Skip to contentSkip to conversation

ARM AT91 new board EB+CPUx9K2

PATCH New Jens Scharsig 4 participants 20 messages Oct 27, 2009 – Oct 30, 2009 Kbuild

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

This patch adds a new ARM AT91RM9200 board, named EB+CPUx9K2.

* support for EB+CPUx9K2 board by BuS Elektronik GmbH & Co. KG
* select via make EB_CPUx9K2_config

---
This patch needs the [PATCH] AT91RM9200 BGA port D defines (http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/50667)

This supports the old CONFIG_DRIVER_ETHER and the new CONFIG_DRIVER_AT91EMAC ([PATCH] AT92RM9200 EMAC driver for NET MULTI API
<http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/69484>)
By default the config use the old CONFIG_DRIVER_ETHER (it can be switch on line 192 in include/configs/EB_CPUx9K2.h)

MAINTAINERS +4 -0
index d70a9d2..6027242 100644
@@ -696,6 +696,10 @@ Steve Sakoman <[email protected]>
 
 	omap3_overo	ARM CORTEX-A8 (OMAP3xx SoC)
 
+Jens Scharsig <[email protected]>
+
+	EB_CPUX9K2	ARM920T (AT91RM9200 SoC)
+
 Robert Schwebel <[email protected]>
 
 	csb226		xscale
MAKEALL +1 -0
index d63c5c2..4cd2a8b 100755
@@ -644,6 +644,7 @@ LIST_at91="			\
 	CPU9260			\
 	CPU9G20			\
 	csb637			\
+	EB_CPUx9K2		\
 	kb9202			\
 	meesc			\
 	mp2usb			\
Makefile +3 -0
index a69e6d5..7fd625f 100644
@@ -2715,6 +2715,9 @@ CPUAT91_config	:	unconfig
 csb637_config	:	unconfig
 	@$(MKCONFIG) $(@:_config=) arm arm920t csb637 NULL at91rm9200
 
+EB_CPUx9K2_config	:	unconfig
+	@$(MKCONFIG) $(@:_config=) arm arm920t EB+CPUx9K2 BuS at91rm9200
+
 kb9202_config	:	unconfig
 	@$(MKCONFIG) $(@:_config=) arm arm920t kb9202 NULL at91rm9200
 
board/BuS/EB+CPUx9K2/Makefile new file +51 -0

51 lines not on this page · raw message

board/BuS/EB+CPUx9K2/config.mk new file +1 -0
new file mode 100644
index 0000000..ff2cfd1
@@ -0,0 +1 @@
+TEXT_BASE = 0x23f00000
board/BuS/EB+CPUx9K2/cpux9k2.c new file +393 -0

393 lines not on this page · raw message

board/BuS/EB+CPUx9K2/u-boot.lds new file +57 -0

57 lines not on this page · raw message

include/configs/EB_CPUx9K2.h new file +457 -0

457 lines not on this page · raw message

#1 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Jens Scharsig,

Quoted 8 lines · In message <[email protected]> you wrote:

This patch adds a new ARM AT91RM9200 board, named EB+CPUx9K2.

* support for EB+CPUx9K2 board by BuS Elektronik GmbH & Co. KG
* select via make EB_CPUx9K2_config

---
This patch needs the [PATCH] AT91RM9200 BGA port D defines (http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/50667)

Please use shorter lines.

...

Quoted 5 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +0 -0
new file mode 100644
index 0000000..6c29d91

...

board/BuS/EB+CPUx9K2/cpux9k2.c line 67

Quoted 7 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +7 -0
+int board_init(void)
+{
+	/* Enable Ctrlc */
+	console_init_f();
+
+	/* Correct IRDA resistor problem / Set PA23_TXD in Output */
+	((AT91PS_PIO) AT91C_BASE_PIOA)->PIO_OER = AT91C_PA23_TXD2;

Please use I/O accessors to access device registers. Please fix globally.

Quoted 2 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +2 -0
+int misc_init_r(void)
+{

...

board/BuS/EB+CPUx9K2/cpux9k2.c line 111

Quoted 5 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +5 -0
+			if (tm) {
+				sprintf(str, "%02x:%02x:%02x:%02x:%02x:%02x",
+					mac[0], mac[1], mac[2],
+					mac[3], mac[4], mac[5]);
+				setenv("ethaddr", str);

Please don't reinvent the wheel. Use eth_setenv_enetaddr().

board/BuS/EB+CPUx9K2/cpux9k2.c line 141

Quoted 9 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +9 -0
+/*
+ * DRAM initialisations
+ */
+
+int dram_init(void)
+{
+	gd->bd->bi_dram[0].start = PHYS_SDRAM;
+	gd->bd->bi_dram[0].size = PHYS_SDRAM_SIZE;
+	return 0;

Please consider using get_ram_size() for auto-sizing the memor and perfoming at least a basic RAM test.

board/BuS/EB+CPUx9K2/cpux9k2.c line 163

Quoted 4 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +4 -0
+unsigned int lxt972_IsPhyConnected(AT91PS_EMAC p_mac);
+UCHAR lxt972_GetLinkSpeed(AT91PS_EMAC p_mac);
+UCHAR lxt972_InitPhy(AT91PS_EMAC p_mac);
+UCHAR lxt972_AutoNegotiate(AT91PS_EMAC p_mac, int *status);

Please don't use non-stanrard data types like UCHAR.

Quoted 5 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +5 -0
+void cpux9k2_nand_hw_init(void)
+{
+	/* Setup Smart Media, fitst enable the address range of CS3 */
+	*AT91C_EBI_CSA |= AT91C_EBI_CS3A_SMC_SmartMedia;
+	/* set the bus interface characteristics based on

board/BuS/EB+CPUx9K2/cpux9k2.c line 203

Quoted 10 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +10 -0
+	   tDS Data Set up Time 30 - ns
+	   tDH Data Hold Time 20 - ns
+	   tALS ALE Set up Time 20 - ns
+	   16ns at 60 MHz ~= 3  */
+	/*memory mapping structures */
+#define SM_ID_RWH	(5 << 28)
+#define SM_RWH		(1 << 28)
+#define SM_RWS		(0 << 24)
+#define SM_TDF		(1 << 8)
+#define SM_NWS		(3)

Please don't add #defines right in the middle of the code. Move to header file, or at least to head of file.

And see comment above - use I/O accessors here and everywhere.

board/BuS/EB+CPUx9K2/cpux9k2.c line 269

Quoted 6 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +6 -0
+#if defined(CONFIG_VIDEO)
+/*
+ ****h* EB+CPUx9K2/drv_video_init
+ * FUNCTION
+ ***
+ */

Incorrect multiline comment style.

Quoted 12 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +12 -0
+int drv_video_init(void)
+{
+	char *s;
+	unsigned long splash;
+
+	printf("Init Video as ");
+	s = getenv("displaywidth");
+	if (s != NULL)
+		display_width = simple_strtoul(s, NULL, 10);
+	else
+		display_width = 256;
+	s = getenv("displayheight");

board/BuS/EB+CPUx9K2/cpux9k2.c line 292

Quoted 10 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +10 -0
+	if (s != NULL)
+		display_height = simple_strtoul(s, NULL, 10);
+	else
+		display_height = 256;
+	printf("%ld x %ld pixel matrix\n", display_width, display_height);
+
+	#define SM2_RWH		(0x7 << 28)
+	#define SM2_RWS		(0x7 << 24)
+	#define SM2_TDF		(15 << 8)
+	#define SM2_NWS		(0x7F)

Please don't add #defines right in the middle of the code. Move to header file, or at least to head of file.

board/BuS/EB+CPUx9K2/u-boot.lds

Quoted 5 lines
board/BuS/EB+CPUx9K2/u-boot.lds new file +0 -0
new file mode 100644
index 0000000..f4fbf96

Do you really need a private linker script? Doesn't look so...

Quoted 5 lines
include/configs/EB_CPUx9K2.h new file +0 -0
new file mode 100644
index 0000000..9ff743d

...

include/configs/EB_CPUx9K2.h line 36

Quoted 6 lines
include/configs/EB_CPUx9K2.h new file +6 -0
+
+/*--------------------------------------------------------------------------*/
+
+#undef DEBUG
+
+/*--------------------------------------------------------------------------*/

Get rid of all this stuff. Don't undef what is not defined in the beginning, and don't meddle with command line options.

include/configs/EB_CPUx9K2.h line 141

Quoted 11 lines
include/configs/EB_CPUx9K2.h new file +11 -0
+#include <config_cmd_default.h>
+
+#define CONFIG_CMD_DHCP
+#define CONFIG_CMD_PING
+#define CONFIG_CMD_NAND
+#define CONFIG_CMD_BMP
+#define	CONFIG_CMD_I2C
+#define	CONFIG_I2C_CMD_TREE
+#define	CONFIG_I2C_CMD_NO_FLAT
+#define	CONFIG_CMD_DATE
+#define	CONFIG_CMD_JFFS2

You may want to keep such lists sorted. And please use a consistent style - either always follow #define with a TAB, or (probably better) always with a space.

include/configs/EB_CPUx9K2.h line 186

Quoted 8 lines
include/configs/EB_CPUx9K2.h new file +8 -0
+#define CONFIG_BAUDRATE 115200
+#define CONFIG_AT91RM9200_USART
+#define CONFIG_DBGU			/* define DBGU as console */
+#undef CONFIG_USART0
+#undef CONFIG_USART1
+
+#undef	CONFIG_HWFLOW
+#undef	CONFIG_MODEM_SUPPORT

Don;t undefine what is not defined.

include/configs/EB_CPUx9K2.h line 199

Quoted 12 lines
include/configs/EB_CPUx9K2.h new file +12 -0
+/*
+ * network
+ */
+
+#if 0
+#define	CONFIG_NET_MULTI		1
+#define	CONFIG_DRIVER_AT91EMAC		1
+#define	CONFIG_DRIVER_AT91EMAC_QUIET	1
+#define	CONFIG_SYS_RX_ETH_BUFFER	8
+#undef	CONFIG_RMII
+#define	CONFIG_MII			1
+#define	CONFIG_CMD_MII			1

Don't add dead code.

include/configs/EB_CPUx9K2.h line 298

Quoted 2 lines
include/configs/EB_CPUx9K2.h new file +2 -0
+#define CONFIG_SYS_FLASH_ERASE_TOUT	(6*CONFIG_SYS_HZ)
+#define CONFIG_SYS_FLASH_WRITE_TOUT	(2*CONFIG_SYS_HZ)

This looks wrong to me. A timeout is a time, but CONFIG_SYS_HZ is a frequency, i. e. the inverse of a time. These don't mix.

Best regards,

Wolfgang Denk

Signature · 6 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
This all sounds complicated, but it mostly does excatly what you  ex-
pect. It's just difficult for us to explain what you expect...
                       - L. Wall & R. L. Schwartz, _Programming Perl_
#2 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Quoted 11 lines · Wolfgang Denk

Dear Jens Scharsig,

board/BuS/EB+CPUx9K2/cpux9k2.c line 67

Quoted 7 lines
board/BuS/EB+CPUx9K2/cpux9k2.c new file +7 -0
+int board_init(void)
+{
+	/* Enable Ctrlc */
+	console_init_f();
+
+	/* Correct IRDA resistor problem / Set PA23_TXD in Output */
+	((AT91PS_PIO) AT91C_BASE_PIOA)->PIO_OER = AT91C_PA23_TXD2;

Please use I/O accessors to access device registers. Please fix globally.

Do you mean change to

writel(AT91C_PA23_TXD2, ((AT91PS_PIO) AT91C_BASE_PIOA)->PIO_OER);

like notation

Best regards Jens

#3 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Jens Scharsig,

Quoted 8 lines · In message <[email protected]> you wrote:
Quoted 3 lines

Please use I/O accessors to access device registers. Please fix globally.

Do you mean change to

writel(AT91C_PA23_TXD2, ((AT91PS_PIO) AT91C_BASE_PIOA)->PIO_OER);

like notation

Close, but please also get rid of this "((AT91PS_PIO) AT91C_BASE_PIOA)" thing, especialy the cast is nasty. Declare a real pointer variable to the respective data structure, and use that.

Best regards,

Wolfgang Denk

--
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de If the odds are a million to one against something occuring, chances are 50-50 it will.

#4 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Wolfgang Denk,

Quoted 8 lines
Quoted 3 lines

writel(AT91C_PA23_TXD2, ((AT91PS_PIO) AT91C_BASE_PIOA)->PIO_OER);

like notation

Close, but please also get rid of this "((AT91PS_PIO) AT91C_BASE_PIOA)" thing, especialy the cast is nasty. Declare a real pointer variable to the respective data structure, and use that.

I, think

writel(AT91C_PA23_TXD2, AT91C_PIOA_OER);

is the most correct way.
But, there are need some additional def's in AT91RM9200.h. Only the port A registers is defined as pointer variable. If you think also this is the right way, I will send a separate patch for AT91RM9200.h (PORTB .. PORTD, SMC registers)

Best regards,

Jens Scharsig

#5 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Jens Scharsig,

Quoted 11 lines · In message <[email protected]> you wrote:
Quoted 5 lines

Close, but please also get rid of this "((AT91PS_PIO) AT91C_BASE_PIOA)" thing, especialy the cast is nasty. Declare a real pointer variable to the respective data structure, and use that.

I, think

writel(AT91C_PA23_TXD2, AT91C_PIOA_OER);

is the most correct way.

"most correct way" are big words. No, this is not "correct" at all.

The whole set of address / offset definitions in include/asm-arm/arch-at91rm9200/AT91RM9200.h should be turned into a C struct.

Best regards,

Wolfgang Denk

Signature · 4 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
When all else fails, read the instructions.
#6 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Wolfgang Denk

Quoted 9 lines
Quoted 3 lines

writel(AT91C_PA23_TXD2, AT91C_PIOA_OER);

is the most correct way.

"most correct way" are big words. No, this is not "correct" at all.

The whole set of address / offset definitions in include/asm-arm/arch-at91rm9200/AT91RM9200.h should be turned into a C struct.

I've got you misunderstood, what real pointer variable means.

If I understand you correctly now?
The train goes in the opposite direction.

in the AT91RM9200.h

a port is defined as

typedef struct _AT91S_PIO

{
	...
	AT91_REG	 PIO_OER;	/* Output Enable Register */
	...
} AT91S_PIO, *AT91PS_PIO;

and

#define AT91C_BASE_PIOC ((AT91PS_PIO) 0xFFFFF800)

So the access should be

	AT91PS_PIO pioa = AT91C_BASE_PIOA;
	...
	writel(AT91C_PA23_TXD2, &pioa->PIO_OER);

or

	writel(AT91C_PA23_TXD2, &AT91C_BASE_PIOA->PIO_OER);

instead of

	writel(AT91C_PA23_TXD2, ((AT91PS_PIO) AT91C_BASE_PIOA)->PIO_OER)

	or 
	
	writel(AT91C_PA23_TXD2, AT91C_PIOA_OER)

The result, if all the board are adjusted, it can be removed individual register definitions like
#define AT91C_PIOD_OWER ((AT91_REG *) 0xFFFFFAA0)

I'm right?

Best regards,

Jens Scharsig

#7 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Jens Scharsig,

Quoted 6 lines · In message <[email protected]> you wrote:

If I understand you correctly now?
The train goes in the opposite direction.

in the AT91RM9200.h

Quoted 8 lines

a port is defined as

typedef struct _AT91S_PIO

{
	...
	AT91_REG	 PIO_OER;	/* Output Enable Register */
	...
} AT91S_PIO, *AT91PS_PIO;

This is close. Of course we should drop the AT91_REG and use standard types instead, and "PIO_OER" is not a logal variable name either because it's all-capitals. So this entry should rather look like this:

	...
	u32 pio_oer;
	...
Quoted 3 lines

and

#define AT91C_BASE_PIOC ((AT91PS_PIO) 0xFFFFF800)

This is definitely deprecated.

Quoted 5 lines

So the access should be

	AT91PS_PIO pioa = AT91C_BASE_PIOA;
	...
	writel(AT91C_PA23_TXD2, &pioa->PIO_OER);

Yes, except for the incorrect variable name.

Quoted 3 lines

or

	writel(AT91C_PA23_TXD2, &AT91C_BASE_PIOA->PIO_OER);

No.

Quoted 1 line

I'm right?

Mostly :-)

Thanks!

Best regards,

Wolfgang Denk

Signature · 4 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
 The software required `Windows 95 or better', so I installed Linux.
#8 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Wolfgang Denk

Quoted 16 lines

Dear Jens Scharsig,

This is close. Of course we should drop the AT91_REG and use standard types instead, and "PIO_OER" is not a logal variable name either because it's all-capitals. So this entry should rather look like this:

	...
	u32 pio_oer;
	...
Quoted 3 lines

and

#define AT91C_BASE_PIOC ((AT91PS_PIO) 0xFFFFF800)

This is definitely deprecated.

Quoted 8 lines
Quoted 5 lines

So the access should be

	AT91PS_PIO pioa = AT91C_BASE_PIOA;
	...
	writel(AT91C_PA23_TXD2, &pioa->PIO_OER);

Yes, except for the incorrect variable name.

By the way, the AT91RM9200.h. has hundreds of style problems. This requires a complete revision of the AT91RM9200.h. I can try this, but will take a while and I can't test other RM9200 boards.

Best regards,

Jens Scharsig

#9 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Quoted 21 lines · Jens Scharsig

Dear Wolfgang Denk

Quoted 20 lines

Dear Jens Scharsig,

This is close. Of course we should drop the AT91_REG and use standard types instead, and "PIO_OER" is not a logal variable name either because it's all-capitals. So this entry should rather look like this:

	...
	u32 pio_oer;
	...
Quoted 3 lines

and

#define AT91C_BASE_PIOC ((AT91PS_PIO) 0xFFFFF800)

This is definitely deprecated.

Quoted 5 lines

So the access should be

	AT91PS_PIO pioa = AT91C_BASE_PIOA;
	...
	writel(AT91C_PA23_TXD2, &pioa->PIO_OER);
Quoted 8 lines
Quoted 2 lines

Yes, except for the incorrect variable name.

By the way, the AT91RM9200.h. has hundreds of style problems. This requires a complete revision of the AT91RM9200.h. I can try this, but will take a while and I can't test other RM9200 boards.

Please limit your changes to what you can test.
Tom

Quoted 8 lines

Best regards,

Jens Scharsig
_______________________________________________
U-Boot mailing list
U-Boot at lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot

#10 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Tom,

Quoted 8 lines · In message <[email protected]> you wrote:
Quoted 5 lines

By the way, the AT91RM9200.h. has hundreds of style problems. This requires a complete revision of the AT91RM9200.h. I can try this, but will take a while and I can't test other RM9200 boards.

Please limit your changes to what you can test.
Tom

Pure coding style changes are easy to verify - they should lead to identical (except for the time stamps) images.

Also, board maintainers can and should be challenged to test such patches.

Please don't stop improving the code just because others need to help with the testing.

Best regards,

Wolfgang Denk

Signature · 4 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
God made the integers; all else is the work of Man.       - Kronecker
#11 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Quoted 10 lines · Wolfgang Denk

Dear Tom,

Quoted 6 lines · In message <[email protected]> you wrote:
Quoted 4 lines

By the way, the AT91RM9200.h. has hundreds of style problems. This requires a complete revision of the AT91RM9200.h. I can try this, but will take a while and I can't test other RM9200 boards.

Please limit your changes to what you can test.
Tom

Quoted 8 lines

Pure coding style changes are easy to verify - they should lead to identical (except for the time stamps) images.

Also, board maintainers can and should be challenged to test such patches.

Please don't stop improving the code just because others need to help with the testing.

I was thinking that this would be converting from #define offsets to structures. It can be easy to get these wrong. Without testing as-you-go these types of bugs can be difficult to resolve.

Tom

Quoted 5 lines

Best regards,

Wolfgang Denk

#12 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

In reply to Jens Scharsig

Dear Jens Scharsig,

Quoted 5 lines · In message <[email protected]> you wrote:

By the way, the AT91RM9200.h. has hundreds of style problems. This requires a complete revision of the AT91RM9200.h. I can try this, but will take a while and I can't test other RM9200 boards.

I am aware of these issues. AT91 is indeed in a poor shape.

Also, it's still lacking a custodian.

Nobody volunteered for this job so far. Seems AT91 is not much used any more these days ?

Best regards,

Wolfgang Denk

Signature · 6 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Every program has at least one bug and can be shortened by  at  least
one instruction - from which, by induction, one can deduce that every
program can be reduced to one instruction which doesn't work.
#13 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Wolfgang Denk

Quoted 8 lines

Dear Jens Scharsig,

Quoted 4 lines · In message <[email protected]> you wrote:

By the way, the AT91RM9200.h. has hundreds of style problems. This requires a complete revision of the AT91RM9200.h. I can try this, but will take a while and I can't test other RM9200 boards.

Quoted 6 lines

I am aware of these issues. AT91 is indeed in a poor shape.

Also, it's still lacking a custodian.

Nobody volunteered for this job so far. Seems AT91 is not much used any more these days ?

That is exactly the problem the AT91RM9200 is around 10 years old. It isn't 100% compatible with the newer AT91SAM derivatives. I think the RM9200 is only used in redesigns of old projects. We ourselves use AT91SAM9xxx for new projects. But we have several customer board that use the same hardware like our reference board EB+CPUx9K2.
I think there is an second reason, why nobody volunteered for this. Some people want to benefit from the open-source projects, but nobody will help.

Best regards
Jens Scharsig

#14 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

In reply to Wolfgang Denk
Quoted 5 lines · Wolfgang Denk

include/configs/EB_CPUx9K2.h line 298

Quoted 2 lines
include/configs/EB_CPUx9K2.h new file +2 -0
+#define CONFIG_SYS_FLASH_ERASE_TOUT	(6*CONFIG_SYS_HZ)
+#define CONFIG_SYS_FLASH_WRITE_TOUT	(2*CONFIG_SYS_HZ)

This looks wrong to me. A timeout is a time, but CONFIG_SYS_HZ is a frequency, i. e. the inverse of a time. These don't mix.

We've been over this:
http://lists.denx.de/pipermail/u-boot/2009-September/059961.html
:-P

C values don't carry units, so you can't assume that multiplying a frequency by 6 means you want another frequency 6 times as fast.

Whether the timeout is supposed to be expressed in ticks or in some other unit such as milliseconds is another matter.

-Scott

#15 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

In reply to Jens Scharsig

Dear Jens Scharsig,

Quoted 1 line · In message <[email protected]> you wrote:

This patch adds a new ARM AT91RM9200 board, named EB+CPUx9K2.

And by the way: please chose a board name that can be used consistently:

Makefile CPUAT91_config : unconfig · line 2719

Quoted 2 lines
Makefile +2 -0
+EB_CPUx9K2_config	:	unconfig
+	@$(MKCONFIG) $(@:_config=) arm arm920t EB+CPUx9K2 BuS at91rm9200

If the target name is EB_CPUx9K2 (why are you shouting here?) then the board directory and config file names shall be based on EB_CPUx9K2, too.

Thanks.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Perl itself is  usually  pretty  good  about  telling  you  what  you
shouldn't do. :-)     - Larry Wall in <[email protected]>
#16 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Quoted 8 lines · Wolfgang Denk

And by the way: please chose a board name that can be used consistently:

Makefile CPUAT91_config : unconfig · line 2719

Quoted 2 lines
Makefile +2 -0
+EB_CPUx9K2_config	:	unconfig
+	@$(MKCONFIG) $(@:_config=) arm arm920t EB+CPUx9K2 BuS at91rm9200

If the target name is EB_CPUx9K2 (why are you shouting here?) then the board directory and config file names shall be based on EB_CPUx9K2, too.

It's a incomplete rename by my side. sorry.

Wath name/syntax do you prefer? eb_cpux9k2... or eb+cpux9k2...

Best regards,

Jens Scharsig

#17 lore raw

PATCH 1/1 ARM AT91 new board EB+CPUx9K2

Dear Jens Scharsig,

Quoted 10 lines · In message <[email protected]> you wrote:
Quoted 8 lines · Wolfgang Denk

And by the way: please chose a board name that can be used consistently:

Makefile CPUAT91_config : unconfig · line 2719

Quoted 2 lines
Makefile +2 -0
+EB_CPUx9K2_config	:	unconfig
+	@$(MKCONFIG) $(@:_config=) arm arm920t EB+CPUx9K2 BuS at91rm9200

If the target name is EB_CPUx9K2 (why are you shouting here?) then the board directory and config file names shall be based on EB_CPUx9K2, too.

Quoted 3 lines

It's a incomplete rename by my side. sorry.

Wath name/syntax do you prefer? eb_cpux9k2... or eb+cpux9k2...

'+' is not a valid character in any identifier, so please avoid this.

Best regards,

Wolfgang Denk

Signature · 4 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
It's not an optical illusion, it just looks like one.   -- Phil White
#18 lore raw

PATCH v2 1/1 ARM AT91 new board EB+CPUx9K2

In reply to Jens Scharsig

This patch adds a new ARM AT91RM9200 board, named EB+CPUx9K2.

* support for EB+CPUx9K2 board by BuS Elektronik GmbH & Co. KG
* select via make eb_cpux9k2_config

---
This patch needs the [PATCH] AT91RM9200: real pointer variable for PORT A to D configuration
By default the config use the old CONFIG_DRIVER_ETHER.

The patch supports also the new CONFIG_DRIVER_AT91EMAC ([PATCH] AT92RM9200 EMAC driver for NET MULTI API). Remove comments in #define CONFIG_NET_MULTI line to use it.

Diffrence to V1:

* use the consistently silently board name eb_cpux9k2
* use I/O accessors
* dram_init using get_ram_size()
* remove non-standard data types
* no #defines in the middle of the code
* use global linker script
* remove undef's
* timeouts do not based on CONFIG_SYS_HZ
* fix several code style problems
* several fixes

MAINTAINERS +4 -0
index d70a9d2..65b06e2 100644
@@ -696,6 +696,10 @@ Steve Sakoman <[email protected]>
 
 	omap3_overo	ARM CORTEX-A8 (OMAP3xx SoC)
 
+Jens Scharsig	<[email protected]>
+
+	eb_cpux9k2	ARM920T (AR91RM9200 SoC)
+
 Robert Schwebel <[email protected]>
 
 	csb226		xscale
MAKEALL +1 -0
index d63c5c2..d7ef41c 100755
@@ -644,6 +644,7 @@ LIST_at91="			\
 	CPU9260			\
 	CPU9G20			\
 	csb637			\
+	eb_cpux9k2		\
 	kb9202			\
 	meesc			\
 	mp2usb			\
Makefile +3 -0
index a69e6d5..d425b1a 100644
@@ -2715,6 +2715,9 @@ CPUAT91_config	:	unconfig
 csb637_config	:	unconfig
 	@$(MKCONFIG) $(@:_config=) arm arm920t csb637 NULL at91rm9200
 
+eb_cpux9k2_config	:	unconfig
+	@$(MKCONFIG) $(@:_config=) arm arm920t eb_cpux9k2 BuS at91rm9200
+
 kb9202_config	:	unconfig
 	@$(MKCONFIG) $(@:_config=) arm arm920t kb9202 NULL at91rm9200
 
board/BuS/eb_cpux9k2/Makefile new file +51 -0

51 lines not on this page · raw message

board/BuS/eb_cpux9k2/config.mk new file +1 -0
new file mode 100644
index 0000000..ff2cfd1
@@ -0,0 +1 @@
+TEXT_BASE = 0x23f00000
board/BuS/eb_cpux9k2/cpux9k2.c new file +376 -0

376 lines not on this page · raw message

include/configs/eb_cpux9k2.h new file +426 -0

426 lines not on this page · raw message

#19 lore raw

PATCH v2 1/1 ARM AT91 new board EB+CPUx9K2

Dear Jens Scharsig,

Quoted 9 lines · In message <[email protected]> you wrote:

This patch adds a new ARM AT91RM9200 board, named EB+CPUx9K2.

* support for EB+CPUx9K2 board by BuS Elektronik GmbH & Co. KG
* select via make eb_cpux9k2_config

---
This patch needs the [PATCH] AT91RM9200: real pointer variable for PORT A to D configuration

...which has been NAKed. Please use C structs!!!

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"Anyone attempting to generate random numbers by deterministic  means
is, of course, living in a state of sin."          - John Von Neumann