From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jingoo Han Date: Tue, 06 Mar 2012 07:26:59 +0000 Subject: RE: [PATCH 1/3] video: s3c-fb: move video interface timing out of window setup data Message-Id: <002501ccfb6a$84a765d0$8df63170$%han@samsung.com> List-Id: References: <1330789808-8253-1-git-send-email-thomas.abraham@linaro.org> <1330789808-8253-2-git-send-email-thomas.abraham@linaro.org> <001f01ccfb53$f034ef00$d09ecd00$%han@samsung.com> <002101ccfb61$710d2c70$53278550$%han@samsung.com> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable To: 'Thomas Abraham' Cc: linux-fbdev@vger.kernel.org, FlorianSchandinat@gmx.de, linux-samsung-soc@vger.kernel.org, kgene.kim@samsung.com, ben-linux@fluff.org, patches@linaro.org, 'Jingoo Han' > -----Original Message----- > From: Thomas Abraham [mailto:thomas.abraham@linaro.org] > Sent: Tuesday, March 06, 2012 3:35 PM > To: Jingoo Han > Cc: linux-fbdev@vger.kernel.org; FlorianSchandinat@gmx.de; linux-samsung-= soc@vger.kernel.org; > kgene.kim@samsung.com; ben-linux@fluff.org; patches@linaro.org > Subject: Re: [PATCH 1/3] video: s3c-fb: move video interface timing out o= f window setup data >=20 > On 6 March 2012 11:52, Jingoo Han wrote: >=20 > >> -----Original Message----- > >> From: Thomas Abraham [mailto:thomas.abraham@linaro.org] > >> Sent: Tuesday, March 06, 2012 2:26 PM > >> To: Jingoo Han > >> Cc: linux-fbdev@vger.kernel.org; FlorianSchandinat@gmx.de; linux-samsu= ng-soc@vger.kernel.org; > >> kgene.kim@samsung.com; ben-linux@fluff.org; patches@linaro.org > >> Subject: Re: [PATCH 1/3] video: s3c-fb: move video interface timing ou= t of window setup data > >> > >> On 6 March 2012 10:15, Jingoo Han wrote: > >> > >> >> -----Original Message----- > >> >> From: Thomas Abraham [mailto:thomas.abraham@linaro.org] > >> >> Sent: Sunday, March 04, 2012 12:50 AM > >> >> To: linux-fbdev@vger.kernel.org > >> >> Cc: FlorianSchandinat@gmx.de; linux-samsung-soc@vger.kernel.org; kg= ene.kim@samsung.com; > >> >> jg1.han@samsung.com; ben-linux@fluff.org; patches@linaro.org > >> >> Subject: [PATCH 1/3] video: s3c-fb: move video interface timing out= of window setup data > >> >> > >> >> The video interface timing is independent of the window setup data. > >> >> The resolution of the window can be smaller than that of the lcd > >> >> panel to which the video data is output. > >> >> > >> >> So move the video timing data from the per-window setup data to the > >> >> platform specific section in the platform data. This also removes > >> >> the restriction that atleast one window should have the same > >> >> resolution as that of the panel attached. > >> >> > >> >> Cc: Ben Dooks > >> >> Cc: Jingoo Han > >> >> Signed-off-by: Thomas Abraham > >> >> --- > >> >> =A0arch/arm/plat-samsung/include/plat/fb.h | =A0 =A09 ++- > >> >> =A0drivers/video/s3c-fb.c =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0| =A01= 06 +++++++++++++++++-------------- > >> >> =A02 files changed, 63 insertions(+), 52 deletions(-) > >> >> > >> >> diff --git a/arch/arm/plat-samsung/include/plat/fb.h b/arch/arm/pla= t-samsung/include/plat/fb.h > >> >> index 0fedf47..39d6bd7 100644 > >> >> --- a/arch/arm/plat-samsung/include/plat/fb.h > >> >> +++ b/arch/arm/plat-samsung/include/plat/fb.h > >> >> @@ -24,15 +24,16 @@ > >> >> > >> >> =A0/** > >> >> =A0 * struct s3c_fb_pd_win - per window setup data > >> >> - * @win_mode: The display parameters to initialise (not for window= 0) > >> >> + * @xres =A0 =A0 : The window X size. > >> >> + * @yres =A0 =A0 : The window Y size. > >> >> =A0 * @virtual_x: The virtual X size. > >> >> =A0 * @virtual_y: The virtual Y size. > >> >> =A0 */ > >> >> =A0struct s3c_fb_pd_win { > >> >> - =A0 =A0 struct fb_videomode =A0 =A0 win_mode; > >> >> - > >> >> =A0 =A0 =A0 unsigned short =A0 =A0 =A0 =A0 =A0default_bpp; > >> >> =A0 =A0 =A0 unsigned short =A0 =A0 =A0 =A0 =A0max_bpp; > >> >> + =A0 =A0 unsigned short =A0 =A0 =A0 =A0 =A0xres; > >> >> + =A0 =A0 unsigned short =A0 =A0 =A0 =A0 =A0yres; > >> >> =A0 =A0 =A0 unsigned short =A0 =A0 =A0 =A0 =A0virtual_x; > >> >> =A0 =A0 =A0 unsigned short =A0 =A0 =A0 =A0 =A0virtual_y; > >> >> =A0}; > >> >> @@ -45,6 +46,7 @@ struct s3c_fb_pd_win { > >> >> =A0 * @default_win: default window layer number to be used for UI l= ayer. > >> >> =A0 * @vidcon0: The base vidcon0 values to control the panel data f= ormat. > >> >> =A0 * @vidcon1: The base vidcon1 values to control the panel data o= utput. > >> >> + * @vtiming: Video timing when connected to a RGB type panel. > >> > > >> > fb_videomode can be set, even if it is not RGB type panel. > >> > In my opinion, it would be better. > >> > + * @vtiming: The video timing values to set the interface timing of= the panel. > >> > >> The other interface that is supported is the i80 interface. Can these > >> timing values be used for i80 interface as well ? > > > > No, you're right. The i80 is not supported. > > Please ignore my comment. > > > >> > >> > > >> >> =A0 * @win: The setup data for each hardware window, or NULL for un= used. > >> >> =A0 * @display_mode: The LCD output display mode. > >> >> =A0 * > >> >> @@ -58,6 +60,7 @@ struct s3c_fb_platdata { > >> >> =A0 =A0 =A0 void =A0 =A0(*setup_gpio)(void); > >> >> > >> >> =A0 =A0 =A0 struct s3c_fb_pd_win =A0 =A0*win[S3C_FB_MAX_WIN]; > >> >> + =A0 =A0 struct fb_videomode =A0 =A0 *vtiming; > >> >> > >> >> =A0 =A0 =A0 u32 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0default_= win; > >> >> > >> >> diff --git a/drivers/video/s3c-fb.c b/drivers/video/s3c-fb.c > >> >> index 1fb7ddf..8e05d4d 100644 > >> >> --- a/drivers/video/s3c-fb.c > >> >> +++ b/drivers/video/s3c-fb.c > >> >> @@ -495,7 +495,6 @@ static int s3c_fb_set_par(struct fb_info *info) > >> >> =A0 =A0 =A0 u32 alpha =3D 0; > >> >> =A0 =A0 =A0 u32 data; > >> >> =A0 =A0 =A0 u32 pagewidth; > >> >> - =A0 =A0 int clkdiv; > >> >> > >> >> =A0 =A0 =A0 dev_dbg(sfb->dev, "setting framebuffer parameters\n"); > >> >> > >> >> @@ -532,46 +531,9 @@ static int s3c_fb_set_par(struct fb_info *info) > >> >> =A0 =A0 =A0 /* disable the window whilst we update it */ > >> >> =A0 =A0 =A0 writel(0, regs + WINCON(win_no)); > >> >> > >> >> - =A0 =A0 /* use platform specified window as the basis for the lcd= timings */ > >> >> - > >> >> - =A0 =A0 if (win_no =3D sfb->pdata->default_win) { > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 clkdiv =3D s3c_fb_calc_pixclk(sfb, var->p= ixclock); > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 data =3D sfb->pdata->vidcon0; > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 data &=3D ~(VIDCON0_CLKVAL_F_MASK | VIDCO= N0_CLKDIR); > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 if (clkdiv > 1) > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 data |=3D VIDCON0_CLKVAL_= F(clkdiv-1) | VIDCON0_CLKDIR; > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 else > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 data &=3D ~VIDCON0_CLKDIR= ; =A0 =A0 =A0 =A0/* 1:1 clock */ > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 /* write the timing data to the panel */ > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 if (sfb->variant.is_2443) > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 data |=3D (1 << 5); > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 writel(data, regs + VIDCON0); > >> >> - > >> >> + =A0 =A0 if (win_no =3D sfb->pdata->default_win) > >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 s3c_fb_enable(sfb, 1); > >> >> > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 data =3D VIDTCON0_VBPD(var->upper_margin = - 1) | > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0VIDTCON0_VFPD(var->lower_m= argin - 1) | > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0VIDTCON0_VSPW(var->vsync_l= en - 1); > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 writel(data, regs + sfb->variant.vidtcon); > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 data =3D VIDTCON1_HBPD(var->left_margin -= 1) | > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0VIDTCON1_HFPD(var->right_m= argin - 1) | > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0VIDTCON1_HSPW(var->hsync_l= en - 1); > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 /* VIDTCON1 */ > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 writel(data, regs + sfb->variant.vidtcon = + 4); > >> >> - > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 data =3D VIDTCON2_LINEVAL(var->yres - 1) | > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0VIDTCON2_HOZVAL(var->xres = - 1); > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 writel(data, regs + sfb->variant.vidtcon = + 8); > >> >> - =A0 =A0 } > >> >> - > >> > > >> > It looks good. > >> > VIDTCON registers don't need to be written whenever s3c_fb_set_par i= s called. > >> > > >> >> =A0 =A0 =A0 /* write the buffer address */ > >> >> > >> >> =A0 =A0 =A0 /* start and end registers stride is 8 */ > >> >> @@ -1136,11 +1098,11 @@ static int __devinit s3c_fb_alloc_memory(st= ruct s3c_fb *sfb, > >> >> > >> >> =A0 =A0 =A0 dev_dbg(sfb->dev, "allocating memory for display\n"); > >> >> > >> >> - =A0 =A0 real_size =3D windata->win_mode.xres * windata->win_mode.= yres; > >> >> + =A0 =A0 real_size =3D windata->xres * windata->yres; > >> >> =A0 =A0 =A0 virt_size =3D windata->virtual_x * windata->virtual_y; > >> >> > >> >> =A0 =A0 =A0 dev_dbg(sfb->dev, "real_size=3D%u (%u.%u), virt_size=3D= %u (%u.%u)\n", > >> >> - =A0 =A0 =A0 =A0 =A0 =A0 real_size, windata->win_mode.xres, windat= a->win_mode.yres, > >> >> + =A0 =A0 =A0 =A0 =A0 =A0 real_size, windata->xres, windata->yres, > >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 virt_size, windata->virtual_x, windata-= >virtual_y); > >> >> > >> >> =A0 =A0 =A0 size =3D (real_size > virt_size) ? real_size : virt_siz= e; > >> >> @@ -1222,7 +1184,7 @@ static int __devinit s3c_fb_probe_win(struct = s3c_fb *sfb, unsigned int > win_no, > >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 struct s3c_fb_win **res) > >> >> =A0{ > >> >> =A0 =A0 =A0 struct fb_var_screeninfo *var; > >> >> - =A0 =A0 struct fb_videomode *initmode; > >> >> + =A0 =A0 struct fb_videomode initmode; > >> > > >> > *initmode cannot be used??? > >> > Can you tell me why pointer type should be changed? > >> > > >> > >> The initmode is used to pass video timing to the fb_videomode_to_var() > >> function. Each window setup data in platform data included video > >> timing information. Since, the video timing data is now moved out of > >> per-window data, the xres and yres values have to be setup based on > >> the window for which the fb_videomode_to_var() is called. Hence, the > >> common timing values is first copied into initmode and then the window > >> specific xres and yres are set. If initmode is a maintained as a > >> pointer (to the video timing data in platform data), then any xres and > >> yres update to initmode would overwrite the 'constant' video timing. > > > > My point is that variable type 'initmode' can be not changed by using p= ointer type as follows: > > > > @@ -1243,11 +1243,11 @@ static int __devinit s3c_fb_probe_win(struct s3= c_fb *sfb, unsigned int win_no, > > =A0 =A0 =A0 =A0} > > > > =A0 =A0 =A0 =A0windata =3D sfb->pdata->win[win_no]; > > - =A0 =A0 =A0 initmode =3D &windata->win_mode; > > + =A0 =A0 =A0 initmode =3D sfb->pdata->vtiming; >=20 > In this case, initmode is not pointing to the 'constant' lcd panel > timing values. >=20 > > > > =A0 =A0 =A0 =A0WARN_ON(windata->max_bpp =3D 0); > > - =A0 =A0 =A0 WARN_ON(windata->win_mode.xres =3D 0); > > - =A0 =A0 =A0 WARN_ON(windata->win_mode.yres =3D 0); > > + =A0 =A0 =A0 WARN_ON(windata->xres =3D 0); > > + =A0 =A0 =A0 WARN_ON(windata->yres =3D 0); > > > > =A0 =A0 =A0 =A0win =3D fbinfo->par; > > =A0 =A0 =A0 =A0*res =3D win; > > @@ -1286,6 +1286,8 @@ static int __devinit s3c_fb_probe_win(struct s3c_= fb *sfb, unsigned int win_no, > > =A0 =A0 =A0 =A0} > > > > =A0 =A0 =A0 =A0/* setup the initial video mode from the window */ > > + =A0 =A0 =A0 initmode->xres =3D windata->xres; > > + =A0 =A0 =A0 initmode->yres =3D windata->yres; >=20 > And then here, the xres and yres of the 'constant' lcd panel timing > value will be changed. >=20 > But, we don't want to change the actual lcd panel timing value. That > is the reason to first copy the lcd panel timing value into a local > copy and then update the xres and yres and then pass this local copy > of the timing value to fb_videomode_to_var() function. OK. I see. You're right. If you add add it in the resume path in the resume path, you can use my acked-by to this 1st patch. Thank you for sending the patch. Best regards, Jingoo Han >=20 > Thanks, > Thomas. >=20 > [...]