drm/loongson: Fix framebuffer selection in async cursor updates
Subsystems (4)
PATCH 1/1 drm/loongson: Fix framebuffer selection in async cursor updates
An asynchronous cursor update swaps the new framebuffer into the live plane state. However, the driver reads both old_fb and new_fb from the live state before this swap. When an enabled cursor changes buffers, the address-update condition is therefore false and the hardware continues scanning the previous buffer.
Read new_fb from the new plane state before swapping the framebuffer references. This updates the hardware address when the framebuffer changes while retaining the existing behavior for cursor movement.
Verified on a Loongson 3A5000 with LS7A2000 using 32x32 and 64x64 legacy cursor updates. Callback counters and read-only register observations confirmed the asynchronous path, the stale address before the change, and the new address afterward. Same-framebuffer movement and cursor disable/re-enable also passed, with the primary framebuffer and mode unchanged.
AI assistance was used to investigate the existing report, prepare the fix and test tools, run the build and hardware comparison, and draft this message.
--- drivers/gpu/drm/loongson/lsdc_plane.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/loongson/lsdc_plane.c +1 -1
@@ -230,7 +230,7 @@ static void lsdc_cursor_plane_atomic_async_update(struct drm_plane *plane, new_state =drm_atomic_get_new_plane_state (state, plane); - new_fb = plane->state->fb; + new_fb = new_state->fb; plane->state->crtc_x = new_state->crtc_x; plane->state->crtc_y = new_state->crtc_y;
base-commit: bca45af5998a05f34b13a2ef11e639bac9c62643
Signature · 1 line
2.43.0
Icenowy Zheng
reviewer(Cc'ing drm-misc maintainers)
Quoted 48 lines · 在 2026-10-03六的 09:00 +0000,Evanshenf写道:
An asynchronous cursor update swaps the new framebuffer into the live plane state. However, the driver reads both
old_fbandnew_fbfrom the
live state before this swap. When an enabled cursor changes buffers, the
address-update condition is therefore false and the hardware continues
scanning the previous buffer.Read
new_fbfrom the new plane state before swapping the framebuffer references. This updates the hardware address when the framebuffer changes while retaining the existing behavior for cursor movement.Verified on a Loongson 3A5000 with LS7A2000 using 32x32 and 64x64 legacy
cursor updates. Callback counters and read-only register observations confirmed the asynchronous path, the stale address before the change, and the new address afterward. Same-framebuffer movement and cursor disable/re-enable also passed, with the primary framebuffer and mode unchanged.AI assistance was used to investigate the existing report, prepare the
fix and test tools, run the build and hardware comparison, and draft this message.Reported-by kernel test robot <[email protected]>Closes:
https://lore.kernel.org/r/[email protected]/Fixes f39db26c5428 ("drm: Add kms driver for loongson displaycontroller")
---
drivers/gpu/drm/loongson/lsdc_plane.c | 2 +-1 file changed, 1 insertion(+), 1 deletion(-)drivers/gpu/drm/loongson/lsdc_plane.c +0
-0@@ -230,7 +230,7 @@ static void lsdc_cursor_plane_atomic_async_update(struct drm_plane *plane,
new_state=drm_atomic_get_new_plane_state(state, plane);
drivers/gpu/drm/loongson/lsdc_plane.c static void lsdc_cursor_plane_atomic_async_update(struct drm_plane *plane, · line 233
Quoted 2 lines
drivers/gpu/drm/loongson/lsdc_plane.c +1
-1- new_fb = plane->state->fb; + new_fb = new_state->fb;
This change looks okay, although I don't know whether the atomic_async_update() hook should be implemented at all now.
``` Reviewed-by: Icenowy Zheng <[email protected]> ```
However, please wait for further clarification of the function itself from drm-misc maintainers.
Thanks,
Icenowy
Quoted 5 lines
plane->state->crtc_x=new_state->crtc_x;
plane->state->crtc_y=new_state->crtc_y;base-commit:
bca45af5998a05f34b13a2ef11e639bac9c62643
The fix is correct.
Thanks,
Qianhai Wu
Quoted 72 lines · Icenowy Zheng
(Cc'ing drm-misc maintainers)
Quoted 44 lines · 在 2026-10-03六的 09:00 +0000,Evanshenf写道:
An asynchronous cursor update swaps the new framebuffer into the live plane state. However, the driver reads both
old_fbandnew_fbfrom the
live state before this swap. When an enabled cursor changes buffers, the
address-update condition is therefore false and the hardware continues
scanning the previous buffer.Read
new_fbfrom the new plane state before swapping the framebuffer references. This updates the hardware address when the framebuffer changes while retaining the existing behavior for cursor movement.Verified on a Loongson 3A5000 with LS7A2000 using 32x32 and 64x64 legacy
cursor updates. Callback counters and read-only register observations confirmed the asynchronous path, the stale address before the change, and the new address afterward. Same-framebuffer movement and cursor disable/re-enable also passed, with the primary framebuffer and mode unchanged.AI assistance was used to investigate the existing report, prepare the
fix and test tools, run the build and hardware comparison, and draft this message.Reported-by kernel test robot <[email protected]>Closes:
https://lore.kernel.org/r/[email protected]/Fixes f39db26c5428 ("drm: Add kms driver for loongson displaycontroller")
--- drivers/gpu/drm/loongson/lsdc_plane.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/gpu/drm/loongson/lsdc_plane.c +0
-0drivers/gpu/drm/loongson/lsdc_plane.c static void · line 234
Quoted 6 lines
drivers/gpu/drm/loongson/lsdc_plane.c +1
-1@@ -230,7 +230,7 @@ static void lsdc_cursor_plane_atomic_async_update(struct drm_plane *plane, new_state = drm_atomic_get_new_plane_state(state, plane); - new_fb = plane->state->fb; + new_fb = new_state->fb;This change looks okay, although I don't know whether the
atomic_async_update()hook should be implemented at all now.``` Reviewed-by: Icenowy Zheng <[email protected]> ```However, please wait for further clarification of the function itself from drm-misc maintainers.
Thanks,
IcenowyQuoted 5 lines
drivers/gpu/drm/loongson/lsdc_plane.c +0
-0plane->state->crtc_x = new_state->crtc_x; plane->state->crtc_y = new_state->crtc_y;base-commit:
bca45af5998a05f34b13a2ef11e639bac9c62643
Quoted 22 lines · Evanshenf
An asynchronous cursor update swaps the new framebuffer into the live plane state. However, the driver reads both
old_fbandnew_fbfrom the live state before this swap. When an enabled cursor changes buffers, the address-update condition is therefore false and the hardware continues scanning the previous buffer.Read
new_fbfrom the new plane state before swapping the framebuffer references. This updates the hardware address when the framebuffer changes while retaining the existing behavior for cursor movement.Verified on a Loongson 3A5000 with LS7A2000 using 32x32 and 64x64 legacy cursor updates. Callback counters and read-only register observations confirmed the asynchronous path, the stale address before the change, and the new address afterward. Same-framebuffer movement and cursor disable/re-enable also passed, with the primary framebuffer and mode unchanged.
AI assistance was used to investigate the existing report, prepare the fix and test tools, run the build and hardware comparison, and draft this message.
Quoted 5 lines
Reported-by kernel test robot <[email protected]>Fixes f39db26c5428 ("drm: Add kms driver for loongson display controller")Assisted-by LLM
Quoted 23 lines
Signed-off-by Evanshenf <[email protected]>--- drivers/gpu/drm/loongson/lsdc_plane.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/gpu/drm/loongson/lsdc_plane.c +1
-1@@ -230,7 +230,7 @@ static void lsdc_cursor_plane_atomic_async_update(struct drm_plane *plane, new_state =drm_atomic_get_new_plane_state (state, plane); - new_fb = plane->state->fb; + new_fb = new_state->fb; plane->state->crtc_x = new_state->crtc_x; plane->state->crtc_y = new_state->crtc_y;base-commit:
bca45af5998a05f34b13a2ef11e639bac9c62643--
2.43.0
Quoted 26 lines · 在 2026-10-03六的 09:00 +0000,Evanshenf写道:
An asynchronous cursor update swaps the new framebuffer into the live plane state. However, the driver reads both
old_fbandnew_fbfrom the
live state before this swap. When an enabled cursor changes buffers, the
address-update condition is therefore false and the hardware continues
scanning the previous buffer.Read
new_fbfrom the new plane state before swapping the framebuffer references. This updates the hardware address when the framebuffer changes while retaining the existing behavior for cursor movement.Verified on a Loongson 3A5000 with LS7A2000 using 32x32 and 64x64 legacy
cursor updates. Callback counters and read-only register observations confirmed the asynchronous path, the stale address before the change, and the new address afterward. Same-framebuffer movement and cursor disable/re-enable also passed, with the primary framebuffer and mode unchanged.AI assistance was used to investigate the existing report, prepare the
fix and test tools, run the build and hardware comparison, and draft this message.
Quoted 9 lines
Reported-by kernel test robot <[email protected]>Closes:
https://lore.kernel.org/r/[email protected]/Fixes f39db26c5428 ("drm: Add kms driver for loongson displaycontroller")
Applied to drm-misc-fixes, Thanks for your contribution!
Icenowy
Quoted 21 lines
---
drivers/gpu/drm/loongson/lsdc_plane.c | 2 +-1 file changed, 1 insertion(+), 1 deletion(-)drivers/gpu/drm/loongson/lsdc_plane.c +0
-0@@ -230,7 +230,7 @@ static void lsdc_cursor_plane_atomic_async_update(struct drm_plane *plane,
new_state=drm_atomic_get_new_plane_state(state, plane);drivers/gpu/drm/loongson/lsdc_plane.c +1
-1- new_fb = plane->state->fb; + new_fb = new_state->fb;
plane->state->crtc_x=new_state->crtc_x;
plane->state->crtc_y=new_state->crtc_y;base-commit:
bca45af5998a05f34b13a2ef11e639bac9c62643