltable for fprobe_ip_table")
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
 kernel/trace/fprobe.c |    5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
index 6a205903b1ed..12ec194fdfed 100644
--- a/kernel/trace/fprobe.c
+++ b/kernel/trace/fprobe.c
@@ -457,8 +457,6 @@ static int fprobe_addr_list_add(struct fprobe_addr_list *alist, unsigned long ad
 static void fprobe_remove_node_in_module(struct module *mod, struct fprobe_hlist_node *node,
 					 struct fprobe_addr_list *alist)
 {
-	int ret = 0;
-
 	if (!within_module(node->addr, mod))
 		return;
 	if (delete_fprobe_node(node))
@@ -467,8 +465,7 @@ static void fprobe_remove_node_in_module(struct module *mod, struct fprobe_hlist
 	 * If failed to update alist, just continue to update hlist.
 	 * Therefore, at list user handler will not hit anymore.
 	 */
-	if (!ret)
-		ret = fprobe_addr_list_add(alist, node->addr);
+	fprobe_addr_list_add(alist, node->addr);
 }
 
 /* Handle module unloading to manage fprobe_ip_table. */

[PATCH 2/2] tracing: fprobe: Fix to remove recorded module addresses from filter

[Masami Hiramatsu (Google)]

From: Masami Hiramatsu (Google) <mhiramat@kernel.org>

Even if there is a memory allocation failure in fprobe_addr_list_add(),
there is a partial list of module addresses. So remove the recorded
addresses from filter if exists.
This also removes the redundant ret local variable.

Fixes: a3dc2983ca7b ("tracing: fprobe: Cleanup fprobe hash when module unloading")
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
 kernel/trace/fprobe.c |    7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
index 12ec194fdfed..95e43814b85b 100644
--- a/kernel/trace/fprobe.c
+++ b/kernel/trace/fprobe.c
@@ -434,8 +434,9 @@ static int fprobe_addr_list_add(struct fprobe_addr_list *alist, unsigned long ad
 {
 	unsigned long *addrs;
 
-	if (alist->index >= alist->size)
-		return -ENOMEM;
+	/* Previously we failed to expand the list. */
+	if (alist->index == alist->size)
+		return -ENOSPC;
 
 	alist->addrs[alist->index++] = addr;
 	if (alist->index < alist->size)
@@ -497,7 +498,7 @@ static int fprobe_module_callback(struct notifier_block *nb,
 	} while (node == ERR_PTR(-EAGAIN));
 	rhashtable_walk_exit(&iter);
 
-	if (alist.index < alist.size && alist.index > 0)
+	if (alist.index > 0)
 		ftrace_set_filter_ips(&fprobe_graph_ops.ops,
 				      alist.addrs, alist.index, 1, 0);
 	mutex_unlock(&fprobe_mutex);

Re: [PATCH 1/2] tracing: fprobe: Remove unused local variable

[menglong.dong]

On 2025/9/24 08:26 Masami Hiramatsu (Google) <mhiramat@kernel.org> write:
> From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> 
> The 'ret' local variable in fprobe_remove_node_in_module() was used
> for checking the error state in the loop, but commit dfe0d675df82
> ("tracing: fprobe: use rhltable for fprobe_ip_table") removed the loop.
> So we don't need it anymore.
> 
> Fixes: dfe0d675df82 ("tracing: fprobe: use rhltable for fprobe_ip_table")
> Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> ---
>  kernel/trace/fprobe.c |    5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
> 
> diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
> index 6a205903b1ed..12ec194fdfed 100644
> --- a/kernel/trace/fprobe.c
> +++ b/kernel/trace/fprobe.c
> @@ -457,8 +457,6 @@ static int fprobe_addr_list_add(struct fprobe_addr_list *alist, unsigned long ad
>  static void fprobe_remove_node_in_module(struct module *mod, struct fprobe_hlist_node *node,
>  					 struct fprobe_addr_list *alist)
>  {
> -	int ret = 0;

Acked-by: Menglong Dong <menglong8.dong@gmail.com>

Thanks~

> -
>  	if (!within_module(node->addr, mod))
>  		return;
>  	if (delete_fprobe_node(node))
> @@ -467,8 +465,7 @@ static void fprobe_remove_node_in_module(struct module *mod, struct fprobe_hlist
>  	 * If failed to update alist, just continue to update hlist.
>  	 * Therefore, at list user handler will not hit anymore.
>  	 */
> -	if (!ret)
> -		ret = fprobe_addr_list_add(alist, node->addr);
> +	fprobe_addr_list_add(alist, node->addr);
>  }
>  
>  /* Handle module unloading to manage fprobe_ip_table. */
> 
> 
> 

Re: [PATCH 2/2] tracing: fprobe: Fix to remove recorded module addresses from filter

[menglong.dong]

On 2025/9/24 08:26 Masami Hiramatsu (Google) <mhiramat@kernel.org> write:
> From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> 
> Even if there is a memory allocation failure in fprobe_addr_list_add(),
> there is a partial list of module addresses. So remove the recorded
> addresses from filter if exists.
> This also removes the redundant ret local variable.

I think this is what you do in the previous patch?

The rest looks good to me.

Reviewed-by: Menglong Dong <menglong8.dong@gmail.com>

(I think I should use Reviewed-by here, rather than Acked-by,
right?)

Thanks!
Menglong Dong

> 
> Fixes: a3dc2983ca7b ("tracing: fprobe: Cleanup fprobe hash when module unloading")
> Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> ---
>  kernel/trace/fprobe.c |    7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
> index 12ec194fdfed..95e43814b85b 100644
> --- a/kernel/trace/fprobe.c
> +++ b/kernel/trace/fprobe.c
> @@ -434,8 +434,9 @@ static int fprobe_addr_list_add(struct fprobe_addr_list *alist, unsigned long ad
>  {
>  	unsigned long *addrs;
>  
> -	if (alist->index >= alist->size)
> -		return -ENOMEM;
> +	/* Previously we failed to expand the list. */
> +	if (alist->index == alist->size)
> +		return -ENOSPC;
>  
>  	alist->addrs[alist->index++] = addr;
>  	if (alist->index < alist->size)
> @@ -497,7 +498,7 @@ static int fprobe_module_callback(struct notifier_block *nb,
>  	} while (node == ERR_PTR(-EAGAIN));
>  	rhashtable_walk_exit(&iter);
>  
> -	if (alist.index < alist.size && alist.index > 0)
> +	if (alist.index > 0)
>  		ftrace_set_filter_ips(&fprobe_graph_ops.ops,
>  				      alist.addrs, alist.index, 1, 0);
>  	mutex_unlock(&fprobe_mutex);
> 
> 
> 

Re: [PATCH 2/2] tracing: fprobe: Fix to remove recorded module addresses from filter

[menglong.dong]

On 2025/9/24 08:26 Masami Hiramatsu (Google) <mhiramat@kernel.org> write:
> From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> 
> Even if there is a memory allocation failure in fprobe_addr_list_add(),
> there is a partial list of module addresses. So remove the recorded
> addresses from filter if exists.
> This also removes the redundant ret local variable.
> 
> Fixes: a3dc2983ca7b ("tracing: fprobe: Cleanup fprobe hash when module unloading")
> Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> ---
>  kernel/trace/fprobe.c |    7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)

Hi, Masami. Should I send the V2 of the patch:

  tracing: fprobe: optimization for entry only case

after this series applied?

Thanks!
Menglong Dong

> 
> diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
> index 12ec194fdfed..95e43814b85b 100644
> --- a/kernel/trace/fprobe.c
> +++ b/kernel/trace/fprobe.c
> @@ -434,8 +434,9 @@ static int fprobe_addr_list_add(struct fprobe_addr_list *alist, unsigned long ad
>  {
>  	unsigned long *addrs;
>  
> -	if (alist->index >= alist->size)
> -		return -ENOMEM;
> +	/* Previously we failed to expand the list. */
> +	if (alist->index == alist->size)
> +		return -ENOSPC;
>  
>  	alist->addrs[alist->index++] = addr;
>  	if (alist->index < alist->size)
> @@ -497,7 +498,7 @@ static int fprobe_module_callback(struct notifier_block *nb,
>  	} while (node == ERR_PTR(-EAGAIN));
>  	rhashtable_walk_exit(&iter);
>  
> -	if (alist.index < alist.size && alist.index > 0)
> +	if (alist.index > 0)
>  		ftrace_set_filter_ips(&fprobe_graph_ops.ops,
>  				      alist.addrs, alist.index, 1, 0);
>  	mutex_unlock(&fprobe_mutex);
> 
> 
> 

Re: [PATCH 2/2] tracing: fprobe: Fix to remove recorded module addresses from filter

[Masami Hiramatsu]

On Wed, 24 Sep 2025 11:49:39 +0800
menglong.dong@linux.dev wrote:

> On 2025/9/24 08:26 Masami Hiramatsu (Google) <mhiramat@kernel.org> write:
> > From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> > 
> > Even if there is a memory allocation failure in fprobe_addr_list_add(),
> > there is a partial list of module addresses. So remove the recorded
> > addresses from filter if exists.
> > This also removes the redundant ret local variable.
> > 
> > Fixes: a3dc2983ca7b ("tracing: fprobe: Cleanup fprobe hash when module unloading")
> > Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> > ---
> >  kernel/trace/fprobe.c |    7 ++++---
> >  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> Hi, Masami. Should I send the V2 of the patch:
> 
>   tracing: fprobe: optimization for entry only case
> 
> after this series applied?

Yeah, I'll push this [2/2] to stable tonight. and
push [1/2] to probes/for-next too. I'll rebase probes/for-next
so that it will have both patch.

Thank you,

> 
> Thanks!
> Menglong Dong
> 
> > 
> > diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
> > index 12ec194fdfed..95e43814b85b 100644
> > --- a/kernel/trace/fprobe.c
> > +++ b/kernel/trace/fprobe.c
> > @@ -434,8 +434,9 @@ static int fprobe_addr_list_add(struct fprobe_addr_list *alist, unsigned long ad
> >  {
> >  	unsigned long *addrs;
> >  
> > -	if (alist->index >= alist->size)
> > -		return -ENOMEM;
> > +	/* Previously we failed to expand the list. */
> > +	if (alist->index == alist->size)
> > +		return -ENOSPC;
> >  
> >  	alist->addrs[alist->index++] = addr;
> >  	if (alist->index < alist->size)
> > @@ -497,7 +498,7 @@ static int fprobe_module_callback(struct notifier_block *nb,
> >  	} while (node == ERR_PTR(-EAGAIN));
> >  	rhashtable_walk_exit(&iter);
> >  
> > -	if (alist.index < alist.size && alist.index > 0)
> > +	if (alist.index > 0)
> >  		ftrace_set_filter_ips(&fprobe_graph_ops.ops,
> >  				      alist.addrs, alist.index, 1, 0);
> >  	mutex_unlock(&fprobe_mutex);
> > 
> > 
> > 
> 
> 
> 
> 


-- 
Masami Hiramatsu (Google) <mhiramat@kernel.org>