Here are some fixes port from rockchip_linux project[0],
Tested on rk3399 and rk3288 board.
[0]: https://github.com/rockchip-linux/kernel
Mark Yao (6):
drm/rockchip: vop: no need wait vblank on crtc enable
drm/rockchip: vop: fix iommu page fault when resume
drm/rockchip: vop: fix NV12 video display error
drm/rockchip: vop: round_up pitches to word align
drm/rockchip: vop: report error when check resource error
drm/rockchip: fix race with kms hotplug and fbdev
drivers/gpu/drm/rockchip/rockchip_drm_fb.c | 2 +-
drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 77 ++++++++---------------------
drivers/gpu/drm/rockchip/rockchip_drm_vop.h | 3 ++
3 files changed, 24 insertions(+), 58 deletions(-)
--
1.9.1
Iommu would get page fault with following path:
vop_disable:
1, disable all windows and set vop config done
2, vop enter to standy, all windows not works, but their registers
are not clean, when you read window's enable bit, may found the
window is enable.
vop_enable:
1, memcpy(vop->regsbak, vop->regs, len)
save current vop registers to vop->regsbak, then you can found
window is enable on regsbak.
2, VOP_WIN_SET(vop, win, gate, 1);
force enable window gate, but gate and enable are on same
hardware register, then window enable bit rewrite to vop hardware.
3, vop power on, and vop might try to scan destroyed buffer,
then iommu get page fault.
Move windows disable after vop regsbak restore, then vop regsbak mechanism
would keep tracing the modify, everything would be safe.
Signed-off-by: Mark Yao <[email protected]>
---
drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 33 +++++++++++++----------------
1 file changed, 15 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
index 0bfd563..0b5fd75 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
@@ -495,7 +495,7 @@ static void vop_line_flag_irq_disable(struct vop *vop)
static int vop_enable(struct drm_crtc *crtc)
{
struct vop *vop = to_vop(crtc);
- int ret;
+ int ret, i;
ret = pm_runtime_get_sync(vop->dev);
if (ret < 0) {
@@ -528,6 +528,20 @@ static int vop_enable(struct drm_crtc *crtc)
}
memcpy(vop->regs, vop->regsbak, vop->len);
+ /*
+ * We need to make sure that all windows are disabled before we
+ * enable the crtc. Otherwise we might try to scan from a destroyed
+ * buffer later.
+ */
+ for (i = 0; i < vop->data->win_size; i++) {
+ struct vop_win *vop_win = &vop->win[i];
+ const struct vop_win_data *win = vop_win->data;
+
+ spin_lock(&vop->reg_lock);
+ VOP_WIN_SET(vop, win, enable, 0);
+ spin_unlock(&vop->reg_lock);
+ }
+
vop_cfg_done(vop);
/*
@@ -561,28 +575,11 @@ static int vop_enable(struct drm_crtc *crtc)
static void vop_crtc_disable(struct drm_crtc *crtc)
{
struct vop *vop = to_vop(crtc);
- int i;
WARN_ON(vop->event);
rockchip_drm_psr_deactivate(&vop->crtc);
- /*
- * We need to make sure that all windows are disabled before we
- * disable that crtc. Otherwise we might try to scan from a destroyed
- * buffer later.
- */
- for (i = 0; i < vop->data->win_size; i++) {
- struct vop_win *vop_win = &vop->win[i];
- const struct vop_win_data *win = vop_win->data;
-
- spin_lock(&vop->reg_lock);
- VOP_WIN_SET(vop, win, enable, 0);
- spin_unlock(&vop->reg_lock);
- }
-
- vop_cfg_done(vop);
-
drm_crtc_vblank_off(crtc);
/*
--
1.9.1
fixup the scale calculation formula on the case
src_height == (dst_height/2).
Signed-off-by: Mark Yao <[email protected]>
---
drivers/gpu/drm/rockchip/rockchip_drm_vop.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
index af1091f..56bbd2e 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
@@ -299,6 +299,9 @@ static inline uint16_t scl_get_bili_dn_vskip(int src_h, int dst_h,
act_height = (src_h + vskiplines - 1) / vskiplines;
+ if (act_height == dst_h)
+ return GET_SCL_FT_BILI_DN(src_h, dst_h) / vskiplines;
+
return GET_SCL_FT_BILI_DN(act_height, dst_h);
}
--
1.9.1
VOP pitch register is word align, need align to word.
VOP_WIN0_VIR:
bit[31:16] win0_vir_stride_uv
Number of words of Win0 uv Virtual width
bit[15:0] win0_vir_width
Number of words of Win0 yrgb Virtual width
ARGB888 : win0_vir_width
RGB888 : (win0_vir_width*3/4) + (win0_vir_width%3)
RGB565 : ceil(win0_vir_width/2)
YUV : ceil(win0_vir_width/4)
Signed-off-by: Mark Yao <[email protected]>
---
drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
index 0b5fd75..fa0d9f7 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
@@ -756,7 +756,7 @@ static void vop_plane_atomic_update(struct drm_plane *plane,
spin_lock(&vop->reg_lock);
VOP_WIN_SET(vop, win, format, format);
- VOP_WIN_SET(vop, win, yrgb_vir, fb->pitches[0] >> 2);
+ VOP_WIN_SET(vop, win, yrgb_vir, DIV_ROUND_UP(fb->pitches[0], 4));
VOP_WIN_SET(vop, win, yrgb_mst, dma_addr);
if (is_yuv_support(fb->format->format)) {
int hsub = drm_format_horz_chroma_subsampling(fb->format->format);
@@ -770,7 +770,7 @@ static void vop_plane_atomic_update(struct drm_plane *plane,
offset += (src->y1 >> 16) * fb->pitches[1] / vsub;
dma_addr = rk_uv_obj->dma_addr + offset + fb->offsets[1];
- VOP_WIN_SET(vop, win, uv_vir, fb->pitches[1] >> 2);
+ VOP_WIN_SET(vop, win, uv_vir, DIV_ROUND_UP(fb->pitches[1], 4));
VOP_WIN_SET(vop, win, uv_mst, dma_addr);
}
--
1.9.1
The user would be confused while facing a error commit without
any error report.
Signed-off-by: Mark Yao <[email protected]>
---
drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
index fa0d9f7..999c2e0 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
@@ -674,8 +674,10 @@ static int vop_plane_atomic_check(struct drm_plane *plane,
* Src.x1 can be odd when do clip, but yuv plane start point
* need align with 2 pixel.
*/
- if (is_yuv_support(fb->format->format) && ((state->src.x1 >> 16) % 2))
+ if (is_yuv_support(fb->format->format) && ((state->src.x1 >> 16) % 2)) {
+ DRM_ERROR("Invalid Source: Yuv format not support odd xpos\n");
return -EINVAL;
+ }
return 0;
}
--
1.9.1
Since fb_helper is not a pointer on rockchip_drm_private, it's no
need to check pointer.
Kms hotplug event may race into fbdev helper initial, and fb_helper->dev
may be NULL pointer, that would cause the bug:
[ 0.735411] [00000200] *pgd=00000000f6ffe003, *pud=00000000f6ffe003, *pmd=0000000000000000
[ 0.736156] Internal error: Oops: 96000005 [#1] PREEMPT SMP
[ 0.736648] Modules linked in:
[ 0.736930] CPU: 2 PID: 20 Comm: kworker/2:0 Not tainted 4.4.41 #20
[ 0.737480] Hardware name: Rockchip RK3399 Board rev2 (BOX) (DT)
[ 0.738020] Workqueue: events cdn_dp_pd_event_work
[ 0.738447] task: ffffffc0f21f3100 ti: ffffffc0f2218000 task.ti: ffffffc0f2218000
[ 0.739109] PC is at mutex_lock+0x14/0x44
[ 0.739469] LR is at drm_fb_helper_hotplug_event+0x30/0x114
[ 0.756253] [<ffffff8008a344f4>] mutex_lock+0x14/0x44
[ 0.756260] [<ffffff8008445708>] drm_fb_helper_hotplug_event+0x30/0x114
[ 0.756271] [<ffffff8008473c84>] rockchip_drm_output_poll_changed+0x18/0x20
[ 0.756280] [<ffffff8008439fcc>] drm_kms_helper_hotplug_event+0x28/0x34
[ 0.756286] [<ffffff800846c444>] cdn_dp_pd_event_work+0x394/0x3c4
[ 0.756295] [<ffffff80080b2b38>] process_one_work+0x218/0x3e0
[ 0.756302] [<ffffff80080b3538>] worker_thread+0x2e8/0x404
[ 0.756308] [<ffffff80080b7e70>] kthread+0xe8/0xf0
[ 0.756316] [<ffffff8008082690>] ret_from_fork+0x10/0x40
Signed-off-by: Mark Yao <[email protected]>
---
drivers/gpu/drm/rockchip/rockchip_drm_fb.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
index 81f9548..e6bd0f4 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
@@ -170,7 +170,7 @@ static void rockchip_drm_output_poll_changed(struct drm_device *dev)
struct rockchip_drm_private *private = dev->dev_private;
struct drm_fb_helper *fb_helper = &private->fbdev_helper;
- if (fb_helper)
+ if (fb_helper->dev)
drm_fb_helper_hotplug_event(fb_helper);
}
--
1.9.1
Since atomic framework, crtc enable and disable are in pairs,
no need to wait vblank.
Signed-off-by: Mark Yao <[email protected]>
---
drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 36 -----------------------------
1 file changed, 36 deletions(-)
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
index 1d42049..0bfd563 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
@@ -893,42 +893,6 @@ static void vop_crtc_enable(struct drm_crtc *crtc)
return;
}
- /*
- * If dclk rate is zero, mean that scanout is stop,
- * we don't need wait any more.
- */
- if (clk_get_rate(vop->dclk)) {
- /*
- * Rk3288 vop timing register is immediately, when configure
- * display timing on display time, may cause tearing.
- *
- * Vop standby will take effect at end of current frame,
- * if dsp hold valid irq happen, it means standby complete.
- *
- * mode set:
- * standby and wait complete --> |----
- * | display time
- * |----
- * |---> dsp hold irq
- * configure display timing --> |
- * standby exit |
- * | new frame start.
- */
-
- reinit_completion(&vop->dsp_hold_completion);
- vop_dsp_hold_valid_irq_enable(vop);
-
- spin_lock(&vop->reg_lock);
-
- VOP_REG_SET(vop, common, standby, 1);
-
- spin_unlock(&vop->reg_lock);
-
- wait_for_completion(&vop->dsp_hold_completion);
-
- vop_dsp_hold_valid_irq_disable(vop);
- }
-
pin_pol = BIT(DCLK_INVERT);
pin_pol |= (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC) ?
BIT(HSYNC_POSITIVE) : 0;
--
1.9.1
On 31 July 2017 at 10:50, Mark Yao <[email protected]> wrote:
> Since fb_helper is not a pointer on rockchip_drm_private, it's no
> need to check pointer.
>
> Kms hotplug event may race into fbdev helper initial, and fb_helper->dev
> may be NULL pointer, that would cause the bug:
>
> [ 0.735411] [00000200] *pgd=00000000f6ffe003, *pud=00000000f6ffe003, *pmd=0000000000000000
> [ 0.736156] Internal error: Oops: 96000005 [#1] PREEMPT SMP
> [ 0.736648] Modules linked in:
> [ 0.736930] CPU: 2 PID: 20 Comm: kworker/2:0 Not tainted 4.4.41 #20
> [ 0.737480] Hardware name: Rockchip RK3399 Board rev2 (BOX) (DT)
> [ 0.738020] Workqueue: events cdn_dp_pd_event_work
> [ 0.738447] task: ffffffc0f21f3100 ti: ffffffc0f2218000 task.ti: ffffffc0f2218000
> [ 0.739109] PC is at mutex_lock+0x14/0x44
> [ 0.739469] LR is at drm_fb_helper_hotplug_event+0x30/0x114
> [ 0.756253] [<ffffff8008a344f4>] mutex_lock+0x14/0x44
> [ 0.756260] [<ffffff8008445708>] drm_fb_helper_hotplug_event+0x30/0x114
> [ 0.756271] [<ffffff8008473c84>] rockchip_drm_output_poll_changed+0x18/0x20
> [ 0.756280] [<ffffff8008439fcc>] drm_kms_helper_hotplug_event+0x28/0x34
> [ 0.756286] [<ffffff800846c444>] cdn_dp_pd_event_work+0x394/0x3c4
> [ 0.756295] [<ffffff80080b2b38>] process_one_work+0x218/0x3e0
> [ 0.756302] [<ffffff80080b3538>] worker_thread+0x2e8/0x404
> [ 0.756308] [<ffffff80080b7e70>] kthread+0xe8/0xf0
> [ 0.756316] [<ffffff8008082690>] ret_from_fork+0x10/0x40
>
> Signed-off-by: Mark Yao <[email protected]>
> ---
> drivers/gpu/drm/rockchip/rockchip_drm_fb.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
> index 81f9548..e6bd0f4 100644
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
> @@ -170,7 +170,7 @@ static void rockchip_drm_output_poll_changed(struct drm_device *dev)
> struct rockchip_drm_private *private = dev->dev_private;
> struct drm_fb_helper *fb_helper = &private->fbdev_helper;
>
> - if (fb_helper)
> + if (fb_helper->dev)
> drm_fb_helper_hotplug_event(fb_helper);
Food for thought:
Quick grep shows that no other drivers have such a ->dev check. Does
this mean that either the issue is rockchip specific?
If not, one could look into resolving the problem directly in drm core.
Or at least update the other users, so they don't stumble upon the problem?
HTH
Emil
On Mon, Jul 31, 2017 at 1:57 PM, Emil Velikov <[email protected]> wrote:
> On 31 July 2017 at 10:50, Mark Yao <[email protected]> wrote:
>> Since fb_helper is not a pointer on rockchip_drm_private, it's no
>> need to check pointer.
>>
>> Kms hotplug event may race into fbdev helper initial, and fb_helper->dev
>> may be NULL pointer, that would cause the bug:
>>
>> [ 0.735411] [00000200] *pgd=00000000f6ffe003, *pud=00000000f6ffe003, *pmd=0000000000000000
>> [ 0.736156] Internal error: Oops: 96000005 [#1] PREEMPT SMP
>> [ 0.736648] Modules linked in:
>> [ 0.736930] CPU: 2 PID: 20 Comm: kworker/2:0 Not tainted 4.4.41 #20
>> [ 0.737480] Hardware name: Rockchip RK3399 Board rev2 (BOX) (DT)
>> [ 0.738020] Workqueue: events cdn_dp_pd_event_work
>> [ 0.738447] task: ffffffc0f21f3100 ti: ffffffc0f2218000 task.ti: ffffffc0f2218000
>> [ 0.739109] PC is at mutex_lock+0x14/0x44
>> [ 0.739469] LR is at drm_fb_helper_hotplug_event+0x30/0x114
>> [ 0.756253] [<ffffff8008a344f4>] mutex_lock+0x14/0x44
>> [ 0.756260] [<ffffff8008445708>] drm_fb_helper_hotplug_event+0x30/0x114
>> [ 0.756271] [<ffffff8008473c84>] rockchip_drm_output_poll_changed+0x18/0x20
>> [ 0.756280] [<ffffff8008439fcc>] drm_kms_helper_hotplug_event+0x28/0x34
>> [ 0.756286] [<ffffff800846c444>] cdn_dp_pd_event_work+0x394/0x3c4
>> [ 0.756295] [<ffffff80080b2b38>] process_one_work+0x218/0x3e0
>> [ 0.756302] [<ffffff80080b3538>] worker_thread+0x2e8/0x404
>> [ 0.756308] [<ffffff80080b7e70>] kthread+0xe8/0xf0
>> [ 0.756316] [<ffffff8008082690>] ret_from_fork+0x10/0x40
>>
>> Signed-off-by: Mark Yao <[email protected]>
>> ---
>> drivers/gpu/drm/rockchip/rockchip_drm_fb.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
>> index 81f9548..e6bd0f4 100644
>> --- a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
>> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
>> @@ -170,7 +170,7 @@ static void rockchip_drm_output_poll_changed(struct drm_device *dev)
>> struct rockchip_drm_private *private = dev->dev_private;
>> struct drm_fb_helper *fb_helper = &private->fbdev_helper;
>>
>> - if (fb_helper)
>> + if (fb_helper->dev)
>> drm_fb_helper_hotplug_event(fb_helper);
> Food for thought:
>
> Quick grep shows that no other drivers have such a ->dev check. Does
> this mean that either the issue is rockchip specific?
> If not, one could look into resolving the problem directly in drm core.
>
> Or at least update the other users, so they don't stumble upon the problem?
The fbdev helpers support already handling hotplug events before you
have finalized the fbdev setup. Please read the kerneldoc for the
various fbdev functions, they explain what you should be doing. This
hack here should indeed not be needed.
-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
On 2017年07月31日 20:28, Daniel Vetter wrote:
> On Mon, Jul 31, 2017 at 1:57 PM, Emil Velikov <[email protected]> wrote:
>> On 31 July 2017 at 10:50, Mark Yao <[email protected]> wrote:
>>> Since fb_helper is not a pointer on rockchip_drm_private, it's no
>>> need to check pointer.
>>>
>>> Kms hotplug event may race into fbdev helper initial, and fb_helper->dev
>>> may be NULL pointer, that would cause the bug:
>>>
>>> [ 0.735411] [00000200] *pgd=00000000f6ffe003, *pud=00000000f6ffe003, *pmd=0000000000000000
>>> [ 0.736156] Internal error: Oops: 96000005 [#1] PREEMPT SMP
>>> [ 0.736648] Modules linked in:
>>> [ 0.736930] CPU: 2 PID: 20 Comm: kworker/2:0 Not tainted 4.4.41 #20
>>> [ 0.737480] Hardware name: Rockchip RK3399 Board rev2 (BOX) (DT)
>>> [ 0.738020] Workqueue: events cdn_dp_pd_event_work
>>> [ 0.738447] task: ffffffc0f21f3100 ti: ffffffc0f2218000 task.ti: ffffffc0f2218000
>>> [ 0.739109] PC is at mutex_lock+0x14/0x44
>>> [ 0.739469] LR is at drm_fb_helper_hotplug_event+0x30/0x114
>>> [ 0.756253] [<ffffff8008a344f4>] mutex_lock+0x14/0x44
>>> [ 0.756260] [<ffffff8008445708>] drm_fb_helper_hotplug_event+0x30/0x114
>>> [ 0.756271] [<ffffff8008473c84>] rockchip_drm_output_poll_changed+0x18/0x20
>>> [ 0.756280] [<ffffff8008439fcc>] drm_kms_helper_hotplug_event+0x28/0x34
>>> [ 0.756286] [<ffffff800846c444>] cdn_dp_pd_event_work+0x394/0x3c4
>>> [ 0.756295] [<ffffff80080b2b38>] process_one_work+0x218/0x3e0
>>> [ 0.756302] [<ffffff80080b3538>] worker_thread+0x2e8/0x404
>>> [ 0.756308] [<ffffff80080b7e70>] kthread+0xe8/0xf0
>>> [ 0.756316] [<ffffff8008082690>] ret_from_fork+0x10/0x40
>>>
>>> Signed-off-by: Mark Yao <[email protected]>
>>> ---
>>> drivers/gpu/drm/rockchip/rockchip_drm_fb.c | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
>>> index 81f9548..e6bd0f4 100644
>>> --- a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
>>> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
>>> @@ -170,7 +170,7 @@ static void rockchip_drm_output_poll_changed(struct drm_device *dev)
>>> struct rockchip_drm_private *private = dev->dev_private;
>>> struct drm_fb_helper *fb_helper = &private->fbdev_helper;
>>>
>>> - if (fb_helper)
>>> + if (fb_helper->dev)
>>> drm_fb_helper_hotplug_event(fb_helper);
>> Food for thought:
>>
>> Quick grep shows that no other drivers have such a ->dev check. Does
>> this mean that either the issue is rockchip specific?
>> If not, one could look into resolving the problem directly in drm core.
>>
>> Or at least update the other users, so they don't stumble upon the problem?
> The fbdev helpers support already handling hotplug events before you
> have finalized the fbdev setup. Please read the kerneldoc for the
> various fbdev functions, they explain what you should be doing. This
> hack here should indeed not be needed.
> -Daniel
Hi Daniel
Right, the doc[0] already detail this:
It is possible, though perhaps somewhat tricky, to implement race-free hotplug detection using
the fbdev helpers. The drm_fb_helper_prepare() helper must be called first to initialize the
minimum required to make hotplug detection work.Drivers also need to make sure to properly
set up the drm_mode_config.funcs member. After calling drm_kms_helper_poll_init() it is safe to
enable interrupts and start processing hotplug events.
The problem is drm/rockchip do the wrong initial, call drm_kms_helper_poll_init before fbdev setup.
will fix it at next version.
[0]: https://01.org/linuxgraphics/gfx-docs/drm/gpu/drm-kms-helpers.html#fbdev-helper-functions-reference
Best regards.
--
Ｍark Yao
According to the kerneldoc[0], should do fbdev setup before calling
drm_kms_helper_poll_init(), otherwise, Kms hotplug event may race
into fbdev helper initial, and fb_helper->dev may be NULL pointer,
that would cause the bug:
[ 0.735411] [00000200] *pgd=00000000f6ffe003, *pud=00000000f6ffe003, *pmd=0000000000000000
[ 0.736156] Internal error: Oops: 96000005 [#1] PREEMPT SMP
[ 0.736648] Modules linked in:
[ 0.736930] CPU: 2 PID: 20 Comm: kworker/2:0 Not tainted 4.4.41 #20
[ 0.737480] Hardware name: Rockchip RK3399 Board rev2 (BOX) (DT)
[ 0.738020] Workqueue: events cdn_dp_pd_event_work
[ 0.738447] task: ffffffc0f21f3100 ti: ffffffc0f2218000 task.ti: ffffffc0f2218000
[ 0.739109] PC is at mutex_lock+0x14/0x44
[ 0.739469] LR is at drm_fb_helper_hotplug_event+0x30/0x114
[ 0.756253] [<ffffff8008a344f4>] mutex_lock+0x14/0x44
[ 0.756260] [<ffffff8008445708>] drm_fb_helper_hotplug_event+0x30/0x114
[ 0.756271] [<ffffff8008473c84>] rockchip_drm_output_poll_changed+0x18/0x20
[ 0.756280] [<ffffff8008439fcc>] drm_kms_helper_hotplug_event+0x28/0x34
[ 0.756286] [<ffffff800846c444>] cdn_dp_pd_event_work+0x394/0x3c4
[ 0.756295] [<ffffff80080b2b38>] process_one_work+0x218/0x3e0
[ 0.756302] [<ffffff80080b3538>] worker_thread+0x2e8/0x404
[ 0.756308] [<ffffff80080b7e70>] kthread+0xe8/0xf0
[ 0.756316] [<ffffff8008082690>] ret_from_fork+0x10/0x40
[0]: https://01.org/linuxgraphics/gfx-docs/drm/gpu/drm-kms-helpers.html
Signed-off-by: Mark Yao <[email protected]>
---
Changes in v1.1:
- According to the kerneldoc, fix the race bug in generic way.
drivers/gpu/drm/rockchip/rockchip_drm_drv.c | 13 ++++++-------
1 file changed, 6 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c
index 848edcf..c41f48a 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c
@@ -161,22 +161,21 @@ static int rockchip_drm_bind(struct device *dev)
*/
drm_dev->irq_enabled = true;
- /* init kms poll for handling hpd */
- drm_kms_helper_poll_init(drm_dev);
-
ret = rockchip_drm_fbdev_init(drm_dev);
if (ret)
- goto err_kms_helper_poll_fini;
+ goto err_unbind_all;
+
+ /* init kms poll for handling hpd */
+ drm_kms_helper_poll_init(drm_dev);
ret = drm_dev_register(drm_dev, 0);
if (ret)
- goto err_fbdev_fini;
+ goto err_kms_helper_poll_fini;
return 0;
-err_fbdev_fini:
- rockchip_drm_fbdev_fini(drm_dev);
err_kms_helper_poll_fini:
drm_kms_helper_poll_fini(drm_dev);
+ rockchip_drm_fbdev_fini(drm_dev);
err_unbind_all:
component_unbind_all(dev, drm_dev);
err_mode_config_cleanup:
--
1.9.1
Hi mark,
在 2017/7/31 17:49, Mark Yao 写道:
> Since atomic framework, crtc enable and disable are in pairs,
> no need to wait vblank.
>
> Signed-off-by: Mark Yao <[email protected]>
> ---
> drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 36 -----------------------------
> 1 file changed, 36 deletions(-)
>
Reviewed-by: Sandy huang <[email protected]>
Hi mark,
在 2017/7/31 17:49, Mark Yao 写道:
> Iommu would get page fault with following path:
> vop_disable:
> 1, disable all windows and set vop config done
> 2, vop enter to standy, all windows not works, but their registers
> are not clean, when you read window's enable bit, may found the
> window is enable.
>
> vop_enable:
> 1, memcpy(vop->regsbak, vop->regs, len)
> save current vop registers to vop->regsbak, then you can found
> window is enable on regsbak.
> 2, VOP_WIN_SET(vop, win, gate, 1);
> force enable window gate, but gate and enable are on same
> hardware register, then window enable bit rewrite to vop hardware.
> 3, vop power on, and vop might try to scan destroyed buffer,
> then iommu get page fault.
>
> Move windows disable after vop regsbak restore, then vop regsbak mechanism
> would keep tracing the modify, everything would be safe.
>
> Signed-off-by: Mark Yao <[email protected]>
> ---
> drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 33 +++++++++++++----------------
> 1 file changed, 15 insertions(+), 18 deletions(-)
>
Reviewed-by: Sandy huang <[email protected]>
Hi mark,
在 2017/7/31 17:49, Mark Yao 写道:
> fixup the scale calculation formula on the case
> src_height == (dst_height/2).
>
> Signed-off-by: Mark Yao <[email protected]>
> ---
> drivers/gpu/drm/rockchip/rockchip_drm_vop.h | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
> index af1091f..56bbd2e 100644
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
> @@ -299,6 +299,9 @@ static inline uint16_t scl_get_bili_dn_vskip(int src_h, int dst_h,
>
> act_height = (src_h + vskiplines - 1) / vskiplines;
>
> + if (act_height == dst_h)
> + return GET_SCL_FT_BILI_DN(src_h, dst_h) / vskiplines;
> +
> return GET_SCL_FT_BILI_DN(act_height, dst_h);
> }
>
>
Reviewed-by: Sandy huang <[email protected]>
在 2017/7/31 17:49, Mark Yao 写道:
> VOP pitch register is word align, need align to word.
>
> VOP_WIN0_VIR:
> bit[31:16] win0_vir_stride_uv
> Number of words of Win0 uv Virtual width
> bit[15:0] win0_vir_width
> Number of words of Win0 yrgb Virtual width
> ARGB888 : win0_vir_width
> RGB888 : (win0_vir_width*3/4) + (win0_vir_width%3)
> RGB565 : ceil(win0_vir_width/2)
> YUV : ceil(win0_vir_width/4)
>
> Signed-off-by: Mark Yao <[email protected]>
> ---
> drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
Reviewed-by: Sandy huang <[email protected]>
Hi mark,
在 2017/7/31 17:49, Mark Yao 写道:
> The user would be confused while facing a error commit without
> any error report.
>
> Signed-off-by: Mark Yao <[email protected]>
> ---
> drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
> index fa0d9f7..999c2e0 100644
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
> @@ -674,8 +674,10 @@ static int vop_plane_atomic_check(struct drm_plane *plane,
> * Src.x1 can be odd when do clip, but yuv plane start point
> * need align with 2 pixel.
> */
> - if (is_yuv_support(fb->format->format) && ((state->src.x1 >> 16) % 2))
> + if (is_yuv_support(fb->format->format) && ((state->src.x1 >> 16) % 2)) {
> + DRM_ERROR("Invalid Source: Yuv format not support odd xpos\n");
> return -EINVAL;
> + }
>
> return 0;
> }
>
Reviewed-by: Sandy huang <[email protected]>
Hi Mark,
在 2017/8/1 16:11, Mark Yao 写道:
> According to the kerneldoc[0], should do fbdev setup before calling
> drm_kms_helper_poll_init(), otherwise, Kms hotplug event may race
> into fbdev helper initial, and fb_helper->dev may be NULL pointer,
> that would cause the bug:
> [ 0.735411] [00000200] *pgd=00000000f6ffe003, *pud=00000000f6ffe003, *pmd=0000000000000000
> [ 0.736156] Internal error: Oops: 96000005 [#1] PREEMPT SMP
> [ 0.736648] Modules linked in:
> [ 0.736930] CPU: 2 PID: 20 Comm: kworker/2:0 Not tainted 4.4.41 #20
> [ 0.737480] Hardware name: Rockchip RK3399 Board rev2 (BOX) (DT)
> [ 0.738020] Workqueue: events cdn_dp_pd_event_work
> [ 0.738447] task: ffffffc0f21f3100 ti: ffffffc0f2218000 task.ti: ffffffc0f2218000
> [ 0.739109] PC is at mutex_lock+0x14/0x44
> [ 0.739469] LR is at drm_fb_helper_hotplug_event+0x30/0x114
> [ 0.756253] [<ffffff8008a344f4>] mutex_lock+0x14/0x44
> [ 0.756260] [<ffffff8008445708>] drm_fb_helper_hotplug_event+0x30/0x114
> [ 0.756271] [<ffffff8008473c84>] rockchip_drm_output_poll_changed+0x18/0x20
> [ 0.756280] [<ffffff8008439fcc>] drm_kms_helper_hotplug_event+0x28/0x34
> [ 0.756286] [<ffffff800846c444>] cdn_dp_pd_event_work+0x394/0x3c4
> [ 0.756295] [<ffffff80080b2b38>] process_one_work+0x218/0x3e0
> [ 0.756302] [<ffffff80080b3538>] worker_thread+0x2e8/0x404
> [ 0.756308] [<ffffff80080b7e70>] kthread+0xe8/0xf0
> [ 0.756316] [<ffffff8008082690>] ret_from_fork+0x10/0x40
>
> [0]: https://01.org/linuxgraphics/gfx-docs/drm/gpu/drm-kms-helpers.html
>
> Signed-off-by: Mark Yao <[email protected]>
> ---
> Changes in v1.1:
> - According to the kerneldoc, fix the race bug in generic way.
>
> drivers/gpu/drm/rockchip/rockchip_drm_drv.c | 13 ++++++-------
> 1 file changed, 6 insertions(+), 7 deletions(-)
>
Reviewed-by: Sandy huang <[email protected]>
On Mon, Jul 31, 2017 at 5:49 AM, Mark Yao <[email protected]> wrote:
> Here are some fixes port from rockchip_linux project[0],
>
> Tested on rk3399 and rk3288 board.
>
> [0]: https://github.com/rockchip-linux/kernel
>
> Mark Yao (6):
> drm/rockchip: vop: no need wait vblank on crtc enable
> drm/rockchip: vop: fix iommu page fault when resume
> drm/rockchip: vop: fix NV12 video display error
> drm/rockchip: vop: round_up pitches to word align
> drm/rockchip: vop: report error when check resource error
> drm/rockchip: fix race with kms hotplug and fbdev
Hi Mark,
I noticed you applied this set to both -misc-next and -misc-fixes. In
the future, please do not apply patches in both places.
Instead, consult the flowchart under "Where Do I Apply My Patch?" at
https://01.org/linuxgraphics/gfx-docs/maintainer-tools/drm-misc.html
and choose the most appropriate place. If you are unsure, please reach
out to a -misc maintainer.
Thank you,
Sean
>
> drivers/gpu/drm/rockchip/rockchip_drm_fb.c | 2 +-
> drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 77 ++++++++---------------------
> drivers/gpu/drm/rockchip/rockchip_drm_vop.h | 3 ++
> 3 files changed, 24 insertions(+), 58 deletions(-)
>
> --
> 1.9.1
>
>
> _______________________________________________
> dri-devel mailing list
> [email protected]
> https://lists.freedesktop.org/mailman/listinfo/dri-devel