linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* mxc_nand.c on mach-imx/imx53
@ 2012-08-23  9:30 Roland Stigge
  2012-08-29  7:41 ` Sascha Hauer
  0 siblings, 1 reply; 8+ messages in thread
From: Roland Stigge @ 2012-08-23  9:30 UTC (permalink / raw)
  To: linux-mtd, linux-kernel, linux-arm-kernel, s.hauer

Hi,

I'd like to know about the current state of mxc_nand.c on imx53, because
I got an error on probe() because the clock that mxc_nand is requesting
is "nfc" but imx53 only defines the clock "mxc_nand".

Even after adjusting the name, probe() couldn't find the clock, so I'm
wondering if mxc_nand.c is supported on imx53 at all because other mx'es
have IMX_HAVE_PLATFORM_MXC_NAND in arch/arm/mach-imx/Kconfig but not imx53.

In a custom board BSP, I found an mxc_nd2.c besides mxc_nand.c but not
in mainline. Are there any plans here?

But maybe I just missed sth. - a hint or link would be very much
appreciated.

Thanks in advance,

Roland

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: mxc_nand.c on mach-imx/imx53
  2012-08-23  9:30 mxc_nand.c on mach-imx/imx53 Roland Stigge
@ 2012-08-29  7:41 ` Sascha Hauer
  2012-08-29 15:20   ` Roland Stigge
  0 siblings, 1 reply; 8+ messages in thread
From: Sascha Hauer @ 2012-08-29  7:41 UTC (permalink / raw)
  To: Roland Stigge; +Cc: linux-mtd, linux-kernel, linux-arm-kernel

Hi Roland,

On Thu, Aug 23, 2012 at 11:30:54AM +0200, Roland Stigge wrote:
> Hi,
> 
> I'd like to know about the current state of mxc_nand.c on imx53, because
> I got an error on probe() because the clock that mxc_nand is requesting
> is "nfc" but imx53 only defines the clock "mxc_nand".

The driver generally works on i.MX53, only some pieces are missing. I
just resent the corresponding series and put you on Cc.

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: mxc_nand.c on mach-imx/imx53
  2012-08-29  7:41 ` Sascha Hauer
@ 2012-08-29 15:20   ` Roland Stigge
  2012-08-29 15:28     ` Eric Bénard
  0 siblings, 1 reply; 8+ messages in thread
From: Roland Stigge @ 2012-08-29 15:20 UTC (permalink / raw)
  To: Sascha Hauer; +Cc: linux-mtd, linux-kernel, linux-arm-kernel

Hi Sascha,

On 08/29/2012 09:41 AM, Sascha Hauer wrote:
> On Thu, Aug 23, 2012 at 11:30:54AM +0200, Roland Stigge wrote:
>> Hi,
>>
>> I'd like to know about the current state of mxc_nand.c on imx53, because
>> I got an error on probe() because the clock that mxc_nand is requesting
>> is "nfc" but imx53 only defines the clock "mxc_nand".
> 
> The driver generally works on i.MX53, only some pieces are missing. I
> just resent the corresponding series and put you on Cc.

Good - thank you for the patches, didn't know about them.

Now, my flash is detected:

NAND device: Manufacturer ID: 0x2c, Chip ID: 0xdc (Micron NAND 512MiB
3,3V 8-bit), page size: 4096, OOB size: 128

However, when I read a partition, kernel (mxc_nand_correct_data_v2_v3())
says:

...
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
...

I'm using the following dt (on an Emtrion DIMM-MX53 board):

                        nand@63fdb000 {
                                status = "okay";
                                #address-cells = <1>;
                                #size-cells = <1>;
                                nand-bus-width = <8>;
                                nand-ecc-mode = "hw";

                                mtd0@00000000 {
                                        label = "bootstrap";
                                        reg = <0x00000000 0x80000>;
                                        read-only;
                                };
                                mtd1@00080000 {
                                        label = "uboot";
                                        reg = <0x00080000 0x80000>;
                                        read-only;
                                };
                                mtd2@00100000 {
                                        label = "NVRAM";
                                        reg = <0x00100000 0x100000>;
                                };
                                mtd3@00200000 {
                                        label = "bootlogo";
                                        reg = <0x00200000 0x80000>;
                                };
                                mtd4@00280000 {
                                        label = "linux";
                                        reg = <0x00280000 0x400000>;
                                };
                                mtd5@00680000 {
                                        label = "rootfs";
                                        reg = <0x00680000 0x8000000>;
                                };
                                mtd6@08680000 {
                                        label = "appfs";
                                        reg = <0x08680000 0x17980000>;
                                };
                        };

NAND info from u-boot:

NAND:  512 MiB
Manufacturer      : Micron (0x2c)
Device Code       : 0xdc
Cell Technology   : SLC
Chip Size         : 512 MiB
Pages per Block   : 64
Page Geometry     : 4096+218
ECC Strength      : 8 bits
ECC Size          : 512 B
Data Setup Time   : 20 ns
Data Hold Time    : 10 ns
Address Setup Time: 10 ns
GPMI Sample Delay : 6 ns
tREA              : Unknown
tRLOH             : Unknown
tRHOH             : Unknown
Description       : MT29F4G08ABAEA

I guess I hit a documented open issue:

/* v3.2b: i.MX53 */
static const struct mxc_nand_devtype_data imx53_nand_devtype_data = {
        .preset = preset_v3,
        .send_cmd = send_cmd_v3,
        .send_addr = send_addr_v3,
        .send_page = send_page_v3,
        .send_read_id = send_read_id_v3,
        .get_dev_status = get_dev_status_v3,
        .check_int = check_int_v3,
        .irq_control = irq_control_v3,
        .get_ecc_status = get_ecc_status_v3,
        .ecclayout_512 = &nandv2_hw_eccoob_smallpage,
        .ecclayout_2k = &nandv2_hw_eccoob_largepage,
        .ecclayout_4k = &nandv2_hw_eccoob_smallpage, /* XXX: needs fix */

Using  ".ecclayout_4k = &nandv2_hw_eccoob_4k" also doesn't work. So is
there any hint how I can proceed from here? Any plans or hint regarding
the "XXX"?

Thanks in advance,

Roland

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: mxc_nand.c on mach-imx/imx53
  2012-08-29 15:20   ` Roland Stigge
@ 2012-08-29 15:28     ` Eric Bénard
  2012-08-31  9:05       ` Roland Stigge
  0 siblings, 1 reply; 8+ messages in thread
From: Eric Bénard @ 2012-08-29 15:28 UTC (permalink / raw)
  To: Roland Stigge; +Cc: Sascha Hauer, linux-mtd, linux-kernel, linux-arm-kernel

Hi Roland,

Le Wed, 29 Aug 2012 17:20:17 +0200,
Roland Stigge <stigge@antcom.de> a écrit :
> Using  ".ecclayout_4k = &nandv2_hw_eccoob_4k" also doesn't work. So is
> there any hint how I can proceed from here? Any plans or hint regarding
> the "XXX"?
> 
you may need something like this patch from barebox :
http://git.pengutronix.de/?p=barebox.git;a=commit;h=632c45795065e6a7471ab82be38e808eb6204341

I may have time to adapt this next week but if you have time before feel
free to do it ;-)

Eric

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: mxc_nand.c on mach-imx/imx53
  2012-08-29 15:28     ` Eric Bénard
@ 2012-08-31  9:05       ` Roland Stigge
  2012-08-31 10:57         ` Roland Stigge
  0 siblings, 1 reply; 8+ messages in thread
From: Roland Stigge @ 2012-08-31  9:05 UTC (permalink / raw)
  To: Eric Bénard; +Cc: Sascha Hauer, linux-mtd, linux-kernel, linux-arm-kernel

Hi Eric,

On 08/29/2012 05:28 PM, Eric Bénard wrote:
> Le Wed, 29 Aug 2012 17:20:17 +0200,
> Roland Stigge <stigge@antcom.de> a écrit :
>> Using  ".ecclayout_4k = &nandv2_hw_eccoob_4k" also doesn't work. So is
>> there any hint how I can proceed from here? Any plans or hint regarding
>> the "XXX"?
>>
> you may need something like this patch from barebox :
> http://git.pengutronix.de/?p=barebox.git;a=commit;h=632c45795065e6a7471ab82be38e808eb6204341

Thank you for the link!

I tried to port this on top of Sascha's patches for imx53, only for mxc v3 for now, as below.

Unfortunately, I still get:

...
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
UnCorrectable RS-ECC Error
...

from mxc_nand.c's mxc_nand_correct_data_v2_v3(). Maybe there's still sth. missing in the
error correction code?

Tracing reveals that in mxc_nand_command()'s switch statement, NAND_CMD_READID is handled,
but the NAND_CMD_PARAM case never gets triggered. (Under which circumstances is this
supposed to happen?) Further, I still wonder why the bootloader identifies the NAND as:

NAND:  512 MiB
Manufacturer      : Micron (0x2c)
Device Code       : 0xdc
Cell Technology   : SLC
Chip Size         : 512 MiB
Pages per Block   : 64
Page Geometry     : 4096+218
ECC Strength      : 8 bits
ECC Size          : 512 B
Data Setup Time   : 20 ns
Data Hold Time    : 10 ns
Address Setup Time: 10 ns
GPMI Sample Delay : 6 ns
tREA              : Unknown
tRLOH             : Unknown
tRHOH             : Unknown
Description       : MT29F4G08ABAEA

But Linux says:

NAND device: Manufacturer ID: 0x2c, Chip ID: 0xdc (Micron NAND 512MiB 3,3V 8-bit), page size: 4096, OOB size: 128

(See OOB size 128 vs. 218).

Thanks in advance,

Roland


---
 drivers/mtd/nand/mxc_nand.c |   36 ++++++++++++++++++++++++++++++++----
 1 file changed, 32 insertions(+), 4 deletions(-)

--- linux-2.6.orig/drivers/mtd/nand/mxc_nand.c
+++ linux-2.6/drivers/mtd/nand/mxc_nand.c
@@ -150,6 +150,7 @@ struct mxc_nand_devtype_data {
        void (*send_addr)(struct mxc_nand_host *, uint16_t, int);
        void (*send_page)(struct mtd_info *, unsigned int);
        void (*send_read_id)(struct mxc_nand_host *);
+       void (*send_read_param)(struct mxc_nand_host *);
        uint16_t (*get_dev_status)(struct mxc_nand_host *);
        int (*check_int)(struct mxc_nand_host *);
        void (*irq_control)(struct mxc_nand_host *, int);
@@ -543,6 +544,16 @@ static void send_read_id_v3(struct mxc_n
        memcpy32_fromio(host->data_buf, host->main_area0, 16);
 }
 
+static void send_read_param_v3(struct mxc_nand_host *host)
+{
+       /* Read param into main buffer */
+       writel(NFC_OUTPUT, NFC_V3_LAUNCH);
+
+       wait_op_done(host, true);
+
+       memcpy32_fromio(host->data_buf, host->main_area0, 1024);
+}
+
 /* Request the NANDFC to perform a read of the NAND device ID. */
 static void send_read_id_v1_v2(struct mxc_nand_host *host)
 {
@@ -741,6 +752,10 @@ static void mxc_nand_read_buf(struct mtd
 
        n = min(n, len);
 
+       /* handle the read param special case */
+       if ((mtd->writesize == 0) && (len != 0))
+               n = len;
+
        memcpy(buf, host->data_buf + col, n);
 
        host->buf_start += n;
@@ -842,9 +857,12 @@ static void mxc_do_addr_cycle(struct mtd
                 * MXC NANDFC can only perform full page+spare or
                 * spare-only read/write.  When the upper layers
                 * perform a read/write buf operation, the saved column
-                 * address is used to index into the full page.
+                * address is used to index into the full page.
+                *
+                * The colum address must be sent to the flash in
+                * order to get the ONFI header (0x20)
                 */
-               host->devtype_data->send_addr(host, 0, page_addr == -1);
+               host->devtype_data->send_addr(host, column, page_addr == -1);
                if (mtd->writesize > 512)
                        /* another col addr cycle for 2k page */
                        host->devtype_data->send_addr(host, 0, false);
@@ -998,9 +1016,11 @@ static void preset_v3(struct mtd_info *m
 
        writel(0, NFC_V3_IPC);
 
+       /* if the flash has a 224 oob, the NFC must be configured to 218 */
        config2 = NFC_V3_CONFIG2_ONE_CYCLE |
                NFC_V3_CONFIG2_2CMD_PHASES |
-               NFC_V3_CONFIG2_SPAS(mtd->oobsize >> 1) |
+               NFC_V3_CONFIG2_SPAS(((mtd->oobsize > 218) ?
+                                    218 : mtd->oobsize) >> 1) |
                NFC_V3_CONFIG2_ST_CMD(0x70) |
                NFC_V3_CONFIG2_INT_MSK |
                NFC_V3_CONFIG2_NUM_ADDR_PHASE0;
@@ -1124,6 +1144,13 @@ static void mxc_nand_command(struct mtd_
                host->buf_start = column;
                break;
 
+       case NAND_CMD_PARAM:
+               host->devtype_data->send_cmd(host, command, true);
+               mxc_do_addr_cycle(mtd, column, page_addr);
+               host->devtype_data->send_read_param(host);
+               host->buf_start = column;
+               break;
+
        case NAND_CMD_ERASE1:
        case NAND_CMD_ERASE2:
                host->devtype_data->send_cmd(host, command, false);
@@ -1271,13 +1298,14 @@ static const struct mxc_nand_devtype_dat
        .send_addr = send_addr_v3,
        .send_page = send_page_v3,
        .send_read_id = send_read_id_v3,
+       .send_read_param = send_read_param_v3,
        .get_dev_status = get_dev_status_v3,
        .check_int = check_int_v3,
        .irq_control = irq_control_v3,
        .get_ecc_status = get_ecc_status_v3,
        .ecclayout_512 = &nandv2_hw_eccoob_smallpage,
        .ecclayout_2k = &nandv2_hw_eccoob_largepage,
-       .ecclayout_4k = &nandv2_hw_eccoob_smallpage, /* XXX: needs fix */
+       .ecclayout_4k = &nandv2_hw_eccoob_4k,
        .select_chip = mxc_nand_select_chip_v1_v3,
        .correct_data = mxc_nand_correct_data_v2_v3,
        .irqpending_quirk = 0,

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: mxc_nand.c on mach-imx/imx53
  2012-08-31  9:05       ` Roland Stigge
@ 2012-08-31 10:57         ` Roland Stigge
  2012-09-03  8:55           ` Sascha Hauer
  0 siblings, 1 reply; 8+ messages in thread
From: Roland Stigge @ 2012-08-31 10:57 UTC (permalink / raw)
  To: Eric Bénard; +Cc: Sascha Hauer, linux-mtd, linux-kernel, linux-arm-kernel

On 08/31/2012 11:05 AM, Roland Stigge wrote:
> I tried to port this on top of Sascha's patches for imx53, only for mxc v3 for now, as below.
> 
> Unfortunately, I still get:
> 
> ...
> UnCorrectable RS-ECC Error
> UnCorrectable RS-ECC Error
> UnCorrectable RS-ECC Error
> UnCorrectable RS-ECC Error
> UnCorrectable RS-ECC Error
> ...
> 
> from mxc_nand.c's mxc_nand_correct_data_v2_v3(). Maybe there's still sth. missing in the
> error correction code?

With the below incremental fix, it works. :-)

Thanks,

Roland

PS: Will you prepare a patch for l2-mtd.git? I only have an imx53, can't test the v1/v2 etc.
configs.


--- linux-2.6.orig/drivers/mtd/nand/mxc_nand.c
+++ linux-2.6/drivers/mtd/nand/mxc_nand.c
@@ -1141,14 +1141,14 @@ static void mxc_nand_command(struct mtd_
                host->devtype_data->send_cmd(host, command, true);
                mxc_do_addr_cycle(mtd, column, page_addr);
                host->devtype_data->send_read_id(host);
-               host->buf_start = column;
+               host->buf_start = 0;
                break;
 
        case NAND_CMD_PARAM:
                host->devtype_data->send_cmd(host, command, true);
                mxc_do_addr_cycle(mtd, column, page_addr);
                host->devtype_data->send_read_param(host);
-               host->buf_start = column;
+               host->buf_start = 0;
                break;
 
        case NAND_CMD_ERASE1:

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: mxc_nand.c on mach-imx/imx53
  2012-08-31 10:57         ` Roland Stigge
@ 2012-09-03  8:55           ` Sascha Hauer
  2012-09-03  9:10             ` Roland Stigge
  0 siblings, 1 reply; 8+ messages in thread
From: Sascha Hauer @ 2012-09-03  8:55 UTC (permalink / raw)
  To: Roland Stigge; +Cc: Eric Bénard, linux-mtd, linux-kernel, linux-arm-kernel

On Fri, Aug 31, 2012 at 12:57:37PM +0200, Roland Stigge wrote:
> On 08/31/2012 11:05 AM, Roland Stigge wrote:
> > I tried to port this on top of Sascha's patches for imx53, only for mxc v3 for now, as below.
> > 
> > Unfortunately, I still get:
> > 
> > ...
> > UnCorrectable RS-ECC Error
> > UnCorrectable RS-ECC Error
> > UnCorrectable RS-ECC Error
> > UnCorrectable RS-ECC Error
> > UnCorrectable RS-ECC Error

Have you written the data you are unable to read with this driver?

> > ...
> > 
> > from mxc_nand.c's mxc_nand_correct_data_v2_v3(). Maybe there's still sth. missing in the
> > error correction code?
> 
> With the below incremental fix, it works. :-)
> 
> Thanks,
> 
> Roland
> 
> PS: Will you prepare a patch for l2-mtd.git? I only have an imx53, can't test the v1/v2 etc.
> configs.
> 
> 
> --- linux-2.6.orig/drivers/mtd/nand/mxc_nand.c
> +++ linux-2.6/drivers/mtd/nand/mxc_nand.c
> @@ -1141,14 +1141,14 @@ static void mxc_nand_command(struct mtd_
>                 host->devtype_data->send_cmd(host, command, true);
>                 mxc_do_addr_cycle(mtd, column, page_addr);
>                 host->devtype_data->send_read_id(host);
> -               host->buf_start = column;
> +               host->buf_start = 0;
>                 break;
>  
>         case NAND_CMD_PARAM:
>                 host->devtype_data->send_cmd(host, command, true);
>                 mxc_do_addr_cycle(mtd, column, page_addr);
>                 host->devtype_data->send_read_param(host);
> -               host->buf_start = column;
> +               host->buf_start = 0;

On what tree are you working? I cannot find a handling of
NAND_CMD_PARAM in my tree.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: mxc_nand.c on mach-imx/imx53
  2012-09-03  8:55           ` Sascha Hauer
@ 2012-09-03  9:10             ` Roland Stigge
  0 siblings, 0 replies; 8+ messages in thread
From: Roland Stigge @ 2012-09-03  9:10 UTC (permalink / raw)
  To: Sascha Hauer; +Cc: Eric Bénard, linux-mtd, linux-kernel, linux-arm-kernel

On 09/03/2012 10:55 AM, Sascha Hauer wrote:
> On Fri, Aug 31, 2012 at 12:57:37PM +0200, Roland Stigge wrote:
>> On 08/31/2012 11:05 AM, Roland Stigge wrote:
>>> I tried to port this on top of Sascha's patches for imx53, only for mxc v3 for now, as below.
>>>
>>> Unfortunately, I still get:
>>>
>>> ...
>>> UnCorrectable RS-ECC Error
>>> UnCorrectable RS-ECC Error
>>> UnCorrectable RS-ECC Error
>>> UnCorrectable RS-ECC Error
>>> UnCorrectable RS-ECC Error
> 
> Have you written the data you are unable to read with this driver?

No, maybe that's one of the reasons. The bootloader also seems to
implement a different on-flash layout. But as written, the problem is
gone with the below fix.

>> --- linux-2.6.orig/drivers/mtd/nand/mxc_nand.c
>> +++ linux-2.6/drivers/mtd/nand/mxc_nand.c
>> @@ -1141,14 +1141,14 @@ static void mxc_nand_command(struct mtd_
>>                 host->devtype_data->send_cmd(host, command, true);
>>                 mxc_do_addr_cycle(mtd, column, page_addr);
>>                 host->devtype_data->send_read_id(host);
>> -               host->buf_start = column;
>> +               host->buf_start = 0;
>>                 break;
>>  
>>         case NAND_CMD_PARAM:
>>                 host->devtype_data->send_cmd(host, command, true);
>>                 mxc_do_addr_cycle(mtd, column, page_addr);
>>                 host->devtype_data->send_read_param(host);
>> -               host->buf_start = column;
>> +               host->buf_start = 0;
> 
> On what tree are you working? I cannot find a handling of
> NAND_CMD_PARAM in my tree.

That's on top of the patch I ported from Eric's barebox driver (other
mail, but only for imx53/v3 for now). Sorry for the confusion.

I can prepare a single patch, but maybe Eric's more knowledgeable on
this driver (including v1/v2 that I can't test), he announced a patch
for this week.

Roland

^ permalink raw reply	[flat|nested] 8+ messages in thread
end of thread, other threads:[~2012-09-03  9:10 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-08-23  9:30 mxc_nand.c on mach-imx/imx53 Roland Stigge
2012-08-29  7:41 ` Sascha Hauer
2012-08-29 15:20   ` Roland Stigge
2012-08-29 15:28     ` Eric Bénard
2012-08-31  9:05       ` Roland Stigge
2012-08-31 10:57         ` Roland Stigge
2012-09-03  8:55           ` Sascha Hauer
2012-09-03  9:10             ` Roland Stigge

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).