| Message ID | 20180204213624.18288-1-philippe.cornu@st.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Archit, Andrzej & Laurent, May I ask you please your feedback on this small patch? Many thanks, Philippe :-) On 02/04/2018 10:36 PM, Philippe Cornu wrote: > The readl_poll_timeout() return value is 0 in case of success > so it is better to detect errors without taking care of the > return value sign. > > Signed-off-by: Philippe Cornu <philippe.cornu@st.com> > --- > drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c | 10 +++++----- > 1 file changed, 5 insertions(+), 5 deletions(-) > > diff --git a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c > index 65aeb3f78b48..4d0e8471a15c 100644 > --- a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c > +++ b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c > @@ -342,7 +342,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) > ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, > val, !(val & GEN_CMD_FULL), 1000, > CMD_PKT_STATUS_TIMEOUT_US); > - if (ret < 0) { > + if (ret) { > dev_err(dsi->dev, "failed to get available command FIFO\n"); > return ret; > } > @@ -353,7 +353,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) > ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, > val, (val & mask) == mask, > 1000, CMD_PKT_STATUS_TIMEOUT_US); > - if (ret < 0) { > + if (ret) { > dev_err(dsi->dev, "failed to write command FIFO\n"); > return ret; > } > @@ -385,7 +385,7 @@ static int dw_mipi_dsi_write(struct dw_mipi_dsi *dsi, > ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, > val, !(val & GEN_PLD_W_FULL), 1000, > CMD_PKT_STATUS_TIMEOUT_US); > - if (ret < 0) { > + if (ret) { > dev_err(dsi->dev, > "failed to get available write payload FIFO\n"); > return ret; > @@ -716,13 +716,13 @@ static void dw_mipi_dsi_dphy_enable(struct dw_mipi_dsi *dsi) > > ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, val, > val & PHY_LOCK, 1000, PHY_STATUS_TIMEOUT_US); > - if (ret < 0) > + if (ret) > DRM_DEBUG_DRIVER("failed to wait phy lock state\n"); > > ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, > val, val & PHY_STOP_STATE_CLK_LANE, 1000, > PHY_STATUS_TIMEOUT_US); > - if (ret < 0) > + if (ret) > DRM_DEBUG_DRIVER("failed to wait phy clk lane stop state\n"); > } > >
On 01.03.2018 10:00, Philippe CORNU wrote: > Hi Archit, Andrzej & Laurent, > > May I ask you please your feedback on this small patch? > Many thanks, > > Philippe :-) > > On 02/04/2018 10:36 PM, Philippe Cornu wrote: >> The readl_poll_timeout() return value is 0 in case of success >> so it is better to detect errors without taking care of the >> return value sign. >> >> Signed-off-by: Philippe Cornu <philippe.cornu@st.com> The patch is of course correct. However I am not sure if necessary. For sure functionally it does not change anything. AFAIK kernel CodingStyle says nothing about it, so I suppose it is matter of personal taste. Anyway I can give it: Reviewed-by: Andrzej Hajda <a.hajda@samsung.com> -- Regards Andrzej >> --- >> drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c | 10 +++++----- >> 1 file changed, 5 insertions(+), 5 deletions(-) >> >> diff --git a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c >> index 65aeb3f78b48..4d0e8471a15c 100644 >> --- a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c >> +++ b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c >> @@ -342,7 +342,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) >> ret = qq(dsi->base + DSI_CMD_PKT_STATUS, >> val, !(val & GEN_CMD_FULL), 1000, >> CMD_PKT_STATUS_TIMEOUT_US); >> - if (ret < 0) { >> + if (ret) { >> dev_err(dsi->dev, "failed to get available command FIFO\n"); >> return ret; >> } >> @@ -353,7 +353,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) >> ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, >> val, (val & mask) == mask, >> 1000, CMD_PKT_STATUS_TIMEOUT_US); >> - if (ret < 0) { >> + if (ret) { >> dev_err(dsi->dev, "failed to write command FIFO\n"); >> return ret; >> } >> @@ -385,7 +385,7 @@ static int dw_mipi_dsi_write(struct dw_mipi_dsi *dsi, >> ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, >> val, !(val & GEN_PLD_W_FULL), 1000, >> CMD_PKT_STATUS_TIMEOUT_US); >> - if (ret < 0) { >> + if (ret) { >> dev_err(dsi->dev, >> "failed to get available write payload FIFO\n"); >> return ret; >> @@ -716,13 +716,13 @@ static void dw_mipi_dsi_dphy_enable(struct dw_mipi_dsi *dsi) >> >> ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, val, >> val & PHY_LOCK, 1000, PHY_STATUS_TIMEOUT_US); >> - if (ret < 0) >> + if (ret) >> DRM_DEBUG_DRIVER("failed to wait phy lock state\n"); >> >> ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, >> val, val & PHY_STOP_STATE_CLK_LANE, 1000, >> PHY_STATUS_TIMEOUT_US); >> - if (ret < 0) >> + if (ret) >> DRM_DEBUG_DRIVER("failed to wait phy clk lane stop state\n"); >> } >>
Hi Andrzej, On 03/02/2018 11:21 AM, Andrzej Hajda wrote: > On 01.03.2018 10:00, Philippe CORNU wrote: >> Hi Archit, Andrzej & Laurent, >> >> May I ask you please your feedback on this small patch? >> Many thanks, >> >> Philippe :-) >> >> On 02/04/2018 10:36 PM, Philippe Cornu wrote: >>> The readl_poll_timeout() return value is 0 in case of success >>> so it is better to detect errors without taking care of the >>> return value sign. >>> >>> Signed-off-by: Philippe Cornu <philippe.cornu@st.com> > > The patch is of course correct. However I am not sure if necessary. For > sure functionally it does not change anything. > AFAIK kernel CodingStyle says nothing about it, so I suppose it is > matter of personal taste. I sent this tiny patch in order to homogenize the dw mipi driver because there were both cases "if (ret)" & "if (ret < 0)" in the source code. I did not really find a preferred way in the kernel source code so I selected what sounds the best to me ie "if (ret)" but it is not a problem to make another patch for "if (ret < 0)" everywhere :-) In any case, the most important from my pov is to have a homogeneous source code :-) Does anyone have a preferred choice between "if (ret)" & "if (ret < 0)" after a "ret = readl_poll_timeout()"? > Anyway I can give it: > Reviewed-by: Andrzej Hajda <a.hajda@samsung.com> Many thanks, Philippe :) > > -- > Regards > Andrzej > >>> --- >>> drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c | 10 +++++----- >>> 1 file changed, 5 insertions(+), 5 deletions(-) >>> >>> diff --git a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c >>> index 65aeb3f78b48..4d0e8471a15c 100644 >>> --- a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c >>> +++ b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c >>> @@ -342,7 +342,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) >>> ret = qq(dsi->base + DSI_CMD_PKT_STATUS, >>> val, !(val & GEN_CMD_FULL), 1000, >>> CMD_PKT_STATUS_TIMEOUT_US); >>> - if (ret < 0) { >>> + if (ret) { >>> dev_err(dsi->dev, "failed to get available command FIFO\n"); >>> return ret; >>> } >>> @@ -353,7 +353,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) >>> ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, >>> val, (val & mask) == mask, >>> 1000, CMD_PKT_STATUS_TIMEOUT_US); >>> - if (ret < 0) { >>> + if (ret) { >>> dev_err(dsi->dev, "failed to write command FIFO\n"); >>> return ret; >>> } >>> @@ -385,7 +385,7 @@ static int dw_mipi_dsi_write(struct dw_mipi_dsi *dsi, >>> ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, >>> val, !(val & GEN_PLD_W_FULL), 1000, >>> CMD_PKT_STATUS_TIMEOUT_US); >>> - if (ret < 0) { >>> + if (ret) { >>> dev_err(dsi->dev, >>> "failed to get available write payload FIFO\n"); >>> return ret; >>> @@ -716,13 +716,13 @@ static void dw_mipi_dsi_dphy_enable(struct dw_mipi_dsi *dsi) >>> >>> ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, val, >>> val & PHY_LOCK, 1000, PHY_STATUS_TIMEOUT_US); >>> - if (ret < 0) >>> + if (ret) >>> DRM_DEBUG_DRIVER("failed to wait phy lock state\n"); >>> >>> ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, >>> val, val & PHY_STOP_STATE_CLK_LANE, 1000, >>> PHY_STATUS_TIMEOUT_US); >>> - if (ret < 0) >>> + if (ret) >>> DRM_DEBUG_DRIVER("failed to wait phy clk lane stop state\n"); >>> } >>> > >
On 02.03.2018 14:15, Philippe CORNU wrote: > Hi Andrzej, > > > On 03/02/2018 11:21 AM, Andrzej Hajda wrote: >> On 01.03.2018 10:00, Philippe CORNU wrote: >>> Hi Archit, Andrzej & Laurent, >>> >>> May I ask you please your feedback on this small patch? >>> Many thanks, >>> >>> Philippe :-) >>> >>> On 02/04/2018 10:36 PM, Philippe Cornu wrote: >>>> The readl_poll_timeout() return value is 0 in case of success >>>> so it is better to detect errors without taking care of the >>>> return value sign. >>>> >>>> Signed-off-by: Philippe Cornu <philippe.cornu@st.com> >> The patch is of course correct. However I am not sure if necessary. For >> sure functionally it does not change anything. >> AFAIK kernel CodingStyle says nothing about it, so I suppose it is >> matter of personal taste. > I sent this tiny patch in order to homogenize the dw mipi driver because > there were both cases "if (ret)" & "if (ret < 0)" in the source code. > > I did not really find a preferred way in the kernel source code so I > selected what sounds the best to me ie "if (ret)" but it is not a > problem to make another patch for "if (ret < 0)" everywhere :-) > > In any case, the most important from my pov is to have a homogeneous > source code :-) OK Queued to drm-misc-next. Regards Andrzej
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c index 65aeb3f78b48..4d0e8471a15c 100644 --- a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c +++ b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c @@ -342,7 +342,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, val, !(val & GEN_CMD_FULL), 1000, CMD_PKT_STATUS_TIMEOUT_US); - if (ret < 0) { + if (ret) { dev_err(dsi->dev, "failed to get available command FIFO\n"); return ret; } @@ -353,7 +353,7 @@ static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val) ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, val, (val & mask) == mask, 1000, CMD_PKT_STATUS_TIMEOUT_US); - if (ret < 0) { + if (ret) { dev_err(dsi->dev, "failed to write command FIFO\n"); return ret; } @@ -385,7 +385,7 @@ static int dw_mipi_dsi_write(struct dw_mipi_dsi *dsi, ret = readl_poll_timeout(dsi->base + DSI_CMD_PKT_STATUS, val, !(val & GEN_PLD_W_FULL), 1000, CMD_PKT_STATUS_TIMEOUT_US); - if (ret < 0) { + if (ret) { dev_err(dsi->dev, "failed to get available write payload FIFO\n"); return ret; @@ -716,13 +716,13 @@ static void dw_mipi_dsi_dphy_enable(struct dw_mipi_dsi *dsi) ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, val, val & PHY_LOCK, 1000, PHY_STATUS_TIMEOUT_US); - if (ret < 0) + if (ret) DRM_DEBUG_DRIVER("failed to wait phy lock state\n"); ret = readl_poll_timeout(dsi->base + DSI_PHY_STATUS, val, val & PHY_STOP_STATE_CLK_LANE, 1000, PHY_STATUS_TIMEOUT_US); - if (ret < 0) + if (ret) DRM_DEBUG_DRIVER("failed to wait phy clk lane stop state\n"); }
The readl_poll_timeout() return value is 0 in case of success so it is better to detect errors without taking care of the return value sign. Signed-off-by: Philippe Cornu <philippe.cornu@st.com> --- drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)