Skip to contentSkip to conversation

fix handling of incorrect arguments by mipi_dsi_msleep

PATCH Applied Tejas Vipin 3 participants 9 messages Jun 12, 2024

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover fix handling of incorrect arguments by mipi_dsi_msleep

mipi_dsi_msleep is currently defined such that it treats ctx as an argument passed by value. In the case of ctx being passed by reference, it doesn't raise an error, but instead evaluates the resulting expression in an undesired manner. Since the majority of the usage of this function passes ctx by reference (similar to other functions), mipi_dsi_msleep can be modified to treat ctx as a pointer and do it correctly, and the other calls to this macro can be adjusted accordingly.

Tejas Vipin (2):
drm/panel : himax-hx83102: fix incorrect argument to mipi_dsi_msleep drm/mipi-dsi: fix handling of ctx in mipi_dsi_msleep

 drivers/gpu/drm/panel/panel-himax-hx83102.c | 6 +++---
 include/drm/drm_mipi_dsi.h                  | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)
Signature · 1 line
2.45.2
#1 lore raw

PATCH 1/2 drm/panel : himax-hx83102: fix incorrect argument to mipi_dsi_msleep

mipi_dsi_msleep should be modified to accept ctx as a pointer and the function call should be adjusted accordingly.

Fixes a2ab7cb169da3 ("drm/panel: himax-hx83102: use wrapped MIPI DCS functions")
---
 drivers/gpu/drm/panel/panel-himax-hx83102.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/gpu/drm/panel/panel-himax-hx83102.c +3 -3
index 6009a3fe1b8f..6e4b7e4644ce 100644
@@ -286,7 +286,7 @@ static int boe_nv110wum_init(struct hx83102 *ctx)
 	mipi_dsi_dcs_write_seq_multi(&dsi_ctx, HX83102_SETBANK, 0x00);
 	hx83102_enable_extended_cmds(&dsi_ctx, false);
 
-	mipi_dsi_msleep(dsi_ctx, 50);
+	mipi_dsi_msleep(&dsi_ctx, 50);
 
 	return dsi_ctx.accum_err;
 };
@@ -391,7 +391,7 @@ static int ivo_t109nw41_init(struct hx83102 *ctx)
 	mipi_dsi_dcs_write_seq_multi(&dsi_ctx, HX83102_SETBANK, 0x00);
 	hx83102_enable_extended_cmds(&dsi_ctx, false);
 
-	mipi_dsi_msleep(dsi_ctx, 60);
+	mipi_dsi_msleep(&dsi_ctx, 60);
 
 	return dsi_ctx.accum_err;
 };
@@ -538,7 +538,7 @@ static int hx83102_prepare(struct drm_panel *panel)
 	dsi_ctx.accum_err = ctx->desc->init(ctx);
 
 	mipi_dsi_dcs_exit_sleep_mode_multi(&dsi_ctx);
-	mipi_dsi_msleep(dsi_ctx, 120);
+	mipi_dsi_msleep(&dsi_ctx, 120);
 	mipi_dsi_dcs_set_display_on_multi(&dsi_ctx);
 	if (dsi_ctx.accum_err)
 		goto poweroff;
Signature · 1 line
2.45.2
Reviewed-by on v1 1/2
#2 lore raw

Re: [PATCH 1/2] drm/panel : himax-hx83102: fix incorrect argument to mipi_dsi_msleep

Hi,

Quoted 9 lines · Tejas Vipin

mipi_dsi_msleep should be modified to accept ctx as a pointer and the function call should be adjusted accordingly.

Fixes a2ab7cb169da3 ("drm/panel: himax-hx83102: use wrapped MIPI DCS functions")
---
 drivers/gpu/drm/panel/panel-himax-hx83102.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Thanks!

#3 lore raw

PATCH 2/2 drm/mipi-dsi: fix handling of ctx in mipi_dsi_msleep

In reply to Tejas Vipin

ctx would be better off treated as a pointer to account for most of its usage so far, and brackets should be added to account for operator precedence for correct evaluation.

Fixes f79d6d28d8fe7 ("drm/mipi-dsi: wrap more functions for streamline handling")
---
 include/drm/drm_mipi_dsi.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/drm/drm_mipi_dsi.h +1 -1
index bd5a0b6d0711..71d121aeef24 100644
@@ -293,7 +293,7 @@ ssize_t mipi_dsi_generic_read(struct mipi_dsi_device *dsi, const void *params,
 
 #define mipi_dsi_msleep(ctx, delay)	\
 	do {				\
-		if (!ctx.accum_err)	\
+		if (!(ctx)->accum_err)	\
 			msleep(delay);	\
 	} while (0)
 
Signature · 1 line
2.45.2
Reviewed-by on v1 2/2
#4 lore raw

Re: [PATCH 2/2] drm/mipi-dsi: fix handling of ctx in mipi_dsi_msleep

Hi,

Quoted 10 lines · Tejas Vipin

ctx would be better off treated as a pointer to account for most of its usage so far, and brackets should be added to account for operator precedence for correct evaluation.

Fixes f79d6d28d8fe7 ("drm/mipi-dsi: wrap more functions for streamline handling")
---
 include/drm/drm_mipi_dsi.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Yeah. Looking closer at the history, it looks like it was always intended to be a pointer since the first users all used it as a pointer.

I've also compile-tested all the panels currently using mipi_dsi_msleep().

Neil: Given that this is a correctness thing, I'd rather see this land sooner rather than later. If you agree, maybe you can land these two patches whenever you're comfortable with them?

-Doug

#5 lore raw

Re: [PATCH 2/2] drm/mipi-dsi: fix handling of ctx in mipi_dsi_msleep

Quoted 18 lines · Doug Anderson

Hi,

Quoted 10 lines · Tejas Vipin

ctx would be better off treated as a pointer to account for most of its usage so far, and brackets should be added to account for operator precedence for correct evaluation.

Fixes f79d6d28d8fe7 ("drm/mipi-dsi: wrap more functions for streamline handling")
---
  include/drm/drm_mipi_dsi.h | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

Yeah. Looking closer at the history, it looks like it was always intended to be a pointer since the first users all used it as a pointer.

Quoted 8 lines

I've also compile-tested all the panels currently using mipi_dsi_msleep().

Neil: Given that this is a correctness thing, I'd rather see this land sooner rather than later. If you agree, maybe you can land these two patches whenever you're comfortable with them?

Applying them, but inverting them, fix should go first.

Neil

Quoted 3 lines

-Doug

#6 lore raw

Re: [PATCH 2/2] drm/mipi-dsi: fix handling of ctx in mipi_dsi_msleep

Hi,

Quoted 23 lines · On Wed, Jun 12, 2024 at 7:34 AM <[email protected]> wrote:
Quoted 21 lines · Doug Anderson

Hi,

Quoted 10 lines · Tejas Vipin

ctx would be better off treated as a pointer to account for most of its usage so far, and brackets should be added to account for operator precedence for correct evaluation.

Fixes f79d6d28d8fe7 ("drm/mipi-dsi: wrap more functions for streamline handling")
---
  include/drm/drm_mipi_dsi.h | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

Yeah. Looking closer at the history, it looks like it was always intended to be a pointer since the first users all used it as a pointer.

Quoted 7 lines
Quoted 5 lines

I've also compile-tested all the panels currently using mipi_dsi_msleep().

Neil: Given that this is a correctness thing, I'd rather see this land sooner rather than later. If you agree, maybe you can land these two patches whenever you're comfortable with them?

Applying them, but inverting them, fix should go first.

Well, they're both fixes, and inverting them means that you get a compile failure across several panels if you happen to be bisecting and land on the first commit, but it doesn't really matter. I guess the compile failure is maybe a benefit given that they were not doing their delays properly... ;-)

-Doug

#7 lore raw

Re: [PATCH 2/2] drm/mipi-dsi: fix handling of ctx in mipi_dsi_msleep

Quoted 32 lines · Doug Anderson

Hi,

Quoted 29 lines · On Wed, Jun 12, 2024 at 7:34 AM <[email protected]> wrote:
Quoted 26 lines · Doug Anderson

Hi,

Quoted 10 lines · Tejas Vipin

ctx would be better off treated as a pointer to account for most of its usage so far, and brackets should be added to account for operator precedence for correct evaluation.

Fixes f79d6d28d8fe7 ("drm/mipi-dsi: wrap more functions for streamline handling")
---
   include/drm/drm_mipi_dsi.h | 2 +-
   1 file changed, 1 insertion(+), 1 deletion(-)

Yeah. Looking closer at the history, it looks like it was always intended to be a pointer since the first users all used it as a pointer.

I've also compile-tested all the panels currently using mipi_dsi_msleep().

Neil: Given that this is a correctness thing, I'd rather see this land sooner rather than later. If you agree, maybe you can land these two patches whenever you're comfortable with them?

Quoted 7 lines
Quoted 1 line

Applying them, but inverting them, fix should go first.

Well, they're both fixes, and inverting them means that you get a compile failure across several panels if you happen to be bisecting and land on the first commit, but it doesn't really matter. I guess the compile failure is maybe a benefit given that they were not doing their delays properly... ;-)

Yes, and thanksfully there's a fix for the build failure!

Quoted 2 lines

-Doug

maintainer Applied
#8 lore raw
In reply to Tejas Vipin

Hi,

Quoted 10 lines · Tejas Vipin

mipi_dsi_msleep is currently defined such that it treats ctx as an argument passed by value. In the case of ctx being passed by reference, it doesn't raise an error, but instead evaluates the resulting expression in an undesired manner. Since the majority of the usage of this function passes ctx by reference (similar to other functions), mipi_dsi_msleep can be modified to treat ctx as a pointer and do it correctly, and the other calls to this macro can be adjusted accordingly.

[...]

Thanks, Applied to https://gitlab.freedesktop.org/drm/misc/kernel.git (drm-misc-next)

[1/2] drm/panel : himax-hx83102: fix incorrect argument to mipi_dsi_msleep

      https://gitlab.freedesktop.org/drm/misc/kernel/-/commit/a13aaf157467e694a3824d81304106b58d4c20d6
[2/2] drm/mipi-dsi: fix handling of ctx in mipi_dsi_msleep
      https://gitlab.freedesktop.org/drm/misc/kernel/-/commit/66055636a146c435cd226fb5a334176304652f3c
Signature · 1 line
Neil