8bit nand bugs

Jiangao,

Can you try the following?

1. Try creating/copying thousands of small files(less than 512bytes) into the partition? 

2. Change the sub-page size from 512 to 2048(page size) and try?

I have done this test:

# mount -t tmpfs -o size=600m tmpfs /var/

# cd var/
# tftp -r u-boot-2010.06-psp04.04.00.01.tar.gz -g 192.168.1.2
# tar jxvf u-boot-2010.06-psp04.04.00.01.tar.gz
# mount -t ubifs ubi0_0 /mnt
# cd /mnt/
# cp /var/u-boot-2010.06-psp04.04.00.01/* ./
# sync

It works well!

I also write a simple nand test application:

  1 #include "stdio.h"
  2 #include "unistd.h"
  3 #include "sys/types.h"
  4 #include "sys/stat.h"
  5 #include "fcntl.h"
  6
  7 #define MAX_SIZE 280
  8
  9 int main()
 10 {
 11     int fd_write, fd_read;
 12     int i = 0,j = 0;
 13     char name[100];
 14     unsigned char buffer[MAX_SIZE];
 15     int size;
 16
 17     for (i=0; i < 10000; i++)
 18     {
 19         for (j=0; j < MAX_SIZE; j++)
 20         {
 21             buffer[j] = rand();
 22         }
 23
 24         sprintf(name, "file%d", i);
 25         fd_write = open(name, O_CREAT|O_RDWR);
 26         write(fd_write, buffer, MAX_SIZE);
 27         close(fd_write);
 28
 29
 30         fd_read = open(name, O_RDWR);
 31         if (read(fd_read, buffer, MAX_SIZE) != MAX_SIZE)
 32             printf("read fail\n");
 33         close(fd_read);
 34
 35         sprintf(name, "newfile%d", i);
 36         fd_write = open(name, O_CREAT|O_RDWR);
 37         write(fd_write, buffer, MAX_SIZE);
 38         close(fd_write);
 39     }
 40
 41     return 0;
 42 }

I run the application on my ubi partion, it works well also.

So, it looks that when i copy lots of files(source code of uboot and linux) to the ubi partion, it remain works well.

But why i tar the u-boot-2010.06-psp04.04.00.01.tar.gz in this partion, it fails ?

By the way, How to change the sub-page size from 512 to 2048(page size) ? I am using BCH8 ecc mode.

Many thanks !

Jiangao,

Can you try switch the IO scheduler in kernel from CFQ to noop or deadline?

To change the sub-page size, when you create the UBI image you can specify, and also you can pass it as part of bootargs also. I can give a sample one. 

ubi.mtd=3,2048 rootfstype=ubifs root=ubi0:rootfs

I have changed IO scheduler to noop or deadline, but it's useless.

I use follows to force subpage as 2048:

# ubiformat /dev/mtd/mtd6 -s 2048 -O 2048
ubiformat: mtd6 (nand), size 268435456 bytes (256.0 MiB), 2048 eraseblocks of 131072 bytes (128.0 KiB), min. I/O size 2048 bytes
libscan: scanning eraseblock 2047 -- 100 % complete
ubiformat: 2047 eraseblocks are supposedly empty
ubiformat: 1 bad eraseblocks found, numbers: 2027
ubiformat: formatting eraseblock 2047 -- 100 % complete
# ubiattach /dev/ubi_ctrl -m 6 -O 2048
UBI: attaching mtd6 to ubi0
UBI: physical eraseblock size:   131072 bytes (128 KiB)
UBI: logical eraseblock size:    126976 bytes
UBI: smallest flash I/O unit:    2048
UBI: sub-page size:              512
UBI: VID header offset:          2048 (aligned 2048)
UBI: data offset:                4096
UBI: max. sequence number:       0
UBI: attached mtd6 to ubi0
UBI: MTD device name:            "File System"
UBI: MTD device size:            256 MiB
UBI: number of good PEBs:        2047
UBI: number of bad PEBs:         1
UBI: number of corrupted PEBs:   0
UBI: max. allowed volumes:       128
UBI: wear-leveling threshold:    4096
UBI: number of internal volumes: 1
UBI: number of user volumes:     0
UBI: available PEBs:             2023
UBI: total number of reserved PEBs: 24
UBI: number of PEBs reserved for bad PEB handling: 20
UBI: max/mean erase counter: 0/0
UBI: image sequence number:  265378814
UBI: background thread "ubi_bgt0d" started, PID 95
UBI device number 0, total 2047 LEBs (259919872 bytes, 247.9 MiB), available 2023 LEBs (256872448 bytes, 245.0 MiB), LEB size 126976 bytes (124.0 KiB)
# ubimkvol /dev/ubi0 -N rootfs -s 80MiB

but it's useless also.

In my options, there is some bugs in nand driver, not ubi. Because when i used another 512MB flash (2K/Page,SLC,Toshiba)，it works very well!

so, it's some difference between K9K8G08U0B and TC58NVG2S3ETA00 ? both of them are SLC and 2KB/Page.

Thanks .

The subpage is not set to 2K properly. You can see from the logs.

jiangao peng101018 said:

UBI: sub-page size:              512

Because of the nand driver, the UBI can find subpages always.

I modify the nand driver, and there is no subpages now.

# ubiformat /dev/mtd/mtd6 -s 2048 -O 2048
ubiformat: mtd6 (nand), size 268435456 bytes (256.0 MiB), 2048 eraseblocks of 131072 bytes (128.0 KiB), min. I/O size 2048 bytes
libscan: scanning eraseblock 95usb 1-1: device descriptor read/64, error -110
libscausb 1-1: new high speed USB device using musb-hdrc and address 3
libscan: scanning eraseblock 2047 -- 100 % complete
ubiformat: 2047 eraseblocks are supposedly empty
ubiformat: 1 bad eraseblocks found, numbers: 2027
ubiformat: formatting eraseblock 2047 -- 100 % complete
# usb 1-1: device descriptor read/64, error -110
ubiattach /dev/ubi_ctrl -m 6 -O 2048
UBI: attaching mtd6 to ubi0
UBI: physical eraseblock size:   131072 bytes (128 KiB)
UBI: logical eraseblock size:    126976 bytes
UBI: smallest flash I/O unit:    2048
UBI: VID header offset:          2048 (aligned 2048)
UBI: data offset:                4096
UBI: max. sequence number:       0
UBI: attached mtd6 to ubi0
UBI: MTD device name:            "File System"
UBI: MTD device size:            256 MiB
UBI: number of good PEBs:        2047
UBI: number of bad PEBs:         1
UBI: number of corrupted PEBs:   0
UBI: max. allowed volumes:       128
UBI: wear-leveling threshold:    4096
UBI: number of internal volumes: 1
UBI: number of user volumes:     0
UBI: available PEBs:             2023
UBI: total number of reserved PEBs: 24
UBI: number of PEBs reserved for bad PEB handling: 20
UBI: max/mean erase counter: 0/0
UBI: image sequence number:  1474856324
UBI: background thread "ubi_bgt0d" started, PID 96
UBI device number 0, total 2047 LEBs (259919872 bytes, 247.9 MiB), available 2023 LEBs (256872448 bytes, 245.0 MiB), LEB size 126976 bytes (124.0 KiB)
# ubimkvol /dev/ubi0 -N rootfs -s 200MiB

But is comes error remain:

# tar jxvf u-boot-2010.06-psp04.04.00.01.tar.gz
......
u-boot-2010.06-psp04.04.00.01/arch/blackfin/include/asm/mach-bf548/BF549_cdef.h
u-boot-2010.06-psp04.04.00.01/arch/blackfin/include/asm/mach-bf548/ADSP-EDN-BF544-extended_cdef.h
UBI error: ubi_io_write: error -5 while writing 2048 bytes to PEB 103:20480, written 0 bytes
UBI warning: ubi_eba_write_leb: failed to write data to PEB 103
UBI: recover PEB 103, move data to PEB 1038
u-boot-2010.06-psp04.04.00.01/arUBI error: ubi_io_write: error -5 while writing 2048 bytes to PEB 1038:2048, written 0 bytes
ch/blackfin/inclUBI warning: recover_peb: failed to write to PEB 1038
ude/asm/mach-bf5UBI: try again
48/ADSP-EDN-BF54UBI: recover PEB 103, move data to PEB 1039
9-extended_cdef.h
UBI: run torture test for PEB 1038
UBI: PEB 1038 passed torture test, do not mark it as bad
UBI: data was successfully recovered
UBI: run torture test for PEB 103

I have added some debug message, and i found "chip->errstat" fails :

nand_do_write_ops -> chip->write_page -> chip->errstat 

if (!cached || !(chip->options & NAND_CACHEPRG)) {

  chip->cmdfunc(mtd, NAND_CMD_PAGEPROG, -1, -1);
  status = chip->waitfunc(mtd, chip);
  /*
   * See if operation failed and additional status checks are
   * available
   */
  if ((status & NAND_STATUS_FAIL) && (chip->errstat))
   status = chip->errstat(mtd, chip, FL_WRITING, status,
            page);

  if (status & NAND_STATUS_FAIL)
  {
   printk("chip->errstat fail %x\n", status);  
   return -EIO;
  }
  
 } else {
  chip->cmdfunc(mtd, NAND_CMD_CACHEDPROG, -1, -1);
  status = chip->waitfunc(mtd, chip);
 }

So , it looks like NAND_CMD_PAGEPROG failed, and chip->errstat failed also.

Can you give me some ideas ? Thank you .

Hi Renjith:

I come into omap_wait function:

static int omap_wait(struct mtd_info *mtd, struct nand_chip *chip) {}

I modify it like this :

if (state == FL_ERASING)
  timeo += (HZ * 400) / 1000;
 else
  timeo += (HZ * 20) / 1000;

to

if (state == FL_ERASING)
  timeo += (HZ * 800) / 1000;
 else
  timeo += (HZ * 200) / 1000;

Now, when i tar the u-boot-2010.06-psp04.01.00.06.patch1.tar.gz in my UBI partion, it works well !

I will do some further testing .

Thanks any way !

Good to know that its solved.

hi TI

I am using am3352 u-boot-2013.01.01-psp05.07.00.00 and linux-3.2.0-psp04.06.00.10

i found nand ubifs error and changed the nand delay as explained before

but still i am facing this issue.. can you suggest any additions ?

log  for reference:

<11>Jan 10 21:28:48 kernel: [ 127.459064] UBI error: ubi_io_write: error -5 while writing 2048 bytes to PEB 263:38912, written 0 bytes
192.168.10.13 15/01 18:05:35.645
<12>Jan 10 21:28:48 kernel: [ 127.469106] UBI warning: ubi_eba_write_leb: failed to write data to PEB 263
192.168.10.13 15/01 18:05:35.646
<13>Jan 10 21:28:48 kernel: [ 127.476464] UBI: recover PEB 263, move data to PEB 264
192.168.10.13 15/01 18:05:35.646
<13>Jan 10 21:28:48 kernel: [ 127.500197] UBI: data was successfully recovered
192.168.10.13 15/01 18:05:35.647
<13>Jan 10 21:28:48 kernel: [ 127.505097] UBI: run torture test for PEB 263
192.168.10.13 15/01 18:05:35.647
<11>Jan 10 21:28:48 kernel: [ 127.511148] UBI error: do_sync_erase: cannot erase PEB 263, error -5
192.168.10.13 15/01 18:05:35.691
<11>Jan 10 21:28:48 kernel: [ 127.517860] UBI error: erase_worker: failed to erase PEB 263, error -5
192.168.10.13 15/01 18:05:35.692
<13>Jan 10 21:28:48 kernel: [ 127.524786] UBI: mark PEB 263 as bad
192.168.10.13 15/01 18:05:35.692
<11>Jan 10 21:28:48 kernel: [ 127.528686] UBI error: ubi_io_mark_bad: cannot mark PEB 263 bad, error -30
192.168.10.13 15/01 18:05:35.692
<12>Jan 10 21:28:48 kernel: [ 127.535933] UBI warning: ubi_ro_mode: switch to read-only mode
192.168.10.13 15/01 18:05:35.692
<11>Jan 10 21:28:48 kernel: [ 127.542094] UBI error: do_work: work failed with error code -30
192.168.10.13 15/01 18:05:35.693
<11>Jan 10 21:28:48 kernel: [ 127.548338] UBI error: ubi_thread: ubi_bgt11d: work failed with error code -30
192.168.10.13 15/01 18:05:35.693
<11>Jan 10 21:29:18 kernel: [ 157.868743] UBIFS error (pid 1299): ubifs_leb_write: writing 2048 bytes to LEB 11:18432 failed, error -30
192.168.10.13 15/01 18:06:06.141
<12>Jan 10 21:29:18 kernel: [ 157.878878] UBIFS warning (pid 1299): ubifs_ro_mode: switched to read-only mode, error -30
192.168.10.13 15/01 18:06:06.142
<12>Jan 10 21:29:18 kernel: [ 157.887588] Backtrace:
192.168.10.13 15/01 18:06:06.142

changes made to nand driver 

------------------------------------------

--- nand_base.c (revision 4)
+++ nand_base.c (revision 457)
@@ -844,11 +844,11 @@

unsigned long timeo = jiffies;
int status, state = chip->state;
-
+ /* FOR WAP, DELAY INCREASED 400 to 800, 20 to 200 */
if (state == FL_ERASING)
- timeo += (HZ * 400) / 1000;
+ timeo += (HZ * 800) / 1000;
else
- timeo += (HZ * 20) / 1000;
+ timeo += (HZ * 200) / 1000;

led_trigger_event(nand_led_trigger, LED_FULL);

regards

Nagendra 

hi  Ti,

i found that the nand ready/busy is not implemented in the kernel linux-3.2.0-psp04.06.00.10 for am335x

is it true ? did i miss something ? 

if not implemented, how to added it ? did we have any patches ?

if rdy/busy is implemented i think the delay implementation is not required. is it ?

for ref:

--------

drivers/mtd/nand/omap2.c

/*
* If RDY/BSY line is connected to OMAP then use the omap ready
* funcrtion and the generic nand_wait function which reads the status
* register after monitoring the RDY/BSY line.Otherwise use a standard
* chip delay which is slightly more than tR (AC Timing) of the NAND
* device and read status register until you get a failure or success
*/
if (pdata->dev_ready) {
info->nand.dev_ready = omap_dev_ready;
info->nand.chip_delay = 0;
} else {  /* controle comes here */
info->nand.waitfunc = omap_wait;
info->nand.chip_delay = 50;
}

can u please calrify this ?

regards,

Nagendra

Nagendra,

nagendra bankupalli said:

i found that the nand ready/busy is not implemented in the kernel linux-3.2.0-psp04.06.00.10 for am335x

This is the DM814x/AM387x forum. For AM335x questions, I would recommend you to ask in the Sitara AM335x forum:

http://e2e.ti.com/support/arm/sitara_arm/f/791.aspx

Regards,
Pavel