[v2] ftrace: fix incorrect hash size in register_ftrace_direct()

[PATCH bpf v2] ftrace: fix incorrect hash size in register_ftrace_direct()

Posted by Menglong Dong 8 months, 1 week ago

The maximum of the ftrace hash bits is made fls(32) in
register_ftrace_direct(), which seems illogical. So, we fix it by making
the max hash bits FTRACE_HASH_MAX_BITS instead.

Fixes: d05cb470663a ("ftrace: Fix modification of direct_function hash while in use")
Signed-off-by: Menglong Dong <dongml2@chinatelecom.cn>
---
v2:
- thanks for Steven's advice, we fix the problem by making the max hash
  bits FTRACE_HASH_MAX_BITS instead.
---
 kernel/trace/ftrace.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index 1a48aedb5255..d153ad13e0e0 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -5914,9 +5914,10 @@ int register_ftrace_direct(struct ftrace_ops *ops, unsigned long addr)
 
 	/* Make a copy hash to place the new and the old entries in */
 	size = hash->count + direct_functions->count;
-	if (size > 32)
-		size = 32;
-	new_hash = alloc_ftrace_hash(fls(size));
+	size = fls(size);
+	if (size > FTRACE_HASH_MAX_BITS)
+		size = FTRACE_HASH_MAX_BITS;
+	new_hash = alloc_ftrace_hash(size);
 	if (!new_hash)
 		goto out_unlock;
 
-- 
2.39.5

Re: [PATCH bpf v2] ftrace: fix incorrect hash size in register_ftrace_direct()

Posted by Masami Hiramatsu (Google) 8 months, 1 week ago

On Sun, 13 Apr 2025 09:44:44 +0800
Menglong Dong <menglong8.dong@gmail.com> wrote:

> The maximum of the ftrace hash bits is made fls(32) in
> register_ftrace_direct(), which seems illogical. So, we fix it by making
> the max hash bits FTRACE_HASH_MAX_BITS instead.
> 

Loogs good to me.

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

Thanks!

> Fixes: d05cb470663a ("ftrace: Fix modification of direct_function hash while in use")
> Signed-off-by: Menglong Dong <dongml2@chinatelecom.cn>
> ---
> v2:
> - thanks for Steven's advice, we fix the problem by making the max hash
>   bits FTRACE_HASH_MAX_BITS instead.
> ---
>  kernel/trace/ftrace.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> index 1a48aedb5255..d153ad13e0e0 100644
> --- a/kernel/trace/ftrace.c
> +++ b/kernel/trace/ftrace.c
> @@ -5914,9 +5914,10 @@ int register_ftrace_direct(struct ftrace_ops *ops, unsigned long addr)
>  
>  	/* Make a copy hash to place the new and the old entries in */
>  	size = hash->count + direct_functions->count;
> -	if (size > 32)
> -		size = 32;
> -	new_hash = alloc_ftrace_hash(fls(size));
> +	size = fls(size);
> +	if (size > FTRACE_HASH_MAX_BITS)
> +		size = FTRACE_HASH_MAX_BITS;
> +	new_hash = alloc_ftrace_hash(size);
>  	if (!new_hash)
>  		goto out_unlock;
>  
> -- 
> 2.39.5
> 
> 


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

Re: [PATCH bpf v2] ftrace: fix incorrect hash size in register_ftrace_direct()

Posted by Andrii Nakryiko 8 months, 1 week ago

On Mon, Apr 14, 2025 at 12:05 AM Masami Hiramatsu <mhiramat@kernel.org> wrote:
>
> On Sun, 13 Apr 2025 09:44:44 +0800
> Menglong Dong <menglong8.dong@gmail.com> wrote:
>
> > The maximum of the ftrace hash bits is made fls(32) in
> > register_ftrace_direct(), which seems illogical. So, we fix it by making
> > the max hash bits FTRACE_HASH_MAX_BITS instead.
> >
>
> Loogs good to me.
>
> Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
>
> Thanks!
>

I'm a bit confused by the "[PATCH bpf]" prefix... This fix doesn't
seem to be BPF-related, so I'm not sure why it would go through the
bpf tree. I presume Masami or Steven will route it through their tree,
is that right?


> > Fixes: d05cb470663a ("ftrace: Fix modification of direct_function hash while in use")
> > Signed-off-by: Menglong Dong <dongml2@chinatelecom.cn>
> > ---
> > v2:
> > - thanks for Steven's advice, we fix the problem by making the max hash
> >   bits FTRACE_HASH_MAX_BITS instead.
> > ---
> >  kernel/trace/ftrace.c | 7 ++++---
> >  1 file changed, 4 insertions(+), 3 deletions(-)
> >
> > diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> > index 1a48aedb5255..d153ad13e0e0 100644
> > --- a/kernel/trace/ftrace.c
> > +++ b/kernel/trace/ftrace.c
> > @@ -5914,9 +5914,10 @@ int register_ftrace_direct(struct ftrace_ops *ops, unsigned long addr)
> >
> >       /* Make a copy hash to place the new and the old entries in */
> >       size = hash->count + direct_functions->count;
> > -     if (size > 32)
> > -             size = 32;
> > -     new_hash = alloc_ftrace_hash(fls(size));
> > +     size = fls(size);
> > +     if (size > FTRACE_HASH_MAX_BITS)
> > +             size = FTRACE_HASH_MAX_BITS;
> > +     new_hash = alloc_ftrace_hash(size);
> >       if (!new_hash)
> >               goto out_unlock;
> >
> > --
> > 2.39.5
> >
> >
>
>
> --
> Masami Hiramatsu (Google) <mhiramat@kernel.org>
>

Re: [PATCH bpf v2] ftrace: fix incorrect hash size in register_ftrace_direct()

Posted by Menglong Dong 8 months, 1 week ago

On Wed, Apr 16, 2025 at 7:14 AM Andrii Nakryiko
<andrii.nakryiko@gmail.com> wrote:
>
> On Mon, Apr 14, 2025 at 12:05 AM Masami Hiramatsu <mhiramat@kernel.org> wrote:
> >
> > On Sun, 13 Apr 2025 09:44:44 +0800
> > Menglong Dong <menglong8.dong@gmail.com> wrote:
> >
> > > The maximum of the ftrace hash bits is made fls(32) in
> > > register_ftrace_direct(), which seems illogical. So, we fix it by making
> > > the max hash bits FTRACE_HASH_MAX_BITS instead.
> > >
> >
> > Loogs good to me.
> >
> > Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> >
> > Thanks!
> >
>
> I'm a bit confused by the "[PATCH bpf]" prefix... This fix doesn't
> seem to be BPF-related, so I'm not sure why it would go through the
> bpf tree. I presume Masami or Steven will route it through their tree,
> is that right?
>

Sorry about the confusing......I throught the register_ftrace_direct()
is mainly used by BPF, and it should go to the ftrace tree :/

Thanks!
Menglong Dong

>
> > > Fixes: d05cb470663a ("ftrace: Fix modification of direct_function hash while in use")
> > > Signed-off-by: Menglong Dong <dongml2@chinatelecom.cn>
> > > ---
> > > v2:
> > > - thanks for Steven's advice, we fix the problem by making the max hash
> > >   bits FTRACE_HASH_MAX_BITS instead.
> > > ---
> > >  kernel/trace/ftrace.c | 7 ++++---
> > >  1 file changed, 4 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> > > index 1a48aedb5255..d153ad13e0e0 100644
> > > --- a/kernel/trace/ftrace.c
> > > +++ b/kernel/trace/ftrace.c
> > > @@ -5914,9 +5914,10 @@ int register_ftrace_direct(struct ftrace_ops *ops, unsigned long addr)
> > >
> > >       /* Make a copy hash to place the new and the old entries in */
> > >       size = hash->count + direct_functions->count;
> > > -     if (size > 32)
> > > -             size = 32;
> > > -     new_hash = alloc_ftrace_hash(fls(size));
> > > +     size = fls(size);
> > > +     if (size > FTRACE_HASH_MAX_BITS)
> > > +             size = FTRACE_HASH_MAX_BITS;
> > > +     new_hash = alloc_ftrace_hash(size);
> > >       if (!new_hash)
> > >               goto out_unlock;
> > >
> > > --
> > > 2.39.5
> > >
> > >
> >
> >
> > --
> > Masami Hiramatsu (Google) <mhiramat@kernel.org>
> >

Re: [PATCH bpf v2] ftrace: fix incorrect hash size in register_ftrace_direct()

Posted by Steven Rostedt 8 months, 1 week ago

On Tue, 15 Apr 2025 16:14:01 -0700
Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote:

> I'm a bit confused by the "[PATCH bpf]" prefix... This fix doesn't
> seem to be BPF-related, so I'm not sure why it would go through the
> bpf tree. I presume Masami or Steven will route it through their tree,
> is that right?

I can take this in my tree.

-- Steve