Message ID | 20201030232324.11170-1-khsieh@codeaurora.org (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | drm/msm/dp: promote irq_hpd handle to handle link trainign correctly | expand |
Subject has a typo in "training". Quoting Kuogee Hsieh (2020-10-30 16:23:24) > Some dongles, such as Apple, required link training done at irq_hpd s/required/require/ > request instead of plugin request. This patch promote irq_hpd hanlder s/hanlder/handler/ > to handle link training and setup hpd_state correctly. > > Signed-off-by: Kuogee Hsieh <khsieh@codeaurora.org> > --- Any Fixes tag? > drivers/gpu/drm/msm/dp/dp_display.c | 20 ++++++++++++++++++-- > 1 file changed, 18 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c > index 13b66266cd69..55627530957c 100644 > --- a/drivers/gpu/drm/msm/dp/dp_display.c > +++ b/drivers/gpu/drm/msm/dp/dp_display.c > @@ -483,10 +485,24 @@ static int dp_display_usbpd_attention_cb(struct device *dev) > return -ENODEV; > } > > + hpd = dp->usbpd; > + > /* check for any test request issued by sink */ > rc = dp_link_process_request(dp->link); > - if (!rc) > - dp_display_handle_irq_hpd(dp); > + if (!rc) { > + sink_request = dp->link->sink_request; > + if (sink_request & DS_PORT_STATUS_CHANGED) { > + dp->hpd_state = ST_CONNECT_PENDING; > + hpd->hpd_high = 1; > + } > + > + rc = dp_display_handle_irq_hpd(dp); > + > + if (rc && sink_request & DS_PORT_STATUS_CHANGED) { Can you add parenthesis around this? if (rc && (sink_request & DS_PORT_STATUS_CHANGED)) { I honestly don't know what's going on in this patch. It talks about making link training happen during irq hpd handler but this is the attention handler and we're checking port status changed? This is related? The code is really not clear. > + hpd->hpd_high = 0; > + dp->hpd_state = ST_DISCONNECTED; > + } > + } > > return rc; > } > > base-commit: 0e162b10644605428cd2596c12f8ed410cf9d2d9 What commit is this?
On Mon, Nov 2, 2020 at 11:29 AM Stephen Boyd <swboyd@chromium.org> wrote: > > Subject has a typo in "training". > > Quoting Kuogee Hsieh (2020-10-30 16:23:24) > > Some dongles, such as Apple, required link training done at irq_hpd > > s/required/require/ > > > request instead of plugin request. This patch promote irq_hpd hanlder > > s/hanlder/handler/ > > > to handle link training and setup hpd_state correctly. > > > > Signed-off-by: Kuogee Hsieh <khsieh@codeaurora.org> > > --- > > Any Fixes tag? > > > drivers/gpu/drm/msm/dp/dp_display.c | 20 ++++++++++++++++++-- > > 1 file changed, 18 insertions(+), 2 deletions(-) > > > > diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c > > index 13b66266cd69..55627530957c 100644 > > --- a/drivers/gpu/drm/msm/dp/dp_display.c > > +++ b/drivers/gpu/drm/msm/dp/dp_display.c > > @@ -483,10 +485,24 @@ static int dp_display_usbpd_attention_cb(struct device *dev) > > return -ENODEV; > > } > > > > + hpd = dp->usbpd; > > + > > /* check for any test request issued by sink */ > > rc = dp_link_process_request(dp->link); > > - if (!rc) > > - dp_display_handle_irq_hpd(dp); > > + if (!rc) { > > + sink_request = dp->link->sink_request; > > + if (sink_request & DS_PORT_STATUS_CHANGED) { > > + dp->hpd_state = ST_CONNECT_PENDING; > > + hpd->hpd_high = 1; > > + } > > + > > + rc = dp_display_handle_irq_hpd(dp); > > + > > + if (rc && sink_request & DS_PORT_STATUS_CHANGED) { > > Can you add parenthesis around this? > > if (rc && (sink_request & DS_PORT_STATUS_CHANGED)) { > > > I honestly don't know what's going on in this patch. It talks about > making link training happen during irq hpd handler but this is the > attention handler and we're checking port status changed? This is > related? The code is really not clear. > > > + hpd->hpd_high = 0; > > + dp->hpd_state = ST_DISCONNECTED; > > + } > > + } > > > > return rc; > > } > > > > base-commit: 0e162b10644605428cd2596c12f8ed410cf9d2d9 > > What commit is this? Note that I skipped over a few dp related patches yesterday while starting to pull things into msm-next-staging. Kuogee, when you send the next version can you make sure it is based against v5.10-rc or msm-next-staging? BR, -R
On 2020-11-02 11:29, Stephen Boyd wrote: > Subject has a typo in "training". > > Quoting Kuogee Hsieh (2020-10-30 16:23:24) >> Some dongles, such as Apple, required link training done at irq_hpd > > s/required/require/ > >> request instead of plugin request. This patch promote irq_hpd hanlder > > s/hanlder/handler/ > >> to handle link training and setup hpd_state correctly. >> >> Signed-off-by: Kuogee Hsieh <khsieh@codeaurora.org> >> --- > > Any Fixes tag? > >> drivers/gpu/drm/msm/dp/dp_display.c | 20 ++++++++++++++++++-- >> 1 file changed, 18 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c >> b/drivers/gpu/drm/msm/dp/dp_display.c >> index 13b66266cd69..55627530957c 100644 >> --- a/drivers/gpu/drm/msm/dp/dp_display.c >> +++ b/drivers/gpu/drm/msm/dp/dp_display.c >> @@ -483,10 +485,24 @@ static int dp_display_usbpd_attention_cb(struct >> device *dev) >> return -ENODEV; >> } >> >> + hpd = dp->usbpd; >> + >> /* check for any test request issued by sink */ >> rc = dp_link_process_request(dp->link); >> - if (!rc) >> - dp_display_handle_irq_hpd(dp); >> + if (!rc) { >> + sink_request = dp->link->sink_request; >> + if (sink_request & DS_PORT_STATUS_CHANGED) { >> + dp->hpd_state = ST_CONNECT_PENDING; >> + hpd->hpd_high = 1; >> + } >> + >> + rc = dp_display_handle_irq_hpd(dp); >> + >> + if (rc && sink_request & DS_PORT_STATUS_CHANGED) { > > Can you add parenthesis around this? > > if (rc && (sink_request & DS_PORT_STATUS_CHANGED)) { > > > I honestly don't know what's going on in this patch. It talks about > making link training happen during irq hpd handler but this is the > attention handler and we're checking port status changed? This is > related? The code is really not clear. irq_hpd request is generated by sinker to ask host attention that something has changed. POST_STATUS_CHNAGED bit set by sinker to indicated link had loss of sync. Therefore host need to restart link retaining to fix the link loss of sync problem. > >> + hpd->hpd_high = 0; >> + dp->hpd_state = ST_DISCONNECTED; >> + } >> + } >> >> return rc; >> } >> >> base-commit: 0e162b10644605428cd2596c12f8ed410cf9d2d9 > > What commit is this?
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c index 13b66266cd69..55627530957c 100644 --- a/drivers/gpu/drm/msm/dp/dp_display.c +++ b/drivers/gpu/drm/msm/dp/dp_display.c @@ -469,7 +469,9 @@ static int dp_display_handle_irq_hpd(struct dp_display_private *dp) static int dp_display_usbpd_attention_cb(struct device *dev) { int rc = 0; + u32 sink_request; struct dp_display_private *dp; + struct dp_usbpd *hpd; if (!dev) { DRM_ERROR("invalid dev\n"); @@ -483,10 +485,24 @@ static int dp_display_usbpd_attention_cb(struct device *dev) return -ENODEV; } + hpd = dp->usbpd; + /* check for any test request issued by sink */ rc = dp_link_process_request(dp->link); - if (!rc) - dp_display_handle_irq_hpd(dp); + if (!rc) { + sink_request = dp->link->sink_request; + if (sink_request & DS_PORT_STATUS_CHANGED) { + dp->hpd_state = ST_CONNECT_PENDING; + hpd->hpd_high = 1; + } + + rc = dp_display_handle_irq_hpd(dp); + + if (rc && sink_request & DS_PORT_STATUS_CHANGED) { + hpd->hpd_high = 0; + dp->hpd_state = ST_DISCONNECTED; + } + } return rc; }
Some dongles, such as Apple, required link training done at irq_hpd request instead of plugin request. This patch promote irq_hpd hanlder to handle link training and setup hpd_state correctly. Signed-off-by: Kuogee Hsieh <khsieh@codeaurora.org> --- drivers/gpu/drm/msm/dp/dp_display.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) base-commit: 0e162b10644605428cd2596c12f8ed410cf9d2d9