2019-09-27 18:07:11

by Stephen Kitt

[permalink] [raw]

Subject: [PATCH v2] clk/ti/adpll: allocate room for terminating null

---

The buffer allocated in ti_adpll_clk_get_name doesn't account for the
terminating null. This patch switches to ka_sprintf to avoid
overflowing.

Signed-off-by: Stephen Kitt <[email protected]>
---
drivers/clk/ti/adpll.c | 10 ++--------
1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/drivers/clk/ti/adpll.c b/drivers/clk/ti/adpll.c
index fdfb90058504..021cf9e2b4db 100644
--- a/drivers/clk/ti/adpll.c
+++ b/drivers/clk/ti/adpll.c
@@ -195,14 +195,8 @@ static const char *ti_adpll_clk_get_name(struct ti_adpll_data *d,
return NULL;
} else {
const char *base_name = "adpll";
- char *buf;
-
- buf = devm_kzalloc(d->dev, 8 + 1 + strlen(base_name) + 1 +
- strlen(postfix), GFP_KERNEL);
- if (!buf)
- return NULL;
- sprintf(buf, "%08lx.%s.%s", d->pa, base_name, postfix);
- name = buf;
+ name = devm_kasprintf(d->dev, GFP_KERNEL, "%08lx.%s.%s",
+ d->pa, base_name, postfix);
}

return name;
--
2.20.1

2019-09-27 18:19:33

by Stephen Kitt

[permalink] [raw]

Subject: Re: [PATCH v2] clk/ti/adpll: allocate room for terminating null

---

Le 27/09/2019 20:05, Stephen Kitt a écrit :
> The buffer allocated in ti_adpll_clk_get_name doesn't account for the
> terminating null. This patch switches to ka_sprintf to avoid

Aargh, devm_kasprintf of course...

> overflowing.
>
> Signed-off-by: Stephen Kitt <[email protected]>
> ---
> drivers/clk/ti/adpll.c | 10 ++--------
> 1 file changed, 2 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/clk/ti/adpll.c b/drivers/clk/ti/adpll.c
> index fdfb90058504..021cf9e2b4db 100644
> --- a/drivers/clk/ti/adpll.c
> +++ b/drivers/clk/ti/adpll.c
> @@ -195,14 +195,8 @@ static const char *ti_adpll_clk_get_name(struct
> ti_adpll_data *d,
> return NULL;
> } else {
> const char *base_name = "adpll";
> - char *buf;
> -
> - buf = devm_kzalloc(d->dev, 8 + 1 + strlen(base_name) + 1 +
> - strlen(postfix), GFP_KERNEL);
> - if (!buf)
> - return NULL;
> - sprintf(buf, "%08lx.%s.%s", d->pa, base_name, postfix);
> - name = buf;
> + name = devm_kasprintf(d->dev, GFP_KERNEL, "%08lx.%s.%s",
> + d->pa, base_name, postfix);
> }
>
> return name;

2019-10-18 20:38:13

by Stephen Boyd

[permalink] [raw]

Subject: Re: [PATCH v2] clk/ti/adpll: allocate room for terminating null

---

Quoting Stephen Kitt (2019-09-27 11:05:59)
> The buffer allocated in ti_adpll_clk_get_name doesn't account for the
> terminating null. This patch switches to ka_sprintf to avoid
> overflowing.
>
> Signed-off-by: Stephen Kitt <[email protected]>
> ---
> drivers/clk/ti/adpll.c | 10 ++--------
> 1 file changed, 2 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/clk/ti/adpll.c b/drivers/clk/ti/adpll.c
> index fdfb90058504..021cf9e2b4db 100644
> --- a/drivers/clk/ti/adpll.c
> +++ b/drivers/clk/ti/adpll.c
> @@ -195,14 +195,8 @@ static const char *ti_adpll_clk_get_name(struct ti_adpll_data *d,
> return NULL;
> } else {
> const char *base_name = "adpll";

This is used once.

> - char *buf;
> -
> - buf = devm_kzalloc(d->dev, 8 + 1 + strlen(base_name) + 1 +
> - strlen(postfix), GFP_KERNEL);
> - if (!buf)
> - return NULL;
> - sprintf(buf, "%08lx.%s.%s", d->pa, base_name, postfix);
> - name = buf;
> + name = devm_kasprintf(d->dev, GFP_KERNEL, "%08lx.%s.%s",

So why not make this "%08lx.adpll.%s"?

> + d->pa, base_name, postfix);
> }

2019-10-19 14:35:47

by Stephen Kitt

[permalink] [raw]

Subject: Re: [PATCH v2] clk/ti/adpll: allocate room for terminating null

---

On Thu, 17 Oct 2019 08:48:53 -0700, Stephen Boyd <[email protected]> wrote:
> Quoting Stephen Kitt (2019-09-27 11:05:59)
> > The buffer allocated in ti_adpll_clk_get_name doesn't account for the
> > terminating null. This patch switches to ka_sprintf to avoid
> > overflowing.
> >
> > Signed-off-by: Stephen Kitt <[email protected]>
> > ---
> > drivers/clk/ti/adpll.c | 10 ++--------
> > 1 file changed, 2 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/clk/ti/adpll.c b/drivers/clk/ti/adpll.c
> > index fdfb90058504..021cf9e2b4db 100644
> > --- a/drivers/clk/ti/adpll.c
> > +++ b/drivers/clk/ti/adpll.c
> > @@ -195,14 +195,8 @@ static const char *ti_adpll_clk_get_name(struct
> > ti_adpll_data *d, return NULL;
> > } else {
> > const char *base_name = "adpll";
>
> This is used once.
>
> > - char *buf;
> > -
> > - buf = devm_kzalloc(d->dev, 8 + 1 + strlen(base_name) + 1 +
> > - strlen(postfix), GFP_KERNEL);
> > - if (!buf)
> > - return NULL;
> > - sprintf(buf, "%08lx.%s.%s", d->pa, base_name, postfix);
> > - name = buf;
> > + name = devm_kasprintf(d->dev, GFP_KERNEL, "%08lx.%s.%s",
>
> So why not make this "%08lx.adpll.%s"?

Thanks for the review! I hesitated to do this because I thought the purely
formatting string "%08lx.%s.%s" made the resulting code easier to understand
than a combined "%08lx.adpll.%s". I’ll follow up with a v3 which merges the
"adpll" string into the format string.

Regards,

Stephen

---

Attachments:

(No filename) (849.00 B)
OpenPGP digital signature