2020-02-20 06:23:09

by Kees Cook

[permalink] [raw]
Subject: [PATCH] drm/edid: Distribute switch variables for initialization

Variables declared in a switch statement before any case statements
cannot be automatically initialized with compiler instrumentation (as
they are not part of any execution flow). With GCC's proposed automatic
stack variable initialization feature, this triggers a warning (and they
don't get initialized). Clang's automatic stack variable initialization
(via CONFIG_INIT_STACK_ALL=y) doesn't throw a warning, but it also
doesn't initialize such variables[1]. Note that these warnings (or silent
skipping) happen before the dead-store elimination optimization phase,
so even when the automatic initializations are later elided in favor of
direct initializations, the warnings remain.

To avoid these problems, move such variables into the "case" where
they're used or lift them up into the main function body.

drivers/gpu/drm/drm_edid.c: In function ‘drm_edid_to_eld’:
drivers/gpu/drm/drm_edid.c:4395:9: warning: statement will never be executed [-Wswitch-unreachable]
4395 | int sad_count;
| ^~~~~~~~~

[1] https://bugs.llvm.org/show_bug.cgi?id=44916

Signed-off-by: Kees Cook <[email protected]>
---
drivers/gpu/drm/drm_edid.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 805fb004c8eb..2941b65b427f 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -4392,9 +4392,9 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
dbl = cea_db_payload_len(db);

switch (cea_db_tag(db)) {
- int sad_count;
+ case AUDIO_BLOCK: {

- case AUDIO_BLOCK:
+ int sad_count;
/* Audio Data Block, contains SADs */
sad_count = min(dbl / 3, 15 - total_sad_count);
if (sad_count >= 1)
@@ -4402,6 +4402,7 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
&db[1], sad_count * 3);
total_sad_count += sad_count;
break;
+ }
case SPEAKER_BLOCK:
/* Speaker Allocation Data Block */
if (dbl >= 1)


2020-03-03 04:40:10

by Kees Cook

[permalink] [raw]
Subject: Re: [PATCH] drm/edid: Distribute switch variables for initialization

On Wed, Feb 19, 2020 at 10:22:29PM -0800, Kees Cook wrote:
> Variables declared in a switch statement before any case statements
> cannot be automatically initialized with compiler instrumentation (as
> they are not part of any execution flow). With GCC's proposed automatic
> stack variable initialization feature, this triggers a warning (and they
> don't get initialized). Clang's automatic stack variable initialization
> (via CONFIG_INIT_STACK_ALL=y) doesn't throw a warning, but it also
> doesn't initialize such variables[1]. Note that these warnings (or silent
> skipping) happen before the dead-store elimination optimization phase,
> so even when the automatic initializations are later elided in favor of
> direct initializations, the warnings remain.
>
> To avoid these problems, move such variables into the "case" where
> they're used or lift them up into the main function body.
>
> drivers/gpu/drm/drm_edid.c: In function ‘drm_edid_to_eld’:
> drivers/gpu/drm/drm_edid.c:4395:9: warning: statement will never be executed [-Wswitch-unreachable]
> 4395 | int sad_count;
> | ^~~~~~~~~
>
> [1] https://bugs.llvm.org/show_bug.cgi?id=44916
>
> Signed-off-by: Kees Cook <[email protected]>

Ping. Can someone pick this up, please?

Thanks!

-Kees

> ---
> drivers/gpu/drm/drm_edid.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
> index 805fb004c8eb..2941b65b427f 100644
> --- a/drivers/gpu/drm/drm_edid.c
> +++ b/drivers/gpu/drm/drm_edid.c
> @@ -4392,9 +4392,9 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
> dbl = cea_db_payload_len(db);
>
> switch (cea_db_tag(db)) {
> - int sad_count;
> + case AUDIO_BLOCK: {
>
> - case AUDIO_BLOCK:
> + int sad_count;
> /* Audio Data Block, contains SADs */
> sad_count = min(dbl / 3, 15 - total_sad_count);
> if (sad_count >= 1)
> @@ -4402,6 +4402,7 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
> &db[1], sad_count * 3);
> total_sad_count += sad_count;
> break;
> + }
> case SPEAKER_BLOCK:
> /* Speaker Allocation Data Block */
> if (dbl >= 1)
>

--
Kees Cook

2020-03-03 08:22:52

by Daniel Vetter

[permalink] [raw]
Subject: Re: [PATCH] drm/edid: Distribute switch variables for initialization

On Tue, Mar 3, 2020 at 5:39 AM Kees Cook <[email protected]> wrote:
>
> On Wed, Feb 19, 2020 at 10:22:29PM -0800, Kees Cook wrote:
> > Variables declared in a switch statement before any case statements
> > cannot be automatically initialized with compiler instrumentation (as
> > they are not part of any execution flow). With GCC's proposed automatic
> > stack variable initialization feature, this triggers a warning (and they
> > don't get initialized). Clang's automatic stack variable initialization
> > (via CONFIG_INIT_STACK_ALL=y) doesn't throw a warning, but it also
> > doesn't initialize such variables[1]. Note that these warnings (or silent
> > skipping) happen before the dead-store elimination optimization phase,
> > so even when the automatic initializations are later elided in favor of
> > direct initializations, the warnings remain.
> >
> > To avoid these problems, move such variables into the "case" where
> > they're used or lift them up into the main function body.
> >
> > drivers/gpu/drm/drm_edid.c: In function ‘drm_edid_to_eld’:
> > drivers/gpu/drm/drm_edid.c:4395:9: warning: statement will never be executed [-Wswitch-unreachable]
> > 4395 | int sad_count;
> > | ^~~~~~~~~
> >
> > [1] https://bugs.llvm.org/show_bug.cgi?id=44916
> >
> > Signed-off-by: Kees Cook <[email protected]>
>
> Ping. Can someone pick this up, please?

Whatever the reasons, but your original patch didn't make it through
to dri-devel. Can you pls resubmit?

Thanks, Daniel

>
> Thanks!
>
> -Kees
>
> > ---
> > drivers/gpu/drm/drm_edid.c | 5 +++--
> > 1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
> > index 805fb004c8eb..2941b65b427f 100644
> > --- a/drivers/gpu/drm/drm_edid.c
> > +++ b/drivers/gpu/drm/drm_edid.c
> > @@ -4392,9 +4392,9 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
> > dbl = cea_db_payload_len(db);
> >
> > switch (cea_db_tag(db)) {
> > - int sad_count;
> > + case AUDIO_BLOCK: {
> >
> > - case AUDIO_BLOCK:
> > + int sad_count;
> > /* Audio Data Block, contains SADs */
> > sad_count = min(dbl / 3, 15 - total_sad_count);
> > if (sad_count >= 1)
> > @@ -4402,6 +4402,7 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
> > &db[1], sad_count * 3);
> > total_sad_count += sad_count;
> > break;
> > + }
> > case SPEAKER_BLOCK:
> > /* Speaker Allocation Data Block */
> > if (dbl >= 1)
> >
>
> --
> Kees Cook



--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch

2020-03-03 18:03:04

by Ville Syrjälä

[permalink] [raw]
Subject: Re: [PATCH] drm/edid: Distribute switch variables for initialization

On Mon, Mar 02, 2020 at 08:39:37PM -0800, Kees Cook wrote:
> On Wed, Feb 19, 2020 at 10:22:29PM -0800, Kees Cook wrote:
> > Variables declared in a switch statement before any case statements
> > cannot be automatically initialized with compiler instrumentation (as
> > they are not part of any execution flow). With GCC's proposed automatic
> > stack variable initialization feature, this triggers a warning (and they
> > don't get initialized). Clang's automatic stack variable initialization
> > (via CONFIG_INIT_STACK_ALL=y) doesn't throw a warning, but it also
> > doesn't initialize such variables[1]. Note that these warnings (or silent
> > skipping) happen before the dead-store elimination optimization phase,
> > so even when the automatic initializations are later elided in favor of
> > direct initializations, the warnings remain.
> >
> > To avoid these problems, move such variables into the "case" where
> > they're used or lift them up into the main function body.
> >
> > drivers/gpu/drm/drm_edid.c: In function ‘drm_edid_to_eld’:
> > drivers/gpu/drm/drm_edid.c:4395:9: warning: statement will never be executed [-Wswitch-unreachable]
> > 4395 | int sad_count;
> > | ^~~~~~~~~
> >
> > [1] https://bugs.llvm.org/show_bug.cgi?id=44916
> >
> > Signed-off-by: Kees Cook <[email protected]>
>
> Ping. Can someone pick this up, please?
>
> Thanks!
>
> -Kees
>
> > ---
> > drivers/gpu/drm/drm_edid.c | 5 +++--
> > 1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
> > index 805fb004c8eb..2941b65b427f 100644
> > --- a/drivers/gpu/drm/drm_edid.c
> > +++ b/drivers/gpu/drm/drm_edid.c
> > @@ -4392,9 +4392,9 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
> > dbl = cea_db_payload_len(db);
> >
> > switch (cea_db_tag(db)) {
> > - int sad_count;
> > + case AUDIO_BLOCK: {

I've never been a fan of {} inside switch statements. I'd just
move this one level up.

> >
> > - case AUDIO_BLOCK:
> > + int sad_count;
> > /* Audio Data Block, contains SADs */
> > sad_count = min(dbl / 3, 15 - total_sad_count);
> > if (sad_count >= 1)
> > @@ -4402,6 +4402,7 @@ static void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
> > &db[1], sad_count * 3);
> > total_sad_count += sad_count;
> > break;
> > + }
> > case SPEAKER_BLOCK:
> > /* Speaker Allocation Data Block */
> > if (dbl >= 1)
> >
>
> --
> Kees Cook
> _______________________________________________
> dri-devel mailing list
> [email protected]
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

--
Ville Syrjälä
Intel