[PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Yizhuo Zhai]

Inside function hideep_nvm_unlock(), variable "unmask_code" could
be uninitialized if hideep_pgm_r_reg() returns error, however, it
is used in the later if statement after an "and" operation, which
is potentially unsafe.

Signed-off-by: Yizhuo <yzhai003@ucr.edu>
---
 drivers/input/touchscreen/hideep.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/input/touchscreen/hideep.c
b/drivers/input/touchscreen/hideep.c
index ddad4a82a5e5..49b713ad4384 100644
--- a/drivers/input/touchscreen/hideep.c
+++ b/drivers/input/touchscreen/hideep.c
@@ -363,7 +363,7 @@ static int hideep_enter_pgm(struct hideep_ts *ts)

 static void hideep_nvm_unlock(struct hideep_ts *ts)
 {
-       u32 unmask_code;
+       u32 unmask_code = 0;

        hideep_pgm_w_reg(ts, HIDEEP_FLASH_CFG, HIDEEP_NVM_SFR_RPAGE);
        hideep_pgm_r_reg(ts, 0x0000000C, &unmask_code);
-- 
2.17.1

Re: [PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Dmitry Torokhov]

Hi Yizhuo,

On Tue, Jun 15, 2021 at 10:26:09AM -0700, Yizhuo Zhai wrote:
> Inside function hideep_nvm_unlock(), variable "unmask_code" could
> be uninitialized if hideep_pgm_r_reg() returns error, however, it
> is used in the later if statement after an "and" operation, which
> is potentially unsafe.

I do not think that simply initializing the variable makes the code
behave any better. If we want to fix this properly we need to check for
errors returned by hideep_pgm_r_reg() and hideep_pgm_w_reg() and exit
this function early, signalling the caller about errors.

> 
> Signed-off-by: Yizhuo <yzhai003@ucr.edu>
> ---
>  drivers/input/touchscreen/hideep.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/input/touchscreen/hideep.c
> b/drivers/input/touchscreen/hideep.c
> index ddad4a82a5e5..49b713ad4384 100644
> --- a/drivers/input/touchscreen/hideep.c
> +++ b/drivers/input/touchscreen/hideep.c
> @@ -363,7 +363,7 @@ static int hideep_enter_pgm(struct hideep_ts *ts)
> 
>  static void hideep_nvm_unlock(struct hideep_ts *ts)
>  {
> -       u32 unmask_code;
> +       u32 unmask_code = 0;
> 
>         hideep_pgm_w_reg(ts, HIDEEP_FLASH_CFG, HIDEEP_NVM_SFR_RPAGE);
>         hideep_pgm_r_reg(ts, 0x0000000C, &unmask_code);
> -- 
> 2.17.1

Thanks.

-- 
Dmitry

Re: [PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Yizhuo Zhai]

Hi Demitry:

Thanks for your quick response, following your advice, a careful way
is changing the return type of  "hideep_nvm_unlock()" from void to
int, and its caller "hideep_program_nvm()" also needs to add the
return check.

If this sounds ok, I would go ahead to modify the patch accordingly.

On Tue, Jun 15, 2021 at 11:15 AM Dmitry Torokhov
<dmitry.torokhov@gmail.com> wrote:
>
> Hi Yizhuo,
>
> On Tue, Jun 15, 2021 at 10:26:09AM -0700, Yizhuo Zhai wrote:
> > Inside function hideep_nvm_unlock(), variable "unmask_code" could
> > be uninitialized if hideep_pgm_r_reg() returns error, however, it
> > is used in the later if statement after an "and" operation, which
> > is potentially unsafe.
>
> I do not think that simply initializing the variable makes the code
> behave any better. If we want to fix this properly we need to check for
> errors returned by hideep_pgm_r_reg() and hideep_pgm_w_reg() and exit
> this function early, signalling the caller about errors.
>
> >
> > Signed-off-by: Yizhuo <yzhai003@ucr.edu>
> > ---
> >  drivers/input/touchscreen/hideep.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/input/touchscreen/hideep.c
> > b/drivers/input/touchscreen/hideep.c
> > index ddad4a82a5e5..49b713ad4384 100644
> > --- a/drivers/input/touchscreen/hideep.c
> > +++ b/drivers/input/touchscreen/hideep.c
> > @@ -363,7 +363,7 @@ static int hideep_enter_pgm(struct hideep_ts *ts)
> >
> >  static void hideep_nvm_unlock(struct hideep_ts *ts)
> >  {
> > -       u32 unmask_code;
> > +       u32 unmask_code = 0;
> >
> >         hideep_pgm_w_reg(ts, HIDEEP_FLASH_CFG, HIDEEP_NVM_SFR_RPAGE);
> >         hideep_pgm_r_reg(ts, 0x0000000C, &unmask_code);
> > --
> > 2.17.1
>
> Thanks.
>
> --
> Dmitry



-- 
Kind Regards,

Yizhuo Zhai

Computer Science, Graduate Student
University of California, Riverside

Re: [PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Dmitry Torokhov]

On Tue, Jun 15, 2021 at 11:57:36AM -0700, Yizhuo Zhai wrote:
> Hi Demitry:
> 
> Thanks for your quick response, following your advice, a careful way
> is changing the return type of  "hideep_nvm_unlock()" from void to
> int, and its caller "hideep_program_nvm()" also needs to add the
> return check.
> 
> If this sounds ok, I would go ahead to modify the patch accordingly.

Yes, this sounds right.

> 
> On Tue, Jun 15, 2021 at 11:15 AM Dmitry Torokhov
> <dmitry.torokhov@gmail.com> wrote:
> >
> > Hi Yizhuo,
> >
> > On Tue, Jun 15, 2021 at 10:26:09AM -0700, Yizhuo Zhai wrote:
> > > Inside function hideep_nvm_unlock(), variable "unmask_code" could
> > > be uninitialized if hideep_pgm_r_reg() returns error, however, it
> > > is used in the later if statement after an "and" operation, which
> > > is potentially unsafe.
> >
> > I do not think that simply initializing the variable makes the code
> > behave any better. If we want to fix this properly we need to check for
> > errors returned by hideep_pgm_r_reg() and hideep_pgm_w_reg() and exit
> > this function early, signalling the caller about errors.
> >
> > >
> > > Signed-off-by: Yizhuo <yzhai003@ucr.edu>
> > > ---
> > >  drivers/input/touchscreen/hideep.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/input/touchscreen/hideep.c
> > > b/drivers/input/touchscreen/hideep.c
> > > index ddad4a82a5e5..49b713ad4384 100644
> > > --- a/drivers/input/touchscreen/hideep.c
> > > +++ b/drivers/input/touchscreen/hideep.c
> > > @@ -363,7 +363,7 @@ static int hideep_enter_pgm(struct hideep_ts *ts)
> > >
> > >  static void hideep_nvm_unlock(struct hideep_ts *ts)
> > >  {
> > > -       u32 unmask_code;
> > > +       u32 unmask_code = 0;
> > >
> > >         hideep_pgm_w_reg(ts, HIDEEP_FLASH_CFG, HIDEEP_NVM_SFR_RPAGE);
> > >         hideep_pgm_r_reg(ts, 0x0000000C, &unmask_code);
> > > --
> > > 2.17.1
> >
> > Thanks.
> >
> > --
> > Dmitry
> 
> 
> 
> -- 
> Kind Regards,
> 
> Yizhuo Zhai
> 
> Computer Science, Graduate Student
> University of California, Riverside

-- 
Dmitry

Re: [PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Dmitry Torokhov]

On Wed, Jun 16, 2021 at 03:48:51PM -0700, Yizhuo Zhai wrote:
> Inside function hideep_nvm_unlock(), variable "unmask_code" could
> be uninitialized if hideep_pgm_r_reg() returns error, however, it
> is used in the later if statement after an "and" operation, which
> is potentially unsafe.

I think this is pretty sensible, but let's see if the original author
has some comments...

> 
> Signed-off-by: Yizhuo <yzhai003@ucr.edu>
> ---
>  drivers/input/touchscreen/hideep.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/input/touchscreen/hideep.c
> b/drivers/input/touchscreen/hideep.c
> index ddad4a82a5e5..f860a815b603 100644
> --- a/drivers/input/touchscreen/hideep.c
> +++ b/drivers/input/touchscreen/hideep.c
> @@ -364,9 +364,13 @@ static int hideep_enter_pgm(struct hideep_ts *ts)
>  static void hideep_nvm_unlock(struct hideep_ts *ts)
>  {
>         u32 unmask_code;
> +       int ret;
> 
>         hideep_pgm_w_reg(ts, HIDEEP_FLASH_CFG, HIDEEP_NVM_SFR_RPAGE);
> -       hideep_pgm_r_reg(ts, 0x0000000C, &unmask_code);
> +       ret = hideep_pgm_r_reg(ts, 0x0000000C, &unmask_code);
> +       if (ret)
> +               return ret;
> +
>         hideep_pgm_w_reg(ts, HIDEEP_FLASH_CFG, HIDEEP_NVM_DEFAULT_PAGE);
> 
>         /* make it unprotected code */
> @@ -462,7 +466,9 @@ static int hideep_program_nvm(struct hideep_ts *ts,
>         u32 addr = 0;
>         int error;
> 
> -       hideep_nvm_unlock(ts);
> +       error = hideep_nvm_unlock(ts);
> +       if (error)
> +               return error;
> 
>         while (ucode_len > 0) {
>                 xfer_len = min_t(size_t, ucode_len, HIDEEP_NVM_PAGE_SIZE);
> -- 
> 2.17.1

-- 
Dmitry

Re: [PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Dmitry Torokhov]

On Sat, Jun 19, 2021 at 10:10:37PM -0700, Dmitry Torokhov wrote:
> On Wed, Jun 16, 2021 at 03:48:51PM -0700, Yizhuo Zhai wrote:
> > Inside function hideep_nvm_unlock(), variable "unmask_code" could
> > be uninitialized if hideep_pgm_r_reg() returns error, however, it
> > is used in the later if statement after an "and" operation, which
> > is potentially unsafe.
> 
> I think this is pretty sensible, but let's see if the original author
> has some comments...

I guess not. Oh well...

Applied, thank you.

-- 
Dmitry

Re: [PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Dmitry Torokhov]

On Sat, Jun 19, 2021 at 10:26:40PM -0700, Dmitry Torokhov wrote:
> On Sat, Jun 19, 2021 at 10:10:37PM -0700, Dmitry Torokhov wrote:
> > On Wed, Jun 16, 2021 at 03:48:51PM -0700, Yizhuo Zhai wrote:
> > > Inside function hideep_nvm_unlock(), variable "unmask_code" could
> > > be uninitialized if hideep_pgm_r_reg() returns error, however, it
> > > is used in the later if statement after an "and" operation, which
> > > is potentially unsafe.
> > 
> > I think this is pretty sensible, but let's see if the original author
> > has some comments...
> 
> I guess not. Oh well...
> 
> Applied, thank you.

Note that I had to make some changes to make it compile. Please next
time try building your changes before posting them,

Thanks.

-- 
Dmitry

Re: [PATCH] Input: hideep - fix the uninitialized use in hideep_nvm_unlock()

[Yizhuo Zhai]

Dimitry:

Sorry for the inconvenience, I would build the changes next time.
Thanks for your help : )

On Sat, Jun 19, 2021 at 10:35 PM Dmitry Torokhov
<dmitry.torokhov@gmail.com> wrote:
>
> On Sat, Jun 19, 2021 at 10:26:40PM -0700, Dmitry Torokhov wrote:
> > On Sat, Jun 19, 2021 at 10:10:37PM -0700, Dmitry Torokhov wrote:
> > > On Wed, Jun 16, 2021 at 03:48:51PM -0700, Yizhuo Zhai wrote:
> > > > Inside function hideep_nvm_unlock(), variable "unmask_code" could
> > > > be uninitialized if hideep_pgm_r_reg() returns error, however, it
> > > > is used in the later if statement after an "and" operation, which
> > > > is potentially unsafe.
> > >
> > > I think this is pretty sensible, but let's see if the original author
> > > has some comments...
> >
> > I guess not. Oh well...
> >
> > Applied, thank you.
>
> Note that I had to make some changes to make it compile. Please next
> time try building your changes before posting them,
>
> Thanks.
>
> --
> Dmitry



-- 
Kind Regards,

Yizhuo Zhai

Computer Science, Graduate Student
University of California, Riverside