Skip to contentSkip to conversation

drm/hyperv: Added error message for fb size greater than allocated

PATCH Superseded Saurabh Sengar 2 participants 5 messages Apr 11, 2022 – Apr 12, 2022

Subsystems (5)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Saurabh Sengar

author maintainer
#0 lore raw

PATCH v3 1/1 drm/hyperv: Added error message for fb size greater than allocated

Added error message when the size of requested framebuffer is more than the allocated size by vmbus mmio region for framebuffer

---
v2 -> v3 : then -> than (typo fix)

 drivers/gpu/drm/hyperv/hyperv_drm_modeset.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +4 -1
index e82b815..6634818 100644
@@ -123,8 +123,11 @@ static int hyperv_pipe_check(struct drm_simple_display_pipe *pipe,
 	if (fb->format->format != DRM_FORMAT_XRGB8888)
 		return -EINVAL;
 
-	if (fb->pitches[0] * fb->height > hv->fb_size)
+	if (fb->pitches[0] * fb->height > hv->fb_size) {
+		drm_err(&hv->dev, "hv->hdev, fb size requested by process %s for %d X %d (pitch %d) is greater than allocated size %ld\n",
+		current->comm, fb->width, fb->height, fb->pitches[0], hv->fb_size);
 		return -EINVAL;
+	}
 
 	return 0;
 }
Signature · 1 line
1.8.3.1
maintainer
#1 lore raw
Quoted 5 lines

Subject: [PATCH v3] drm/hyperv: Added error message for fb size greater than allocated

Added error message when the size of requested framebuffer is more than the allocated size by vmbus mmio region for framebuffer

"Added" --> "Add"? My impression is that we don't use past tense in the Subject and the commit message. See
"git log drivers/gpu/drm/hyperv/hyperv_drm_modeset.c".

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c static int hyperv_pipe_check(struct · line 129

Quoted 11 lines
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -1
@@ -123,8 +123,11 @@ static int hyperv_pipe_check(struct
 drm_simple_display_pipe *pipe,
 	if (fb->format->format != DRM_FORMAT_XRGB8888)
 		return -EINVAL;
 
-	if (fb->pitches[0] * fb->height > hv->fb_size)
+	if (fb->pitches[0] * fb->height > hv->fb_size) {
+		drm_err(&hv->dev, "hv->hdev, fb size requested by process %s
 for %d X %d (pitch %d) is greater than allocated size %ld\n",

Should we use drm_err_ratelimited() instead of drm_err()?

The line exceeds 80 chars.

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c static int hyperv_pipe_check(struct drm_simple_display_pipe *pipe, · line 130

Quoted 3 lines
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -0
+		current->comm, fb->width, fb->height, fb->pitches[0], hv->fb_size);
 		return -EINVAL;
+	}

Maybe we can use the below:

	drm_err_ratelimited(&hv->dev, "%s: requested %dX%d (pitch %d) "
                     "exceeds fb_size %ld\n",
                     current->comm, fb->width, fb->height,
                     fb->pitches[0], hv->fb_size);

Note: the first chars of last 3 lines should align with the "&" in the same column. Please run "scripts/checkpatch.pl" against the patch.

author maintainer
#2 lore raw
Quoted 7 lines · Dexuan Cui
Quoted 5 lines

Subject: [PATCH v3] drm/hyperv: Added error message for fb size greater than allocated

Added error message when the size of requested framebuffer is more than the allocated size by vmbus mmio region for framebuffer

"Added" --> "Add"? My impression is that we don't use past tense in the

Ok.

Quoted 10 lines

Subject and the commit message. See
"git log drivers/gpu/drm/hyperv/hyperv_drm_modeset.c".

Quoted 7 lines
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +0 -0
@@ -123,8 +123,11 @@ static int hyperv_pipe_check(struct
 drm_simple_display_pipe *pipe,
 	if (fb->format->format != DRM_FORMAT_XRGB8888)
 		return -EINVAL;
 
Quoted 5 lines

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c static int hyperv_pipe_check(struct drm_simple_display_pipe *pipe,

Quoted 4 lines
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -1
-	if (fb->pitches[0] * fb->height > hv->fb_size)
+	if (fb->pitches[0] * fb->height > hv->fb_size) {
+		drm_err(&hv->dev, "hv->hdev, fb size requested by process %s
 for %d X %d (pitch %d) is greater than allocated size %ld\n",

Should we use drm_err_ratelimited() instead of drm_err()?

The error is not produced in good cases, for every resolution change which is violating this error should print once. I suggest having it print every time some application tries to violate the policy set at boot time. If we use ratelimit many resolutions error change will be suppressed. Please let me know your thoughts.

Quoted 2 lines

The line exceeds 80 chars.

At first I tried braking the line to respect 80 character boundary, but checkpatch.pl reported it as a problem.
And these new lines are suggested by checkpatch.pl itself.
Looks the recent rule realted to 80 charachters are changed. Ref : https://www.theregister.com/2020/06/01/linux_5_7/#:~:text=Linux%20kernel%20overlord%20Linus%20Torvalds,the%20topic%20of%20line%20lengths.

Quoted 6 lines

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c static int hyperv_pipe_check(struct drm_simple_display_pipe *pipe, · line 130

Quoted 4 lines
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -0
+		current->comm, fb->width, fb->height, fb->pitches[0], hv->fb_size);
 		return -EINVAL;
+	}

Maybe we can use the below:

Quoted 7 lines
	drm_err_ratelimited(&hv->dev, "%s: requested %dX%d (pitch %d) "
                     "exceeds fb_size %ld\n",
                     current->comm, fb->width, fb->height,
                     fb->pitches[0], hv->fb_size);

Note: the first chars of last 3 lines should align with the "&" in the same column. Please run "scripts/checkpatch.pl" against the patch.

I have tested checkpatch.pl before sending, for the current patch there is no problem from checkpatch.pl

maintainer
#3 lore raw
Quoted 11 lines

From: Saurabh Singh Sengar <[email protected]>
Sent: Monday, April 11, 2022 12:56 AM

Quoted 6 lines
Quoted 5 lines

...

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -1
-	if (fb->pitches[0] * fb->height > hv->fb_size)
+	if (fb->pitches[0] * fb->height > hv->fb_size) {
+		drm_err(&hv->dev, "hv->hdev, fb size requested by process %s

for %d X %d (pitch %d) is greater than allocated size %ld\n",

Should we use drm_err_ratelimited() instead of drm_err()?

The error is not produced in good cases, for every resolution change which is violating this error should print once.

Thanks for the clarification! Then drm_err() looks good to me.

Quoted 6 lines

I suggest having it print every time some application tries to violate the policy set at boot time.
If we use ratelimit many resolutions error change will be suppressed. Please let me know your thoughts.

Quoted 1 line
Quoted 7 lines
Quoted 1 line

The line exceeds 80 chars.

At first I tried braking the line to respect 80 character boundary, but checkpatch.pl reported it as a problem.
And these new lines are suggested by checkpatch.pl itself. Looks the recent rule realted to 80 charachters are changed. Ref :
...

Good to know! Thanks for sharing the link!

FYI, the default max_line_length in scripts/checkpatch.pl is 100 now:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=bdc48fa11e46f

"80-chars" is still preferred, but isn't a hard limit. I also noticed "never break user-visible strings such as printk messages ", so yes you're correct. It's perfectly fine to have a not-too-long string that exceeds 80 chars.

Quoted 10 lines
Quoted 10 lines

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c static int hyperv_pipe_check(struct drm_simple_display_pipe *pipe, · line 130

Quoted 3 lines
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -0
+		current->comm, fb->width, fb->height, fb->pitches[0], hv->fb_size);
 		return -EINVAL;
+	}

Maybe we can use the below:

	drm_err_ratelimited(&hv->dev, "%s: requested %dX%d (pitch %d) "
                     "exceeds fb_size %ld\n",
                     current->comm, fb->width, fb->height,
                     fb->pitches[0], hv->fb_size);
Quoted 5 lines
Quoted 2 lines

Note: the first chars of last 3 lines should align with the "&" in the same column. Please run "scripts/checkpatch.pl" against the patch.

I have tested checkpatch.pl before sending, for the current patch there is no problem from checkpatch.pl

The line is 138-char long, which seems a little longer to me :-) IMO we can make it shorter, e.g. be removing the part "hv->hdev as the "drm_err(&hv->dev," already tells us which device it's.

BTW, if we run the script with --strict, it reports the below:

# scripts/checkpatch.pl --strict 0001-drm-hyperv-Added-error-message-for-fb-size-greater-t.patch
CHECK: Alignment should match open parenthesis
#28: FILE: drivers/gpu/drm/hyperv/hyperv_drm_modeset.c:127:
+               drm_err(&hv->dev, "hv->hdev, fb size requested by process %s for %d X %d (pitch %d) is greater than allocated size %ld\n",
+               current->comm, fb->width, fb->height, fb->pitches[0], hv->fb_size);
author maintainer
#4 lore raw
Quoted 28 lines · Dexuan Cui
Quoted 11 lines

From: Saurabh Singh Sengar <[email protected]>
Sent: Monday, April 11, 2022 12:56 AM

Quoted 6 lines
Quoted 5 lines

...

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -1
-	if (fb->pitches[0] * fb->height > hv->fb_size)
+	if (fb->pitches[0] * fb->height > hv->fb_size) {
+		drm_err(&hv->dev, "hv->hdev, fb size requested by process %s

for %d X %d (pitch %d) is greater than allocated size %ld\n",

Should we use drm_err_ratelimited() instead of drm_err()?

The error is not produced in good cases, for every resolution change which is violating this error should print once.

Thanks for the clarification! Then drm_err() looks good to me.

Quoted 6 lines

I suggest having it print every time some application tries to violate the policy set at boot time.
If we use ratelimit many resolutions error change will be suppressed. Please let me know your thoughts.

Quoted 1 line
Quoted 7 lines
Quoted 1 line

The line exceeds 80 chars.

At first I tried braking the line to respect 80 character boundary, but checkpatch.pl reported it as a problem.
And these new lines are suggested by checkpatch.pl itself. Looks the recent rule realted to 80 charachters are changed. Ref :
...

Quoted 9 lines

Good to know! Thanks for sharing the link!

FYI, the default max_line_length in scripts/checkpatch.pl is 100 now:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=bdc48fa11e46f

"80-chars" is still preferred, but isn't a hard limit. I also noticed "never break user-visible strings such as printk messages ", so yes you're correct. It's perfectly fine to have a not-too-long string that exceeds 80 chars.

Good information ! thank you for digging this.

Quoted 13 lines
Quoted 13 lines
Quoted 12 lines

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c static int hyperv_pipe_check(struct drm_simple_display_pipe *pipe, · line 130

Quoted 3 lines
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -0
+		current->comm, fb->width, fb->height, fb->pitches[0], hv->fb_size);
 		return -EINVAL;
+	}

Maybe we can use the below:

	drm_err_ratelimited(&hv->dev, "%s: requested %dX%d (pitch %d) "
                     "exceeds fb_size %ld\n",
                     current->comm, fb->width, fb->height,
                     fb->pitches[0], hv->fb_size);

Note: the first chars of last 3 lines should align with the "&" in the same column. Please run "scripts/checkpatch.pl" against the patch.

Quoted 6 lines
Quoted 2 lines

I have tested checkpatch.pl before sending, for the current patch there is no problem from checkpatch.pl

The line is 138-char long, which seems a little longer to me :-) IMO we can make it shorter, e.g. be removing the part "hv->hdev as the "drm_err(&hv->dev," already tells us which device it's.

Ok, will make it shorter.

drivers/gpu/drm/hyperv/hyperv_drm_modeset.c static int hyperv_pipe_check(struct drm_simple_display_pipe *pipe, · line 128

Quoted 8 lines

BTW, if we run the script with --strict, it reports the below:

# scripts/checkpatch.pl --strict 0001-drm-hyperv-Added-error-message-for-fb-size-greater-t.patch
CHECK: Alignment should match open parenthesis
#28: FILE: drivers/gpu/drm/hyperv/hyperv_drm_modeset.c:127:
drivers/gpu/drm/hyperv/hyperv_drm_modeset.c +2 -0
+               drm_err(&hv->dev, "hv->hdev, fb size requested by process %s for %d X %d (pitch %d) is greater than allocated size %ld\n",
+               current->comm, fb->width, fb->height, fb->pitches[0], hv->fb_size);

Sure, will fix this.