Skip to contentSkip to conversation

nand merge problem

U-Boot xiangfu_gmail 2 participants 11 messages May 26, 2009 – Jun 10, 2009 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

Hi
I have try to merge Ingenic U-boot (1.1.6) with upstream. but I have some problem
1. there is no
--
NAND_CTL_SETNCE:
NAND_CTL_CLRNCE:
NAND_CTL_SETCLE:
NAND_CTL_CLRCLE:
NAND_CTL_SETALE:
NAND_CTL_CLRALE:
instead of :
--
NAND_NCE:
NAND_CLE:
NAND_ALE:
NAND_CTRL_CLE:
NAND_CTRL_ALE:
NAND_CTRL_CHANGE:

I use the NAND_CTL_SETNCE... in the jz_hwcontrol function. i use "nand->hwcontrol = jz_hwcontrol" in the u-boot 1.1.6 but in the U-boot V2009-06 the nand_chip structure does not have "hwcontrol" function.

what am I gonna do?

thanks for help.

--

Best Regards
Xiangfu Liu

jabber : xiangfu.z at gmail.com
skype : xiangfu.z

Please avoid sending me Word or PowerPoint attachments. See http://www.gnu.org/philosophy/no-word-attachments.html

#1 lore raw
Quoted 3 lines · xiangfu_gmail

Hi
I have try to merge Ingenic U-boot (1.1.6) with upstream. but I have some problem

Hi I rewrite the jz_hwcontrol to :
--
static void jz_hwcontrol(......)
{ ......

	if (ctrl & NAND_CTRL_CHANGE) {
		if (ctrl & NAND_NCE)
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1;
		else
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;

		if (ctrl & NAND_CLE)
		......
	}
}

and
" nand->cmd_ctrl = jz_hwcontrol; " in board_nand_init

but it's still not work:
the error message:
--
NAND: nand_get_flash_type: second ID read did not match ff,ff against 00,00
No NAND device found!!!

0 MiB
--

what can cause this problem?

thanks

#2 lore raw
Quoted 28 lines · xiangfu_gmail
Quoted 3 lines · xiangfu_gmail

Hi
I have try to merge Ingenic U-boot (1.1.6) with upstream. but I have some problem

Hi I rewrite the jz_hwcontrol to :
--
static void jz_hwcontrol(......)
{ ......

	if (ctrl & NAND_CTRL_CHANGE) {
		if (ctrl & NAND_NCE)
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1;
		else
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;

		if (ctrl & NAND_CLE)
		......
	}
}

and
" nand->cmd_ctrl = jz_hwcontrol; " in board_nand_init

but it's still not work:
the error message:
--
NAND: nand_get_flash_type: second ID read did not match ff,ff against 00,00
No NAND device found!!!

Quoted 4 lines

0 MiB
--

what can cause this problem?

It would be helpful if you were to post the entire hwcontrol, so we can see what's hiding behind those "......"s.

Posting the original legacy NAND defines for this board would be helpful as well.

-Scott

#3 lore raw

Hi Scott

thanks for the reply

Quoted 20 lines · Scott Wood
Quoted 19 lines · xiangfu_gmail
Quoted 3 lines · xiangfu_gmail

Hi
I have try to merge Ingenic U-boot (1.1.6) with upstream. but I have some problem

Hi I rewrite the jz_hwcontrol to :
--
static void jz_hwcontrol(......)
{ ......
}
and
" nand->cmd_ctrl = jz_hwcontrol; " in board_nand_init

but it's still not work:
the error message:
--
NAND: nand_get_flash_type: second ID read did not match ff,ff against 00,00
No NAND device found!!!

Quoted 7 lines
Quoted 4 lines

0 MiB
--

what can cause this problem?

It would be helpful if you were to post the entire hwcontrol, so we can see what's hiding behind those "......"s.

here is the entire hwontrol
static void jz_hwcontrol(struct mtd_info *mtd, int cmd, unsigned int ctrl)

{
	struct nand_chip *this = (struct nand_chip *)(mtd->priv);

	if (ctrl & NAND_CTRL_CHANGE) {
		if (ctrl & NAND_NCE)
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1;
		else
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;

		if (ctrl & NAND_CLE)
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) | 0x00008000);
		else			
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) & ~0x00008000);

		if (ctrl & NAND_ALE)
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) | 0x00010000);
		else
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) & ~0x00010000);

	}
}
Quoted 3 lines

Posting the original legacy NAND defines for this board would be helpful as well.

I don't know which code should I post. so I post the git URL:
http://github.com/xiangfu/pi-u-boot/tree/xiangfu

Quoted 2 lines

-Scott

thanks

Signature · 5 lines
Best Regards
Xiangfu Liu

jabber : xiangfu.z at gmail.com
skype  : xiangfu.z
#4 lore raw
Quoted 18 lines · xiangfu_gmail

static void jz_hwcontrol(struct mtd_info *mtd, int cmd, unsigned int ctrl)

{
	struct nand_chip *this = (struct nand_chip *)(mtd->priv);

	if (ctrl & NAND_CTRL_CHANGE) {
		if (ctrl & NAND_NCE)
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1;
		else
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;

		if (ctrl & NAND_CLE)
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) | 0x00008000);
		else			
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) & ~0x00008000);

		if (ctrl & NAND_ALE)
Quoted 8 lines
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) | 0x00010000);
		else
			this->IO_ADDR_W = (void __iomem *)
		((unsigned long)(this->IO_ADDR_W) & ~0x00010000);

	}
}

Nowhere in this function do you issue the command... See cpu/ppc4xx/ndfc.c for a simple hwcontrol function.

-Scott

author
#6 lore raw
In reply to a message not in the archive
Quoted 6 lines · zyliu at ingenic.cn

Hi, xiangfu

Maybe the drivers of mtd in linux-2.6.24.3 can be as reference.

Best Regards

thanks, now it's output :
"nand_get_flash_type: second ID read did not match 43,20 against 84,84"

I will let you know the progress.

now the hwcontrol is
--
static void jz_hwcontrol(struct mtd_info *mtd, int cmd, unsigned int ctrl)

{
	struct nand_chip *this = (struct nand_chip *)(mtd->priv);
        unsigned int nandaddr = (unsigned int)this->IO_ADDR_W;
	if (ctrl & NAND_CTRL_CHANGE) {
		if (ctrl & NAND_CLE)
			nandaddr = nandaddr | 0x00008000;
                else
                        nandaddr = nandaddr & ~0x00008000;
			
		if (ctrl & NAND_ALE)
			this->IO_ADDR_W = (void __iomem *)((unsigned long)(this->IO_ADDR_W) |
0x00010000);
		else
			this->IO_ADDR_W = (void __iomem *)((unsigned long)(this->IO_ADDR_W) &
~0x00010000);
		if (ctrl & NAND_NCE) {
			this->IO_ADDR_W = this->IO_ADDR_R = (void __iomem *)NAND_DATA_PORT1;
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1;
		} else {
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;
                        REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE2;
                        REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE3;
                        REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE4;
		}
	}

	this->IO_ADDR_W = (void __iomem *)nandaddr;
        if (cmd != NAND_CMD_NONE)
                writeb(cmd, this->IO_ADDR_W);
}
Signature · 5 lines
Best Regards
Xiangfu Liu

jabber : xiangfu.z at gmail.com
skype  : xiangfu.z
#7 lore raw
Quoted 23 lines · xiangfu

static void jz_hwcontrol(struct mtd_info *mtd, int cmd, unsigned int ctrl)

{
	struct nand_chip *this = (struct nand_chip *)(mtd->priv);
        unsigned int nandaddr = (unsigned int)this->IO_ADDR_W;
	if (ctrl & NAND_CTRL_CHANGE) {
		if (ctrl & NAND_CLE)
			nandaddr = nandaddr | 0x00008000;
                else
                        nandaddr = nandaddr & ~0x00008000;
			
		if (ctrl & NAND_ALE)
			this->IO_ADDR_W = (void __iomem *)((unsigned long)(this->IO_ADDR_W) |
0x00010000);
		else
			this->IO_ADDR_W = (void __iomem *)((unsigned long)(this->IO_ADDR_W) &
~0x00010000);
		if (ctrl & NAND_NCE) {
			this->IO_ADDR_W = this->IO_ADDR_R = (void __iomem *)NAND_DATA_PORT1;
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1;
		} else {
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;
                        REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE2;
                        REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE3;
Quoted 8 lines
                        REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE4;
		}
	}

	this->IO_ADDR_W = (void __iomem *)nandaddr;
        if (cmd != NAND_CMD_NONE)
                writeb(cmd, this->IO_ADDR_W);
}

Try something like this instead:

static void jz_hwcontrol(struct mtd_info *mtd, int cmd, unsigned int ctrl)

{
	struct nand_chip *this = mtd->priv;
	unsigned long nandaddr = (unsigned long)this->IO_ADDR_W;

	if (ctrl & NAND_CTRL_CHANGE) {
		/* Change this to use I/O accessors. */
		if (ctrl & NAND_NCE) {
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1; 
		} else {
			/*
			 * Why set only one bit when NCE is high, but clear
			 * four when low?  Why clear separate bits in the same
			 * register one at a time?
			 */
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE2;
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE3;
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE4;
		}
	}

	if (cmd == NAND_CMD_NONE)
		return;

	if (ctrl & NAND_CLE)
		nandaddr |= 0x00008000;
	else /* must be ALE */
		nandaddr |= 0x00010000;

	writeb(cmd, (uint8_t *)nandaddr);
}

-Scott

author
#8 lore raw

Thanks Scott,
it's work. :-)

Quoted 17 lines · Scott Wood

Try something like this instead:

static void jz_hwcontrol(struct mtd_info *mtd, int cmd, unsigned int ctrl)

{
	struct nand_chip *this = mtd->priv;
	unsigned long nandaddr = (unsigned long)this->IO_ADDR_W;

	if (ctrl & NAND_CTRL_CHANGE) {
		/* Change this to use I/O accessors. */
		if (ctrl & NAND_NCE) {
			REG_EMC_NFCSR |= EMC_NFCSR_NFCE1; 
		} else {
			/*
			 * Why set only one bit when NCE is high, but clear
			 * four when low?  Why clear separate bits in the same
			 * register one at a time?
			 */

my mistake. I copy those code form the device's kernel source code.

Quoted 19 lines
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE1;
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE2;
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE3;
			REG_EMC_NFCSR &= ~EMC_NFCSR_NFCE4;
		}
	}

	if (cmd == NAND_CMD_NONE)
		return;

	if (ctrl & NAND_CLE)
		nandaddr |= 0x00008000;
	else /* must be ALE */
		nandaddr |= 0x00010000;

	writeb(cmd, (uint8_t *)nandaddr);
}

-Scott

Signature · 5 lines
Best Regards
Xiangfu Liu

jabber : xiangfu.z at gmail.com
skype  : xiangfu.z
author
#9 lore raw

[U-Boot] console_init_r problem with ingenic jz4740 cpu

In reply to Scott Wood

Hi

the u-boot stop at console_init_r function.

here is part of board_init_r function code.

/** leave this here (after malloc(), environment and PCI are working) **/

	/* Initialize devices */
	devices_init ();

	jumptable_init ();
	debug("DEBUG: jumptable over\n"); [1]

	/* Initialize the console (after the relocation and devices init) */
	console_init_r ();
	debug("DEBUG: console_init_r over\n");[2]

/** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** **/

I add two lines for debug.

the [1] is output form serial console. but [2] never show up.
I also add some "puts" in the console_init_r function. it's not show in the serial console

give me some advice

thanks

Best Regards
Xiangfu

author
#10 lore raw

[U-Boot] lcd logo color not correct

In reply to Scott Wood

Hi
1. I have try to port u-boot to jz4740 cpu device. now the lcd can show the LOGO.
but the color is not correct.

2.

	gd->flags |= GD_FLG_DEVINIT;	/* device initialization completed */

when I comment this line. the device can boot to main_loop. when don't comment this line as normal. the system crash. give me some help about this problem

thanks.

attach is the jz_lcd.c file.

Signature · 12 lines
Best Regards
Xiangfu Liu

jabber : xiangfu.z at gmail.com
skype  : xiangfu.z
-------------- next part --------------
A non-text attachment was scrubbed...
Name: jz_lcd.c
Type: text/x-csrc
Size: 18368 bytes
Desc: not available
Url : http://lists.denx.de/pipermail/u-boot/attachments/20090610/8464b93f/attachment.c 
author
#11 lore raw

[U-Boot] lcd logo color not correct

Quoted 5 lines · xiangfu

Hi
1. I have try to port u-boot to jz4740 cpu device. now the lcd can show the LOGO.
but the color is not correct.

I fix this. now the LCD is correct.

Quoted 7 lines

2.

	gd->flags |= GD_FLG_DEVINIT;	/* device initialization completed */

when I comment this line. the device can boot to main_loop. when don't comment this line as normal. the system crash. give me some help about this problem

thanks.

Sorry for stupid question :-)
the system not crash
because the stdout=lcd. so the serial have nothing output.