[U-Boot] [PATCH 1/3] BeagleBoard: turn off clocks in ehci_stop
From: Koen Kooi koen@dominion.thruhere.net
This fixes display problems in linux
Signed-off-by: Koen Kooi koen@dominion.thruhere.net Signed-off-by: Joel A Fernandes agnel.joel@gmail.com --- board/ti/beagle/beagle.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/board/ti/beagle/beagle.c b/board/ti/beagle/beagle.c index 7768901..a958545 100644 --- a/board/ti/beagle/beagle.c +++ b/board/ti/beagle/beagle.c @@ -356,6 +356,12 @@ int ehci_hcd_stop(void) pr_debug("Resetting OMAP3 EHCI\n"); omap_set_gpio_dataout(GPIO_PHY_RESET, 0); writel(OMAP_UHH_SYSCONFIG_SOFTRESET, OMAP3_UHH_BASE + OMAP_UHH_SYSCONFIG); + /* disable USB clocks */ + struct prcm *prcm_base = (struct prcm *)PRCM_BASE; + sr32(&prcm_base->iclken_usbhost, 0, 1, 0); + sr32(&prcm_base->fclken_usbhost, 0, 2, 0); + sr32(&prcm_base->iclken3_core, 2, 1, 0); + sr32(&prcm_base->fclken3_core, 2, 1, 0); return 0; }
From: Jason Kridner jkridner@beagleboard.org
Based on commit f1099c7c43caf5bac3bf6a65aa266fade4747072 Author: Greg Turner gregturner@ti.com Date: Tue May 25 09:19:06 2010 -0500
New u-boot command for status of USER button on BeagleBoard-xM
Modified bootcmd to check the staus at boot time and set filename of the boot script.
* Moved to a BeagleBoard specific file. * Removed changes to default boot command from adding userbutton command. * Made to handle pre-xM boards. * Flipped polarity of the return value to avoid confusion. Success (0) is when the button is pressed. Failure (1) is when the button is NOT pressed. * Used latest revision getting function. * Used latest macros for board revision. * Added xM-C revision definition (optional, since it was default) * updated default configuration with UserButton functionality * Added a separate bootenv variable to load a user defined .txt file * Added an example, showing how a different environment file can be loaded with the user button pressed
Signed-off-by: Jason Kridner jkridner@beagleboard.org Signed-off-by: Koen Kooi koen@dominion.thruhere.net Signed-off-by: Joel A Fernandes agnel.joel@gmail.com Cc: Greg Turner gregturner@ti.com --- board/ti/beagle/beagle.c | 56 ++++++++++++++++++++++++++++++++++++++++ include/configs/omap3_beagle.h | 7 ++++- 2 files changed, 62 insertions(+), 1 deletions(-)
diff --git a/board/ti/beagle/beagle.c b/board/ti/beagle/beagle.c index a958545..8dc8dfa 100644 --- a/board/ti/beagle/beagle.c +++ b/board/ti/beagle/beagle.c @@ -50,6 +50,7 @@ extern struct ehci_hccr *hccr; extern volatile struct ehci_hcor *hcor; #endif #include "beagle.h" +#include <command.h>
#define pr_debug(fmt, args...) debug(fmt, ##args)
@@ -446,3 +447,58 @@ int ehci_hcd_init(void) }
#endif /* CONFIG_USB_EHCI */ + +/* + * This command returns the status of the user button on beagle xM + * Input - none + * Returns - 1 if button is held down + * 0 if button is not held down + */ +int do_userbutton (cmd_tbl_t *cmdtp, int flag, int argc, char *argv[]) +{ + int button = 0; + int gpio; + + /* + * pass address parameter as argv[0] (aka command name), + * and all remaining args + */ + switch (get_board_revision()) { + case REVISION_AXBX: + case REVISION_CX: + case REVISION_C4: + gpio = 7; + break; + case REVISION_XM_A: + case REVISION_XM_B: + case REVISION_XM_C: + default: + gpio = 4; + break; + } + omap_request_gpio(gpio); + omap_set_gpio_direction(gpio, 1); + printf("The user button is currently "); + if(omap_get_gpio_datain(gpio)) + { + button = 1; + printf("PRESSED.\n"); + } + else + { + button = 0; + printf("NOT pressed.\n"); + } + + omap_free_gpio(gpio); + + return !button; +} + +/* -------------------------------------------------------------------- */ + +U_BOOT_CMD( + userbutton, CONFIG_SYS_MAXARGS, 1, do_userbutton, + "Return the status of the BeagleBoard USER button", + "" +); diff --git a/include/configs/omap3_beagle.h b/include/configs/omap3_beagle.h index c41d48e..a494f99 100644 --- a/include/configs/omap3_beagle.h +++ b/include/configs/omap3_beagle.h @@ -227,7 +227,8 @@ "omapdss.def_disp=${defaultdisplay} " \ "root=${nandroot} " \ "rootfstype=${nandrootfstype}\0" \ - "loadbootenv=fatload mmc ${mmcdev} ${loadaddr} uEnv.txt\0" \ + "bootenv=uEnv.txt\0" \ + "loadbootenv=fatload mmc ${mmcdev} ${loadaddr} ${bootenv}\0" \ "importbootenv=echo Importing environment from mmc ...; " \ "env import -t $loadaddr $filesize\0" \ "loaduimage=fatload mmc ${mmcdev} ${loadaddr} uImage\0" \ @@ -241,8 +242,12 @@
#define CONFIG_BOOTCOMMAND \ "if mmc rescan ${mmcdev}; then " \ + "if userbutton; then " \ + "setenv bootenv user.txt;" \ + "fi;" \ "echo SD/MMC found on device ${mmcdev};" \ "if run loadbootenv; then " \ + "echo Loaded environment from ${bootenv};" \ "run importbootenv;" \ "fi;" \ "if test -n $uenvcmd; then " \
Le 16/08/2011 04:36, Joel A Fernandes a écrit :
From: Jason Kridnerjkridner@beagleboard.org
Based on commit f1099c7c43caf5bac3bf6a65aa266fade4747072
Can(t find this one in my tree.
Author: Greg Turner<gregturner@ti.com> Date: Tue May 25 09:19:06 2010 -0500 New u-boot command for status of USER button on BeagleBoard-xM Modified bootcmd to check the staus at boot time and set filename of the boot script.
Generally speaking, is a 'userbutton' command' really needed when a gpio command exists? The notion of 'user button' isn't really something fundamentally meaningful to a bootloader -- We don't have a concept of 'user', even.
Amicalement,
Dear Joel A Fernandes,
In message 1313462214-3716-2-git-send-email-agnel.joel@gmail.com you wrote:
From: Jason Kridner jkridner@beagleboard.org
Based on commit f1099c7c43caf5bac3bf6a65aa266fade4747072 Author: Greg Turner gregturner@ti.com Date: Tue May 25 09:19:06 2010 -0500
New u-boot command for status of USER button on BeagleBoard-xM Modified bootcmd to check the staus at boot time and set filename of the boot script.
- Moved to a BeagleBoard specific file.
- Removed changes to default boot command from adding userbutton command.
- Made to handle pre-xM boards.
- Flipped polarity of the return value to avoid confusion. Success (0) is when the button is pressed. Failure (1) is when the button is NOT pressed.
- Used latest revision getting function.
- Used latest macros for board revision.
- Added xM-C revision definition (optional, since it was default)
- updated default configuration with UserButton functionality
- Added a separate bootenv variable to load a user defined .txt file
- Added an example, showing how a different environment file can be loaded with the user button pressed
Your patch has a large number of cding style issues; please always run checkpatch before submitting patches.
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
Best regards,
Wolfgang Denk
On Thu, Sep 8, 2011 at 3:05 AM, Wolfgang Denk wd@denx.de wrote:
Dear Joel A Fernandes,
In message 1313462214-3716-2-git-send-email-agnel.joel@gmail.com you wrote:
From: Jason Kridner jkridner@beagleboard.org
Based on commit f1099c7c43caf5bac3bf6a65aa266fade4747072 Author: Greg Turner gregturner@ti.com Date: Tue May 25 09:19:06 2010 -0500
New u-boot command for status of USER button on BeagleBoard-xM
..
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
Thanks, Joel
Hi Joel,
Le 08/09/2011 16:56, Joel A Fernandes a écrit :
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
I gather the "resubmit" means "resubmit the patch set minus this patch and renumbered accordingly".
Thanks, Joel
Amicalement,
On Sep 8, 2011, at 11:03 AM, "Albert ARIBAUD" albert.u.boot@aribaud.net wrote:
Hi Joel,
Le 08/09/2011 16:56, Joel A Fernandes a écrit :
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
I gather the "resubmit" means "resubmit the patch set minus this patch and renumbered accordingly".
One thing that would be necessary is to replace the default bootcmd with one that uses the gpio command and to enable the gpio command.
Thanks, Joel
Amicalement,
Albert.
On Thu, Sep 8, 2011 at 10:10 AM, Kridner, Jason jdk@ti.com wrote:
On Sep 8, 2011, at 11:03 AM, "Albert ARIBAUD" albert.u.boot@aribaud.net wrote:
Hi Joel,
Le 08/09/2011 16:56, Joel A Fernandes a écrit :
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
I gather the "resubmit" means "resubmit the patch set minus this patch and renumbered accordingly".
One thing that would be necessary is to replace the default bootcmd with one that uses the gpio command and to enable the gpio command.
Sounds like a good idea, I will work on this and submit a patch soon.
Thanks, Joel
On Thu, Sep 8, 2011 at 10:10 AM, Kridner, Jason jdk@ti.com wrote:
On Sep 8, 2011, at 11:03 AM, "Albert ARIBAUD" albert.u.boot@aribaud.net wrote:
Hi Joel,
Le 08/09/2011 16:56, Joel A Fernandes a écrit :
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
I gather the "resubmit" means "resubmit the patch set minus this patch and renumbered accordingly".
One thing that would be necessary is to replace the default bootcmd with one that uses the gpio command and to enable the gpio command.
Actually the functions in arch/arm/cpu/armv7/omap-common/gpio.c are not compatible with the ones used by common/cmd_led.c
What would be the right place to implement the required functions for the gpio command?
1. In the (beagle)board file as wrappers to the omap gpio functions 2. In the gpio.c file in omap-common 3. Any other option?
Thanks, Joel
On Thu, Sep 8, 2011 at 9:41 PM, Joel A Fernandes agnel.joel@gmail.com wrote:
On Thu, Sep 8, 2011 at 10:10 AM, Kridner, Jason jdk@ti.com wrote:
On Sep 8, 2011, at 11:03 AM, "Albert ARIBAUD" albert.u.boot@aribaud.net wrote:
Hi Joel,
Le 08/09/2011 16:56, Joel A Fernandes a écrit :
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
I gather the "resubmit" means "resubmit the patch set minus this patch and renumbered accordingly".
One thing that would be necessary is to replace the default bootcmd with one that uses the gpio command and to enable the gpio command.
Actually the functions in arch/arm/cpu/armv7/omap-common/gpio.c are not compatible with the ones used by common/cmd_led.c
The patches from Sanjeev Premi seemed to me to handle the deltas. Are you applying those? They are working for me for use with the led command.
What would be the right place to implement the required functions for the gpio command?
- In the (beagle)board file as wrappers to the omap gpio functions
- In the gpio.c file in omap-common
- Any other option?
Do you mean to replace userbutton? As long as the gpio command returns a status, then you should be able to remove userbutton from the board.c file and add usage of gpio to replace userbutton in the omap3_beagle.h config header file. (Of course, the patches should apply those changes in opposite order to make sure git bisect doesn't show breakage.)
If you are saying that gpio.c is missing functionality, please describe the missing functionality.
Thanks, Joel _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
On Fri, Sep 9, 2011 at 9:19 AM, Jason Kridner jkridner@beagleboard.org wrote:
On Thu, Sep 8, 2011 at 9:41 PM, Joel A Fernandes agnel.joel@gmail.com wrote:
On Thu, Sep 8, 2011 at 10:10 AM, Kridner, Jason jdk@ti.com wrote:
On Sep 8, 2011, at 11:03 AM, "Albert ARIBAUD" albert.u.boot@aribaud.net wrote:
Hi Joel,
Le 08/09/2011 16:56, Joel A Fernandes a écrit :
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
I gather the "resubmit" means "resubmit the patch set minus this patch and renumbered accordingly".
One thing that would be necessary is to replace the default bootcmd with one that uses the gpio command and to enable the gpio command.
Actually the functions in arch/arm/cpu/armv7/omap-common/gpio.c are not compatible with the ones used by common/cmd_led.c
The patches from Sanjeev Premi seemed to me to handle the deltas. Are you applying those? They are working for me for use with the led command.
I missed those patches! I have updated my tree and can see them now, thanks.
Regards, Joel
On Thu, Sep 8, 2011 at 10:10 AM, Kridner, Jason jdk@ti.com wrote:
On Sep 8, 2011, at 11:03 AM, "Albert ARIBAUD" albert.u.boot@aribaud.net wrote:
Hi Joel,
Le 08/09/2011 16:56, Joel A Fernandes a écrit :
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
If this patch is to be dropped, then I'm not sure why the need to resubmit?
I gather the "resubmit" means "resubmit the patch set minus this patch and renumbered accordingly".
One thing that would be necessary is to replace the default bootcmd with one that uses the gpio command and to enable the gpio command.
One of the difficulties with this is the GPIO value for userbutton depends on the board rev which is what your patch does. Hardcoding gpio values for one will break support for other boards
Without adding any additional commands, the only way I see this can be done is: 1. Use the gpio and test commands to get the board rev in omap3_beagle.h 2. Based on this, use the right gpio value to pass to another gpio command to check userbutton state
This is quite messy IMO and is redundant because the revision detection logic is already in the board/ file.
Another way is to add a command to the board/ file like "get_userbutton_gpio" and use the value it returns to pass to the gpio command but this seems to defeat the purpose of the cleanup itself.
If you think one of the above methods is acceptable, please ACK them and I will start developing the patch.
Thanks, Joel
On Thu, Sep 8, 2011 at 3:05 AM, Wolfgang Denk wd@denx.de wrote:
Dear Joel A Fernandes,
In message 1313462214-3716-2-git-send-email-agnel.joel@gmail.com you wrote:
From: Jason Kridner jkridner@beagleboard.org
Based on commit f1099c7c43caf5bac3bf6a65aa266fade4747072 Author: Greg Turner gregturner@ti.com Date: Tue May 25 09:19:06 2010 -0500
New u-boot command for status of USER button on BeagleBoard-xM
Modified bootcmd to check the staus at boot time and set filename of the boot script.
- Moved to a BeagleBoard specific file.
- Removed changes to default boot command from adding userbutton
command.
- Made to handle pre-xM boards.
- Flipped polarity of the return value to avoid confusion. Success (0)
is when the button is pressed. Failure (1) is when the button is NOT pressed.
- Used latest revision getting function.
- Used latest macros for board revision.
- Added xM-C revision definition (optional, since it was default)
- updated default configuration with UserButton functionality
* Added a separate bootenv variable to load a user defined .txt file * Added an example, showing how a different environment file can be loaded with the user button pressed
Your patch has a large number of cding style issues; please always run checkpatch before submitting patches.
Very sorry, I might have missed running it on this patch. I will correct it and resubmit it
Also, I agree with Albert: there should be no need for a separate userbutton command.
Please fix and resubmit.
Sure. Thanks, Joel
From: Jason Kridner jkridner@beagleboard.org
Based on patches from Syed Mohammed Khasim (khasim@ti.com).
Configures the output of the BeagleBoard DVI to be orange. Configures the output of the BeagleBoard S-Video to be a colorbar.
Changed display_init to beagle_display_init as suggested by Igor Grinberg: http://www.mail-archive.com/u-boot@lists.denx.de/msg51446.html
Signed-off-by: Jason Kridner jkridner@beagleboard.org Signed-off-by: Joel A Fernandes agnel.joel@gmail.com --- board/ti/beagle/beagle.c | 24 +++++++++++++ board/ti/beagle/beagle.h | 86 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 110 insertions(+), 0 deletions(-)
diff --git a/board/ti/beagle/beagle.c b/board/ti/beagle/beagle.c index 8dc8dfa..13fe39b 100644 --- a/board/ti/beagle/beagle.c +++ b/board/ti/beagle/beagle.c @@ -165,6 +165,28 @@ unsigned int get_expansion_id(void) }
/* + * Configure DSS to display background color on DVID + * Configure VENC to display color bar on S-Video + */ +void beagle_display_init(void) +{ + omap3_dss_venc_config(&venc_config_std_tv, VENC_HEIGHT, VENC_WIDTH); + switch (get_board_revision()) { + case REVISION_AXBX: + case REVISION_CX: + case REVISION_C4: + omap3_dss_panel_config(&dvid_cfg); + break; + case REVISION_XM_A: + case REVISION_XM_B: + case REVISION_XM_C: + default: + omap3_dss_panel_config(&dvid_cfg_xm); + break; + } +} + +/* * Routine: misc_init_r * Description: Configure board specific parts */ @@ -324,6 +346,8 @@ int misc_init_r(void) GPIO15 | GPIO14 | GPIO13 | GPIO12), &gpio5_base->oe);
dieid_num_r(); + beagle_display_init(); + omap3_dss_enable();
return 0; } diff --git a/board/ti/beagle/beagle.h b/board/ti/beagle/beagle.h index 04247cd..18bfaa8 100644 --- a/board/ti/beagle/beagle.h +++ b/board/ti/beagle/beagle.h @@ -23,6 +23,8 @@ #ifndef _BEAGLE_H_ #define _BEAGLE_H_
+#include <asm/arch/dss.h> + const omap3_sysinfo sysinfo = { DDR_STACKED, "OMAP3 Beagle board", @@ -472,4 +474,88 @@ const omap3_sysinfo sysinfo = { MUX_VAL(CP(MMC2_DAT6), (IDIS | PTU | EN | M4)) /*GPIO_138 BT_EN*/\ MUX_VAL(CP(MMC2_DAT7), (IDIS | PTU | EN | M4)) /*GPIO_139 WLAN_EN*/
+/* + * Display Configuration + */ + +#define DVI_BEAGLE_ORANGE_COL 0x00FF8000 +#define VENC_HEIGHT 0x00ef +#define VENC_WIDTH 0x027f + +/* + * Configure VENC in DSS for Beagle to generate Color Bar + * + * Kindly refer to OMAP TRM for definition of these values. + */ +static const struct venc_regs venc_config_std_tv = { + .status = 0x0000001B, + .f_control = 0x00000040, + .vidout_ctrl = 0x00000000, + .sync_ctrl = 0x00008000, + .llen = 0x00008359, + .flens = 0x0000020C, + .hfltr_ctrl = 0x00000000, + .cc_carr_wss_carr = 0x043F2631, + .c_phase = 0x00000024, + .gain_u = 0x00000130, + .gain_v = 0x00000198, + .gain_y = 0x000001C0, + .black_level = 0x0000006A, + .blank_level = 0x0000005C, + .x_color = 0x00000000, + .m_control = 0x00000001, + .bstamp_wss_data = 0x0000003F, + .s_carr = 0x21F07C1F, + .line21 = 0x00000000, + .ln_sel = 0x00000015, + .l21__wc_ctl = 0x00001400, + .htrigger_vtrigger = 0x00000000, + .savid__eavid = 0x069300F4, + .flen__fal = 0x0016020C, + .lal__phase_reset = 0x00060107, + .hs_int_start_stop_x = 0x008D034E, + .hs_ext_start_stop_x = 0x000F0359, + .vs_int_start_x = 0x01A00000, + .vs_int_stop_x__vs_int_start_y = 0x020501A0, + .vs_int_stop_y__vs_ext_start_x = 0x01AC0024, + .vs_ext_stop_x__vs_ext_start_y = 0x020D01AC, + .vs_ext_stop_y = 0x00000006, + .avid_start_stop_x = 0x03480079, + .avid_start_stop_y = 0x02040024, + .fid_int_start_x__fid_int_start_y = 0x0001008A, + .fid_int_offset_y__fid_ext_start_x = 0x01AC0106, + .fid_ext_start_y__fid_ext_offset_y = 0x01060006, + .tvdetgp_int_start_stop_x = 0x00140001, + .tvdetgp_int_start_stop_y = 0x00010001, + .gen_ctrl = 0x00FF0000, + .output_control = 0x0000000D, + .dac_b__dac_c = 0x00000000 +}; + +/* + * Configure Timings for DVI D + */ +static const struct panel_config dvid_cfg = { + .timing_h = 0x0ff03f31, /* Horizantal timing */ + .timing_v = 0x01400504, /* Vertical timing */ + .pol_freq = 0x00007028, /* Pol Freq */ + .divisor = 0x00010006, /* 72Mhz Pixel Clock */ + .lcd_size = 0x02ff03ff, /* 1024x768 */ + .panel_type = 0x01, /* TFT */ + .data_lines = 0x03, /* 24 Bit RGB */ + .load_mode = 0x02, /* Frame Mode */ + .panel_color = DVI_BEAGLE_ORANGE_COL /* ORANGE */ +}; + +static const struct panel_config dvid_cfg_xm = { + .timing_h = 0x1a4024c9, /* Horizantal timing */ + .timing_v = 0x02c00509, /* Vertical timing */ + .pol_freq = 0x00007028, /* Pol Freq */ + .divisor = 0x00010001, /* 96MHz Pixel Clock */ + .lcd_size = 0x02ff03ff, /* 1024x768 */ + .panel_type = 0x01, /* TFT */ + .data_lines = 0x03, /* 24 Bit RGB */ + .load_mode = 0x02, /* Frame Mode */ + .panel_color = DVI_BEAGLE_ORANGE_COL /* ORANGE */ +}; #endif
participants (5)
-
Albert ARIBAUD -
Jason Kridner -
Joel A Fernandes -
Kridner, Jason -
Wolfgang Denk