* omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 @ 2013-05-15 13:34 jean-philippe francois 2013-05-16 7:54 ` jean-philippe francois 2013-05-16 8:21 ` jean-philippe francois 0 siblings, 2 replies; 5+ messages in thread From: jean-philippe francois @ 2013-05-15 13:34 UTC (permalink / raw) To: linux-omap, linux-media; +Cc: Laurent Pinchart Hi, I am working on a dm3730 based camera. The sensor input clock is provided by the cpu via the CAM_XCLK pin. Here is the corresponding log : [ 9.115966] Entering cam_set_xclk [ 9.119781] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz [ 9.121337] ov10x33 1-0010: sensor id : 0xa630 [ 10.293640] Entering cam_set_xclk [ 10.297149] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 0 Hz [ 10.393920] Entering cam_set_xclk [ 10.397979] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz However, when mesured on the actual pin, the frequency is around 30 MHz The crystal clock is 19.2 MHz All this was correct with 3.6.11. Jean-Philippe Francois ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 2013-05-15 13:34 omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 jean-philippe francois @ 2013-05-16 7:54 ` jean-philippe francois 2013-05-16 8:21 ` jean-philippe francois 1 sibling, 0 replies; 5+ messages in thread From: jean-philippe francois @ 2013-05-16 7:54 UTC (permalink / raw) To: linux-omap, linux-media; +Cc: Laurent Pinchart 2013/5/15 jean-philippe francois <jp.francois@cynove.com>: > Hi, > > I am working on a dm3730 based camera. > The sensor input clock is provided by the cpu via the CAM_XCLK pin. > Here is the corresponding log : > > [ 9.115966] Entering cam_set_xclk > [ 9.119781] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz > [ 9.121337] ov10x33 1-0010: sensor id : 0xa630 > [ 10.293640] Entering cam_set_xclk > [ 10.297149] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 0 Hz > [ 10.393920] Entering cam_set_xclk > [ 10.397979] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz > > However, when mesured on the actual pin, the frequency is around 30 MHz > > The crystal clock is 19.2 MHz > All this was correct with 3.6.11. It seems the dpll4_m5_ck is not correctly set, 3.6.11 code in isp.c (without error handling) r = clk_set_rate(isp->clock[ISP_CLK_DPLL4_M5_CK], CM_CAM_MCLK_HZ/divisor); ... r = clk_enable(isp->clock[ISP_CLK_CAM_MCLK]); 3.9 code in isp.c (without error handling) r = clk_set_rate(isp->clock[ISP_CLK_CAM_MCLK], CM_CAM_MCLK_HZ/divisor); > > Jean-Philippe Francois ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 2013-05-15 13:34 omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 jean-philippe francois 2013-05-16 7:54 ` jean-philippe francois @ 2013-05-16 8:21 ` jean-philippe francois 2013-05-16 11:21 ` Laurent Pinchart 1 sibling, 1 reply; 5+ messages in thread From: jean-philippe francois @ 2013-05-16 8:21 UTC (permalink / raw) To: linux-omap, linux-media; +Cc: Laurent Pinchart 2013/5/15 jean-philippe francois <jp.francois@cynove.com>: > Hi, > > I am working on a dm3730 based camera. > The sensor input clock is provided by the cpu via the CAM_XCLK pin. > Here is the corresponding log : > > [ 9.115966] Entering cam_set_xclk > [ 9.119781] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz > [ 9.121337] ov10x33 1-0010: sensor id : 0xa630 > [ 10.293640] Entering cam_set_xclk > [ 10.297149] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 0 Hz > [ 10.393920] Entering cam_set_xclk > [ 10.397979] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz > > However, when mesured on the actual pin, the frequency is around 30 MHz > > The crystal clock is 19.2 MHz > All this was correct with 3.6.11. > Sorry for the resend, wrong tab and enter key sequence in gmail ... It seems the dpll4_m5_ck is not correctly set, 3.6.11 code in isp.c (without error handling) r = clk_set_rate(isp->clock[ISP_CLK_DPLL4_M5_CK], CM_CAM_MCLK_HZ/divisor); ... r = clk_enable(isp->clock[ISP_CLK_CAM_MCLK]); 3.9 code in isp.c (without error handling) r = clk_set_rate(isp->clock[ISP_CLK_CAM_MCLK], CM_CAM_MCLK_HZ/divisor); r = clk_prepare_enable(isp->clock[ISP_CLK_CAM_MCLK]); The PLL settings ie multiplier and divisor are the same in each case, but the dmesg output differ : Here is what happens when isp_enable_clock is called on 3.6 3.6 [ 10.133697] Entering cam_set_xclk [ 10.137573] clock: clksel_round_rate_div: dpll4_m5_ck target_rate 172800000 [ 10.137573] clock: new_div = 5, new_rate = 172800000 [ 10.137603] clock: dpll4_m5_ck: set rate to 172800000 [ 10.138061] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz 3.9 [ 9.095581] Entering cam_set_xclk [ 9.102661] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 24685714 Hz So the frequency setting register are correctly set, but the actual output frequency is not. maybe dpll4 is not correctly locked ? I will also check isp_enable_clock is really called. But I suppose it is, because except for the frequency, everything is working correctly. ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 2013-05-16 8:21 ` jean-philippe francois @ 2013-05-16 11:21 ` Laurent Pinchart 2013-05-16 13:52 ` jean-philippe francois 0 siblings, 1 reply; 5+ messages in thread From: Laurent Pinchart @ 2013-05-16 11:21 UTC (permalink / raw) To: jean-philippe francois; +Cc: linux-omap, linux-media Hi Jean-Philippe, On Thursday 16 May 2013 10:21:14 jean-philippe francois wrote: > 2013/5/15 jean-philippe francois <jp.francois@cynove.com>: > > Hi, > > > > I am working on a dm3730 based camera. > > The sensor input clock is provided by the cpu via the CAM_XCLK pin. > > Here is the corresponding log : > > > > [ 9.115966] Entering cam_set_xclk > > [ 9.119781] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to > > 24685714 Hz > > [ 9.121337] ov10x33 1-0010: sensor id : 0xa630 > > [ 10.293640] Entering cam_set_xclk > > [ 10.297149] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 0 Hz > > [ 10.393920] Entering cam_set_xclk > > [ 10.397979] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to > > 24685714 Hz > > > > However, when mesured on the actual pin, the frequency is around 30 MHz > > > > The crystal clock is 19.2 MHz > > All this was correct with 3.6.11. > > Sorry for the resend, wrong tab and enter key sequence in gmail ... > > It seems the dpll4_m5_ck is not correctly set, > 3.6.11 code in isp.c (without error handling) > > r = clk_set_rate(isp->clock[ISP_CLK_DPLL4_M5_CK], > CM_CAM_MCLK_HZ/divisor); > ... > r = clk_enable(isp->clock[ISP_CLK_CAM_MCLK]); > > 3.9 code in isp.c (without error handling) > > r = clk_set_rate(isp->clock[ISP_CLK_CAM_MCLK], > CM_CAM_MCLK_HZ/divisor); > > r = clk_prepare_enable(isp->clock[ISP_CLK_CAM_MCLK]); > > The PLL settings ie multiplier and divisor are the same in each case, > but the dmesg output differ : > Here is what happens when isp_enable_clock is called on 3.6 > > 3.6 > [ 10.133697] Entering cam_set_xclk > [ 10.137573] clock: clksel_round_rate_div: dpll4_m5_ck > target_rate 172800000 > [ 10.137573] clock: new_div = 5, new_rate = 172800000 > [ 10.137603] clock: dpll4_m5_ck: set rate to 172800000 > [ 10.138061] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to > 24685714 Hz > > 3.9 > [ 9.095581] Entering cam_set_xclk > [ 9.102661] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to > 24685714 Hz > > So the frequency setting register are correctly set, but the actual > output frequency is not. > maybe dpll4 is not correctly locked ? I will also check > isp_enable_clock is really called. > But I suppose it is, because except for the frequency, everything is > working correctly. Does the following patch fix the issue ? commit 577f8ea9ba7b1276096713b8148b3a8fca96d805 Author: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Date: Thu May 16 13:15:40 2013 +0200 ARM: OMAP3: clock: Back-propagate rate change from cam_mclk to dpll4_m5 on all OMAP3 platforms Commit 7b2e1277598e4187c9be3e61fd9b0f0423f97986 ("ARM: OMAP3: clock: Back-propagate rate change from cam_mclk to dpll4_m5") enabled clock rate back-propagation from cam_mclk do dpll4_m5 on OMAP3630 only. Perform back-propagation on other OMAP3 platforms as well. Reported-by: Jean-Philippe François <jp.francois@cynove.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> diff --git a/arch/arm/mach-omap2/cclock3xxx_data.c b/arch/arm/mach-omap2/cclock3xxx_data.c index 4579c3c..c21065a 100644 --- a/arch/arm/mach-omap2/cclock3xxx_data.c +++ b/arch/arm/mach-omap2/cclock3xxx_data.c @@ -418,7 +418,8 @@ static struct clk_hw_omap dpll4_m5x2_ck_hw = { .clkdm_name = "dpll4_clkdm", }; -DEFINE_STRUCT_CLK(dpll4_m5x2_ck, dpll4_m5x2_ck_parent_names, dpll4_m5x2_ck_ops); +DEFINE_STRUCT_CLK_FLAGS(dpll4_m5x2_ck, dpll4_m5x2_ck_parent_names, + dpll4_m5x2_ck_ops, CLK_SET_RATE_PARENT); static struct clk dpll4_m5x2_ck_3630 = { .name = "dpll4_m5x2_ck", -- Regards, Laurent Pinchart ^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 2013-05-16 11:21 ` Laurent Pinchart @ 2013-05-16 13:52 ` jean-philippe francois 0 siblings, 0 replies; 5+ messages in thread From: jean-philippe francois @ 2013-05-16 13:52 UTC (permalink / raw) To: Laurent Pinchart; +Cc: linux-omap, linux-media 2013/5/16 Laurent Pinchart <laurent.pinchart@ideasonboard.com>: > Hi Jean-Philippe, > > On Thursday 16 May 2013 10:21:14 jean-philippe francois wrote: >> 2013/5/15 jean-philippe francois <jp.francois@cynove.com>: >> > Hi, >> > >> > I am working on a dm3730 based camera. >> > The sensor input clock is provided by the cpu via the CAM_XCLK pin. >> > Here is the corresponding log : >> > >> > [ 9.115966] Entering cam_set_xclk >> > [ 9.119781] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to >> > 24685714 Hz >> > [ 9.121337] ov10x33 1-0010: sensor id : 0xa630 >> > [ 10.293640] Entering cam_set_xclk >> > [ 10.297149] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to 0 Hz >> > [ 10.393920] Entering cam_set_xclk >> > [ 10.397979] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to >> > 24685714 Hz >> > >> > However, when mesured on the actual pin, the frequency is around 30 MHz >> > >> > The crystal clock is 19.2 MHz >> > All this was correct with 3.6.11. >> >> Sorry for the resend, wrong tab and enter key sequence in gmail ... >> >> It seems the dpll4_m5_ck is not correctly set, >> 3.6.11 code in isp.c (without error handling) >> >> r = clk_set_rate(isp->clock[ISP_CLK_DPLL4_M5_CK], >> CM_CAM_MCLK_HZ/divisor); >> ... >> r = clk_enable(isp->clock[ISP_CLK_CAM_MCLK]); >> >> 3.9 code in isp.c (without error handling) >> >> r = clk_set_rate(isp->clock[ISP_CLK_CAM_MCLK], >> CM_CAM_MCLK_HZ/divisor); >> >> r = clk_prepare_enable(isp->clock[ISP_CLK_CAM_MCLK]); >> >> The PLL settings ie multiplier and divisor are the same in each case, >> but the dmesg output differ : >> Here is what happens when isp_enable_clock is called on 3.6 >> >> 3.6 >> [ 10.133697] Entering cam_set_xclk >> [ 10.137573] clock: clksel_round_rate_div: dpll4_m5_ck >> target_rate 172800000 >> [ 10.137573] clock: new_div = 5, new_rate = 172800000 >> [ 10.137603] clock: dpll4_m5_ck: set rate to 172800000 >> [ 10.138061] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to >> 24685714 Hz >> >> 3.9 >> [ 9.095581] Entering cam_set_xclk >> [ 9.102661] omap3isp omap3isp: isp_set_xclk(): cam_xclka set to >> 24685714 Hz >> >> So the frequency setting register are correctly set, but the actual >> output frequency is not. >> maybe dpll4 is not correctly locked ? I will also check >> isp_enable_clock is really called. >> But I suppose it is, because except for the frequency, everything is >> working correctly. > > Does the following patch fix the issue ? > Hi Laurent, thank you for your help. I am on a dm3730, si I already had the back propagation thing. I tried the patch anyway, but it did not solve the issue. If I look at the registers, on both version : 0x48004d44 contains 0x04816807 whiche gives a dpll4 frequency of 864 MHz 0x48004f40 contains 0x5 which gives a cam_mclk frequency of 172,8 MHz 0x480bc050 (ISP.TCTRL_CTRL) contains 0x7 which should give a 24 MHz frequency In my dmesg output, I never see a message like the following but for dpll4_ck. omap3_noncore_dpll_set_rate: dpll5_ck: set rate: locking rate to 120000000. I wonder why. > commit 577f8ea9ba7b1276096713b8148b3a8fca96d805 > Author: Laurent Pinchart <laurent.pinchart@ideasonboard.com> > Date: Thu May 16 13:15:40 2013 +0200 > > ARM: OMAP3: clock: Back-propagate rate change from cam_mclk to dpll4_m5 on all OMAP3 platforms > > Commit 7b2e1277598e4187c9be3e61fd9b0f0423f97986 ("ARM: OMAP3: clock: > Back-propagate rate change from cam_mclk to dpll4_m5") enabled clock > rate back-propagation from cam_mclk do dpll4_m5 on OMAP3630 only. > Perform back-propagation on other OMAP3 platforms as well. > > Reported-by: Jean-Philippe François <jp.francois@cynove.com> > Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> > > diff --git a/arch/arm/mach-omap2/cclock3xxx_data.c b/arch/arm/mach-omap2/cclock3xxx_data.c > index 4579c3c..c21065a 100644 > --- a/arch/arm/mach-omap2/cclock3xxx_data.c > +++ b/arch/arm/mach-omap2/cclock3xxx_data.c > @@ -418,7 +418,8 @@ static struct clk_hw_omap dpll4_m5x2_ck_hw = { > .clkdm_name = "dpll4_clkdm", > }; > > -DEFINE_STRUCT_CLK(dpll4_m5x2_ck, dpll4_m5x2_ck_parent_names, dpll4_m5x2_ck_ops); > +DEFINE_STRUCT_CLK_FLAGS(dpll4_m5x2_ck, dpll4_m5x2_ck_parent_names, > + dpll4_m5x2_ck_ops, CLK_SET_RATE_PARENT); > > static struct clk dpll4_m5x2_ck_3630 = { > .name = "dpll4_m5x2_ck", > > > -- > Regards, > > Laurent Pinchart > ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2013-05-16 13:52 UTC | newest] Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2013-05-15 13:34 omap3 : isp clock a : Difference between dmesg frequency and actual frequency with 3.9 jean-philippe francois 2013-05-16 7:54 ` jean-philippe francois 2013-05-16 8:21 ` jean-philippe francois 2013-05-16 11:21 ` Laurent Pinchart 2013-05-16 13:52 ` jean-philippe francois
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.