On msm8998, vblank timeouts are observed because the DSI controller is not reset properly, which ends up stalling the MDP. This is because the reset logic is not correct per the hardware documentation.
The documentation states that after asserting reset, software should wait some time (no indication of how long), or poll the status register until it returns 0 before deasserting reset.
wmb() is insufficient for this purpose since it just ensures ordering, not timing between writes. Since asserting and deasserting reset occurs on the same register, ordering is already guaranteed by the architecture, making the wmb extraneous.
Since we would define a timeout for polling the status register to avoid a possible infinite loop, lets just use a static delay of 20 ms, since 16.666 ms is the time available to process one frame at 60 fps.
Fixes: a689554ba6ed (drm/msm: Initial add DSI connector support) Signed-off-by: Jeffrey Hugo jeffrey.l.hugo@gmail.com ---
Rob et al, is it possible for this to go into a 5.4-rc?
drivers/gpu/drm/msm/dsi/dsi_host.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c index 663ff9f4fac9..68ded9b4735d 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_host.c +++ b/drivers/gpu/drm/msm/dsi/dsi_host.c @@ -986,7 +986,7 @@ static void dsi_sw_reset(struct msm_dsi_host *msm_host) wmb(); /* clocks need to be enabled before reset */
dsi_write(msm_host, REG_DSI_RESET, 1); - wmb(); /* make sure reset happen */ + msleep(20); /* make sure reset happen */ dsi_write(msm_host, REG_DSI_RESET, 0); }
@@ -1396,7 +1396,7 @@ static void dsi_sw_reset_restore(struct msm_dsi_host *msm_host)
/* dsi controller can only be reset while clocks are running */ dsi_write(msm_host, REG_DSI_RESET, 1); - wmb(); /* make sure reset happen */ + msleep(20); /* make sure reset happen */ dsi_write(msm_host, REG_DSI_RESET, 0); wmb(); /* controller out of reset */ dsi_write(msm_host, REG_DSI_CTRL, data0);
On Wed, Oct 09, 2019 at 02:34:54PM -0700, Jeffrey Hugo wrote:
On msm8998, vblank timeouts are observed because the DSI controller is not reset properly, which ends up stalling the MDP. This is because the reset logic is not correct per the hardware documentation.
The documentation states that after asserting reset, software should wait some time (no indication of how long), or poll the status register until it returns 0 before deasserting reset.
wmb() is insufficient for this purpose since it just ensures ordering, not timing between writes. Since asserting and deasserting reset occurs on the same register, ordering is already guaranteed by the architecture, making the wmb extraneous.
Since we would define a timeout for polling the status register to avoid a possible infinite loop, lets just use a static delay of 20 ms, since 16.666 ms is the time available to process one frame at 60 fps.
Fixes: a689554ba6ed (drm/msm: Initial add DSI connector support) Signed-off-by: Jeffrey Hugo jeffrey.l.hugo@gmail.com
Rob et al, is it possible for this to go into a 5.4-rc?
drivers/gpu/drm/msm/dsi/dsi_host.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c index 663ff9f4fac9..68ded9b4735d 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_host.c +++ b/drivers/gpu/drm/msm/dsi/dsi_host.c @@ -986,7 +986,7 @@ static void dsi_sw_reset(struct msm_dsi_host *msm_host) wmb(); /* clocks need to be enabled before reset */
dsi_write(msm_host, REG_DSI_RESET, 1);
- wmb(); /* make sure reset happen */
- msleep(20); /* make sure reset happen */
Could you please pull this out into a #define used for both in case we decide to tweak it? I don't want these 2 values to drift.
Thanks, Sean
dsi_write(msm_host, REG_DSI_RESET, 0); }
@@ -1396,7 +1396,7 @@ static void dsi_sw_reset_restore(struct msm_dsi_host *msm_host)
/* dsi controller can only be reset while clocks are running */ dsi_write(msm_host, REG_DSI_RESET, 1);
- wmb(); /* make sure reset happen */
- msleep(20); /* make sure reset happen */ dsi_write(msm_host, REG_DSI_RESET, 0); wmb(); /* controller out of reset */ dsi_write(msm_host, REG_DSI_CTRL, data0);
-- 2.17.1
On Thu, Oct 10, 2019 at 2:45 PM Sean Paul sean@poorly.run wrote:
On Wed, Oct 09, 2019 at 02:34:54PM -0700, Jeffrey Hugo wrote:
On msm8998, vblank timeouts are observed because the DSI controller is not reset properly, which ends up stalling the MDP. This is because the reset logic is not correct per the hardware documentation.
The documentation states that after asserting reset, software should wait some time (no indication of how long), or poll the status register until it returns 0 before deasserting reset.
wmb() is insufficient for this purpose since it just ensures ordering, not timing between writes. Since asserting and deasserting reset occurs on the same register, ordering is already guaranteed by the architecture, making the wmb extraneous.
Since we would define a timeout for polling the status register to avoid a possible infinite loop, lets just use a static delay of 20 ms, since 16.666 ms is the time available to process one frame at 60 fps.
Fixes: a689554ba6ed (drm/msm: Initial add DSI connector support) Signed-off-by: Jeffrey Hugo jeffrey.l.hugo@gmail.com
Rob et al, is it possible for this to go into a 5.4-rc?
Sorry, I missed this on the first go-around, I'm Ok with this getting into 5.4. Rob, if you're Ok with this, I can send it through -misc unless you're planning an msm-fixes PR.
drivers/gpu/drm/msm/dsi/dsi_host.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c index 663ff9f4fac9..68ded9b4735d 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_host.c +++ b/drivers/gpu/drm/msm/dsi/dsi_host.c @@ -986,7 +986,7 @@ static void dsi_sw_reset(struct msm_dsi_host *msm_host) wmb(); /* clocks need to be enabled before reset */
dsi_write(msm_host, REG_DSI_RESET, 1);
wmb(); /* make sure reset happen */
msleep(20); /* make sure reset happen */Could you please pull this out into a #define used for both in case we decide to tweak it? I don't want these 2 values to drift.
oh yeah, and with that fixed,
Reviewed-by: Sean Paul sean@poorly.run
Thanks, Sean
dsi_write(msm_host, REG_DSI_RESET, 0);}
@@ -1396,7 +1396,7 @@ static void dsi_sw_reset_restore(struct msm_dsi_host *msm_host)
/* dsi controller can only be reset while clocks are running */ dsi_write(msm_host, REG_DSI_RESET, 1);
wmb(); /* make sure reset happen */
msleep(20); /* make sure reset happen */ dsi_write(msm_host, REG_DSI_RESET, 0); wmb(); /* controller out of reset */ dsi_write(msm_host, REG_DSI_CTRL, data0);-- 2.17.1
-- Sean Paul, Software Engineer, Google / Chromium OS
On Thu, Oct 10, 2019 at 12:49 PM Sean Paul sean@poorly.run wrote:
On Thu, Oct 10, 2019 at 2:45 PM Sean Paul sean@poorly.run wrote:
On Wed, Oct 09, 2019 at 02:34:54PM -0700, Jeffrey Hugo wrote:
On msm8998, vblank timeouts are observed because the DSI controller is not reset properly, which ends up stalling the MDP. This is because the reset logic is not correct per the hardware documentation.
The documentation states that after asserting reset, software should wait some time (no indication of how long), or poll the status register until it returns 0 before deasserting reset.
wmb() is insufficient for this purpose since it just ensures ordering, not timing between writes. Since asserting and deasserting reset occurs on the same register, ordering is already guaranteed by the architecture, making the wmb extraneous.
Since we would define a timeout for polling the status register to avoid a possible infinite loop, lets just use a static delay of 20 ms, since 16.666 ms is the time available to process one frame at 60 fps.
Fixes: a689554ba6ed (drm/msm: Initial add DSI connector support) Signed-off-by: Jeffrey Hugo jeffrey.l.hugo@gmail.com
Rob et al, is it possible for this to go into a 5.4-rc?
Sorry, I missed this on the first go-around, I'm Ok with this getting into 5.4. Rob, if you're Ok with this, I can send it through -misc unless you're planning an msm-fixes PR.
drivers/gpu/drm/msm/dsi/dsi_host.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c index 663ff9f4fac9..68ded9b4735d 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_host.c +++ b/drivers/gpu/drm/msm/dsi/dsi_host.c @@ -986,7 +986,7 @@ static void dsi_sw_reset(struct msm_dsi_host *msm_host) wmb(); /* clocks need to be enabled before reset */
dsi_write(msm_host, REG_DSI_RESET, 1);
wmb(); /* make sure reset happen */
msleep(20); /* make sure reset happen */Could you please pull this out into a #define used for both in case we decide to tweak it? I don't want these 2 values to drift.
Oh, yeah. That's a really good point. Will fix.
oh yeah, and with that fixed,
Reviewed-by: Sean Paul sean@poorly.run
Thanks.
Thanks, Sean
dsi_write(msm_host, REG_DSI_RESET, 0);}
@@ -1396,7 +1396,7 @@ static void dsi_sw_reset_restore(struct msm_dsi_host *msm_host)
/* dsi controller can only be reset while clocks are running */ dsi_write(msm_host, REG_DSI_RESET, 1);
wmb(); /* make sure reset happen */
msleep(20); /* make sure reset happen */ dsi_write(msm_host, REG_DSI_RESET, 0); wmb(); /* controller out of reset */ dsi_write(msm_host, REG_DSI_CTRL, data0);-- 2.17.1
-- Sean Paul, Software Engineer, Google / Chromium OS
dri-devel@lists.freedesktop.org
-
Jeffrey Hugo -
Sean Paul