RE: [PATCH 1/1] OMAP: DSS2: RFBI driver update
On Mon, 2009-10-12 at 17:25 +0200, ext Christensen, Mikkel wrote: On Mon, Oct 12, 2009 at 09:17:27, Tomi Valkeinen wrote: Subject: RE: [PATCH 1/1] OMAP: DSS2: RFBI driver update On Mon, 2009-10-12 at 16:03 +0200, ext Christensen, Mikkel wrote: On Mon, Oct 12, 2009 at 03:25:27, Tomi Valkeinen wrote: Subject: Re: [PATCH 1/1] OMAP: DSS2: RFBI driver update On Fri, 2009-10-09 at 23:08 +0200, ext Mikkel Christensen wrote: This patch adds suspend / resume functionality to the RFBI driver along with missing callback functions needed by OMAP Frame buffer. Signed-off-by: Mikkel Christensen m...@ti.com --- drivers/video/omap2/dss/rfbi.c | 76 1 files changed, 76 insertions(+), 0 deletions(-) snip +static int rfbi_display_suspend(struct omap_dss_device *dssdev) { + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return -EINVAL; + + DSSDBG(rfbi_display_suspend\n); + + if (dssdev-driver-suspend) + dssdev-driver-suspend(dssdev); + + dispc_enable_lcd_out(0); I don't think this is needed. DSS hardware disables lcd_out when the transfer has finished. Although for correct operation suspend() should wait until the last transfer has been done before disabling clocks, which is something it currently doesn't. This was done with reference to the DPI and SDI files that do the same thing. It can be removed if not necessary. DPI and SDI work quite differently than RFBI, you can't just copy their functionality =). It is more correct without disable/enable_lcd_out, but it's not correct as there may be a transfer ongoing when suspend call comes. That's why there should be some mechanism to wait until the transfer has been done. DSI tries to do this (and hopefully correctly does =). OK. Thanks I will try to incorporate this. Where in the DSI driver is this mechanism implemented? It's the dsi_bus_lock mechanism. But it doesn't work as such for RFBI, because the DSI uses a bit different mechanism for updates. But somehow the suspend function needs to wait until the transfer has been completed, and suspend only after that (and also make sure that no new transfer starts before the suspend). Tomi -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/1] OMAP: DSS2: RFBI driver update
Mikkel Christensen m...@ti.com writes: This patch adds suspend / resume functionality to the RFBI driver along with missing callback functions needed by OMAP Frame buffer. Signed-off-by: Mikkel Christensen m...@ti.com --- drivers/video/omap2/dss/rfbi.c | 76 1 files changed, 76 insertions(+), 0 deletions(-) diff --git a/drivers/video/omap2/dss/rfbi.c b/drivers/video/omap2/dss/rfbi.c index 9dd2349..ddfc472 100644 --- a/drivers/video/omap2/dss/rfbi.c +++ b/drivers/video/omap2/dss/rfbi.c @@ -1181,6 +1181,7 @@ int rfbi_init(void) /* Enable autoidle and smart-idle */ l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~((0x03 3)|(0x01 0)); Please use symbolic names for these SYSCONFIG values. I realize you inherited this from other drivers, but it should be fixed. While you're at it, fixing the others would be nice too. You could do a cleanup patch before your actual driver update. Ideally, DSS needs to create an omap_hwmod, but in the meantime you could use the defines from mach/omap_hwmod.h l |= (1 0) | (2 3); rfbi_write_reg(RFBI_SYSCONFIG, l); @@ -1208,6 +1209,9 @@ static int rfbi_display_update(struct omap_dss_device *dssdev, { int rfbi_module; + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return 0; + if (w == 0 || h == 0) return 0; @@ -1239,6 +1243,18 @@ static int rfbi_display_enable_te(struct omap_dss_device *dssdev, bool enable) return 0; } +static enum omap_dss_update_mode rfbi_display_get_update_mode( + struct omap_dss_device *dssdev) +{ + return OMAP_DSS_UPDATE_MANUAL; +} + +static int rfbi_display_set_update_mode(struct omap_dss_device *dssdev, + enum omap_dss_update_mode mode) +{ + return 0; +} + static int rfbi_display_enable(struct omap_dss_device *dssdev) { int r; @@ -1269,6 +1285,7 @@ static int rfbi_display_enable(struct omap_dss_device *dssdev) rfbi_set_timings(dssdev-phy.rfbi.channel, dssdev-ctrl.rfbi_timings); + dssdev-state = OMAP_DSS_DISPLAY_ACTIVE; if (dssdev-driver-enable) { r = dssdev-driver-enable(dssdev); @@ -1288,12 +1305,67 @@ err0: static void rfbi_display_disable(struct omap_dss_device *dssdev) { + dssdev-state = OMAP_DSS_DISPLAY_DISABLED; + dssdev-driver-disable(dssdev); omap_dispc_unregister_isr(framedone_callback, NULL, DISPC_IRQ_FRAMEDONE); omap_dss_stop_device(dssdev); } +static int rfbi_display_suspend(struct omap_dss_device *dssdev) +{ + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return -EINVAL; + + DSSDBG(rfbi_display_suspend\n); + + if (dssdev-driver-suspend) + dssdev-driver-suspend(dssdev); + + dispc_enable_lcd_out(0); + + /* Force idle */ + rfbi_enable_clocks(1); + l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~(0x03 3); Again, use SYSC_SIDLEMODE_SHIFT, SYSC_SIDLEMODE_MASK. More of these follow. Kevin -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/1] OMAP: DSS2: RFBI driver update
On Fri, 2009-10-09 at 23:08 +0200, ext Mikkel Christensen wrote: This patch adds suspend / resume functionality to the RFBI driver along with missing callback functions needed by OMAP Frame buffer. Signed-off-by: Mikkel Christensen m...@ti.com --- drivers/video/omap2/dss/rfbi.c | 76 1 files changed, 76 insertions(+), 0 deletions(-) diff --git a/drivers/video/omap2/dss/rfbi.c b/drivers/video/omap2/dss/rfbi.c index 9dd2349..ddfc472 100644 --- a/drivers/video/omap2/dss/rfbi.c +++ b/drivers/video/omap2/dss/rfbi.c @@ -1181,6 +1181,7 @@ int rfbi_init(void) /* Enable autoidle and smart-idle */ l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~((0x03 3)|(0x01 0)); l |= (1 0) | (2 3); rfbi_write_reg(RFBI_SYSCONFIG, l); @@ -1208,6 +1209,9 @@ static int rfbi_display_update(struct omap_dss_device *dssdev, { int rfbi_module; + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return 0; + if (w == 0 || h == 0) return 0; @@ -1239,6 +1243,18 @@ static int rfbi_display_enable_te(struct omap_dss_device *dssdev, bool enable) return 0; } +static enum omap_dss_update_mode rfbi_display_get_update_mode( + struct omap_dss_device *dssdev) +{ + return OMAP_DSS_UPDATE_MANUAL; +} + +static int rfbi_display_set_update_mode(struct omap_dss_device *dssdev, + enum omap_dss_update_mode mode) +{ + return 0; +} + static int rfbi_display_enable(struct omap_dss_device *dssdev) { int r; @@ -1269,6 +1285,7 @@ static int rfbi_display_enable(struct omap_dss_device *dssdev) rfbi_set_timings(dssdev-phy.rfbi.channel, dssdev-ctrl.rfbi_timings); + dssdev-state = OMAP_DSS_DISPLAY_ACTIVE; if (dssdev-driver-enable) { r = dssdev-driver-enable(dssdev); @@ -1288,12 +1305,67 @@ err0: static void rfbi_display_disable(struct omap_dss_device *dssdev) { + dssdev-state = OMAP_DSS_DISPLAY_DISABLED; + dssdev-driver-disable(dssdev); omap_dispc_unregister_isr(framedone_callback, NULL, DISPC_IRQ_FRAMEDONE); omap_dss_stop_device(dssdev); } +static int rfbi_display_suspend(struct omap_dss_device *dssdev) +{ + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return -EINVAL; + + DSSDBG(rfbi_display_suspend\n); + + if (dssdev-driver-suspend) + dssdev-driver-suspend(dssdev); + + dispc_enable_lcd_out(0); I don't think this is needed. DSS hardware disables lcd_out when the transfer has finished. Although for correct operation suspend() should wait until the last transfer has been done before disabling clocks, which is something it currently doesn't. + + /* Force idle */ + rfbi_enable_clocks(1); + l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~(0x03 3); + rfbi_write_reg(RFBI_SYSCONFIG, l); + rfbi_enable_clocks(0); Why force idle? + + dssdev-state = OMAP_DSS_DISPLAY_SUSPENDED; + + return 0; +} + +static int rfbi_display_resume(struct omap_dss_device *dssdev) +{ + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_SUSPENDED) + return -EINVAL; + + DSSDBG(rfbi_display_resume\n); + + /* Re-enable autoidle */ + rfbi_enable_clocks(1); + l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~(0x03 3); + l |= (0x02 3); + rfbi_write_reg(RFBI_SYSCONFIG, l); + rfbi_enable_clocks(0); + + dispc_enable_lcd_out(1); This is not needed. lcd_out is enabled when the transfer is started. + + if (dssdev-driver-resume) + dssdev-driver-resume(dssdev); + + dssdev-state = OMAP_DSS_DISPLAY_ACTIVE; + + return 0; +} + int rfbi_init_display(struct omap_dss_device *dssdev) { dssdev-enable = rfbi_display_enable; @@ -1301,6 +1373,10 @@ int rfbi_init_display(struct omap_dss_device *dssdev) dssdev-update = rfbi_display_update; dssdev-sync = rfbi_display_sync; dssdev-enable_te = rfbi_display_enable_te; + dssdev-get_update_mode = rfbi_display_get_update_mode; + dssdev-set_update_mode = rfbi_display_set_update_mode; + dssdev-suspend = rfbi_display_suspend; + dssdev-resume = rfbi_display_resume; rfbi.dssdev[dssdev-phy.rfbi.channel] = dssdev; -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
RE: [PATCH 1/1] OMAP: DSS2: RFBI driver update
On Mon, Oct 12, 2009 at 03:25:27, Tomi Valkeinen wrote: Subject: Re: [PATCH 1/1] OMAP: DSS2: RFBI driver update On Fri, 2009-10-09 at 23:08 +0200, ext Mikkel Christensen wrote: This patch adds suspend / resume functionality to the RFBI driver along with missing callback functions needed by OMAP Frame buffer. Signed-off-by: Mikkel Christensen m...@ti.com --- drivers/video/omap2/dss/rfbi.c | 76 1 files changed, 76 insertions(+), 0 deletions(-) diff --git a/drivers/video/omap2/dss/rfbi.c b/drivers/video/omap2/dss/rfbi.c index 9dd2349..ddfc472 100644 --- a/drivers/video/omap2/dss/rfbi.c +++ b/drivers/video/omap2/dss/rfbi.c @@ -1181,6 +1181,7 @@ int rfbi_init(void) /* Enable autoidle and smart-idle */ l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~((0x03 3)|(0x01 0)); l |= (1 0) | (2 3); rfbi_write_reg(RFBI_SYSCONFIG, l); @@ -1208,6 +1209,9 @@ static int rfbi_display_update(struct omap_dss_device *dssdev, { int rfbi_module; + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return 0; + if (w == 0 || h == 0) return 0; @@ -1239,6 +1243,18 @@ static int rfbi_display_enable_te(struct omap_dss_device *dssdev, bool enable) return 0; } +static enum omap_dss_update_mode rfbi_display_get_update_mode( + struct omap_dss_device *dssdev) +{ + return OMAP_DSS_UPDATE_MANUAL; +} + +static int rfbi_display_set_update_mode(struct omap_dss_device *dssdev, + enum omap_dss_update_mode mode) +{ + return 0; +} + static int rfbi_display_enable(struct omap_dss_device *dssdev) { int r; @@ -1269,6 +1285,7 @@ static int rfbi_display_enable(struct omap_dss_device *dssdev) rfbi_set_timings(dssdev-phy.rfbi.channel, dssdev-ctrl.rfbi_timings); + dssdev-state = OMAP_DSS_DISPLAY_ACTIVE; if (dssdev-driver-enable) { r = dssdev-driver-enable(dssdev); @@ -1288,12 +1305,67 @@ err0: static void rfbi_display_disable(struct omap_dss_device *dssdev) { + dssdev-state = OMAP_DSS_DISPLAY_DISABLED; + dssdev-driver-disable(dssdev); omap_dispc_unregister_isr(framedone_callback, NULL, DISPC_IRQ_FRAMEDONE); omap_dss_stop_device(dssdev); } +static int rfbi_display_suspend(struct omap_dss_device *dssdev) { + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return -EINVAL; + + DSSDBG(rfbi_display_suspend\n); + + if (dssdev-driver-suspend) + dssdev-driver-suspend(dssdev); + + dispc_enable_lcd_out(0); I don't think this is needed. DSS hardware disables lcd_out when the transfer has finished. Although for correct operation suspend() should wait until the last transfer has been done before disabling clocks, which is something it currently doesn't. This was done with reference to the DPI and SDI files that do the same thing. It can be removed if not necessary. + + /* Force idle */ + rfbi_enable_clocks(1); + l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~(0x03 3); + rfbi_write_reg(RFBI_SYSCONFIG, l); + rfbi_enable_clocks(0); Why force idle? The RFBI module must be forced to idle on suspend to allow for the DSS module to idle. The CM_CLKSTST_DSS[CLKACTIVITY_DSS] bit does not change if the RFBI module was configured to autoidle, preventing DSS from entering retention or off. + + dssdev-state = OMAP_DSS_DISPLAY_SUSPENDED; + + return 0; +} + +static int rfbi_display_resume(struct omap_dss_device *dssdev) { + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_SUSPENDED) + return -EINVAL; + + DSSDBG(rfbi_display_resume\n); + + /* Re-enable autoidle */ + rfbi_enable_clocks(1); + l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~(0x03 3); + l |= (0x02 3); + rfbi_write_reg(RFBI_SYSCONFIG, l); + rfbi_enable_clocks(0); + + dispc_enable_lcd_out(1); This is not needed. lcd_out is enabled when the transfer is started. This was done with reference to the DPI and SDI files that do the same thing. It can be removed if not necessary. + + if (dssdev-driver-resume) + dssdev-driver-resume(dssdev); + + dssdev-state = OMAP_DSS_DISPLAY_ACTIVE; + + return 0; +} + int rfbi_init_display(struct omap_dss_device *dssdev) { dssdev-enable = rfbi_display_enable; @@ -1301,6 +1373,10 @@ int rfbi_init_display(struct omap_dss_device *dssdev) dssdev-update = rfbi_display_update; dssdev-sync = rfbi_display_sync; dssdev-enable_te = rfbi_display_enable_te; + dssdev-get_update_mode = rfbi_display_get_update_mode; + dssdev-set_update_mode = rfbi_display_set_update_mode; + dssdev-suspend = rfbi_display_suspend; + dssdev-resume = rfbi_display_resume
RE: [PATCH 1/1] OMAP: DSS2: RFBI driver update
On Mon, 2009-10-12 at 16:03 +0200, ext Christensen, Mikkel wrote: On Mon, Oct 12, 2009 at 03:25:27, Tomi Valkeinen wrote: Subject: Re: [PATCH 1/1] OMAP: DSS2: RFBI driver update On Fri, 2009-10-09 at 23:08 +0200, ext Mikkel Christensen wrote: This patch adds suspend / resume functionality to the RFBI driver along with missing callback functions needed by OMAP Frame buffer. Signed-off-by: Mikkel Christensen m...@ti.com --- drivers/video/omap2/dss/rfbi.c | 76 1 files changed, 76 insertions(+), 0 deletions(-) snip +static int rfbi_display_suspend(struct omap_dss_device *dssdev) { + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return -EINVAL; + + DSSDBG(rfbi_display_suspend\n); + + if (dssdev-driver-suspend) + dssdev-driver-suspend(dssdev); + + dispc_enable_lcd_out(0); I don't think this is needed. DSS hardware disables lcd_out when the transfer has finished. Although for correct operation suspend() should wait until the last transfer has been done before disabling clocks, which is something it currently doesn't. This was done with reference to the DPI and SDI files that do the same thing. It can be removed if not necessary. DPI and SDI work quite differently than RFBI, you can't just copy their functionality =). It is more correct without disable/enable_lcd_out, but it's not correct as there may be a transfer ongoing when suspend call comes. That's why there should be some mechanism to wait until the transfer has been done. DSI tries to do this (and hopefully correctly does =). + + /* Force idle */ + rfbi_enable_clocks(1); + l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~(0x03 3); + rfbi_write_reg(RFBI_SYSCONFIG, l); + rfbi_enable_clocks(0); Why force idle? The RFBI module must be forced to idle on suspend to allow for the DSS module to idle. The CM_CLKSTST_DSS[CLKACTIVITY_DSS] bit does not change if the RFBI module was configured to autoidle, preventing DSS from entering retention or off. Why must it? Is there a hardware bug (what errata number?)? Tomi -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
RE: [PATCH 1/1] OMAP: DSS2: RFBI driver update
On Mon, Oct 12, 2009 at 09:17:27, Tomi Valkeinen wrote: Subject: RE: [PATCH 1/1] OMAP: DSS2: RFBI driver update On Mon, 2009-10-12 at 16:03 +0200, ext Christensen, Mikkel wrote: On Mon, Oct 12, 2009 at 03:25:27, Tomi Valkeinen wrote: Subject: Re: [PATCH 1/1] OMAP: DSS2: RFBI driver update On Fri, 2009-10-09 at 23:08 +0200, ext Mikkel Christensen wrote: This patch adds suspend / resume functionality to the RFBI driver along with missing callback functions needed by OMAP Frame buffer. Signed-off-by: Mikkel Christensen m...@ti.com --- drivers/video/omap2/dss/rfbi.c | 76 1 files changed, 76 insertions(+), 0 deletions(-) snip +static int rfbi_display_suspend(struct omap_dss_device *dssdev) { + unsigned long l; + + if (dssdev-state != OMAP_DSS_DISPLAY_ACTIVE) + return -EINVAL; + + DSSDBG(rfbi_display_suspend\n); + + if (dssdev-driver-suspend) + dssdev-driver-suspend(dssdev); + + dispc_enable_lcd_out(0); I don't think this is needed. DSS hardware disables lcd_out when the transfer has finished. Although for correct operation suspend() should wait until the last transfer has been done before disabling clocks, which is something it currently doesn't. This was done with reference to the DPI and SDI files that do the same thing. It can be removed if not necessary. DPI and SDI work quite differently than RFBI, you can't just copy their functionality =). It is more correct without disable/enable_lcd_out, but it's not correct as there may be a transfer ongoing when suspend call comes. That's why there should be some mechanism to wait until the transfer has been done. DSI tries to do this (and hopefully correctly does =). OK. Thanks I will try to incorporate this. Where in the DSI driver is this mechanism implemented? + + /* Force idle */ + rfbi_enable_clocks(1); + l = rfbi_read_reg(RFBI_SYSCONFIG); + l = ~(0x03 3); + rfbi_write_reg(RFBI_SYSCONFIG, l); + rfbi_enable_clocks(0); Why force idle? The RFBI module must be forced to idle on suspend to allow for the DSS module to idle. The CM_CLKSTST_DSS[CLKACTIVITY_DSS] bit does not change if the RFBI module was configured to autoidle, preventing DSS from entering retention or off. Why must it? Is there a hardware bug (what errata number?)? Tomi -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html