lfs error:969 Corrupted dir pair at 0 1

[safirxps]

I got strange error at the begining of my software my code is as follow :

int filesystem_init(void)
{
  int err = 0;
  err = lfs_mount(&Filesystem_Handle, &Filesystem_Config);
  if(err != 0)
  {
   err = lfs_format(&Filesystem_Handle, &Filesystem_Config);
    if(err == 0)
      err = lfs_mount(&Filesystem_Handle, &Filesystem_Config);
  }
  return err;
}

using latest Littlefs version

my config is :

#define BASE_BLOCK_NUM 64 // this is the number of block on the lower side of the flash not included in Filesystem each Block is 4096 Bytes

#define READ_SIZE 1
#define PROG_SIZE 1
#define ERASE_SIZE 4096
#define APP_SIZE BASE_BLOCK_NUM*ERASE_SIZE
#define FLASH_SIZE_ 16384 * ERASE_SIZE // There are 16384 blocks of 4096 Bytes in 64 MBytes Flash

const struct lfs_config Filesystem_Config =
{
  // block device driver context data
  .context = NULL,

  // block device operations
  .read = lfs_read_page,
  .prog = lfs_device_prog,
  .erase = lfs_erase_page,
  .sync = lfs_device_sync,

  // block device configuration
  .read_size = READ_SIZE,
  .prog_size = PROG_SIZE,
  .block_size = ERASE_SIZE,
  .cache_size = 256,
  .block_count = (FLASH_SIZE_-APP_SIZE) / ERASE_SIZE,
  .lookahead_size = 256,
  .block_cycles = 100,
};

[lsilvaalmeida]

In the first mount is normal have this error because lfs can't found the superblock. But, after lfs_format, this error should go away. When are you having this error?

[safirxps]

Hi, Thanks it is working now. New problem after filewrite, I did the following :

filesystem_filewrite("TestPol1",&polo,sizeof(testva));

with :

int filesystem_filewrite(const char* filename, void *buffer, lfs_size_t size)
{
  lfs_file_t Filesystem_Filetype;
  int err = 0;
  err=lfs_file_open(&Filesystem_Handle, &Filesystem_Filetype, filename, LFS_O_WRONLY  | LFS_O_CREAT);
  if(err == 0)
  {
    lfs_size_t WriteSize=lfs_file_write(&Filesystem_Handle, &Filesystem_Filetype, buffer, size);
    err = WriteSize;
  }
  lfs_file_close(&Filesystem_Handle, &Filesystem_Filetype);
  return err;
}

The result of filewrite is ok, with the size of the block, however when calling file open to read the size just after it can not find the file, I look at in Flash with debugger , and I can find file name I was using, what can be wrong ?

lfs_file_open(&Filesystem_Handle, &Filesystem_Filetype, "TestPol1", LFS_O_RDONLY) 

return file name not existing.

[lsilvaalmeida]

Hello, First of all, the size in lfs_file_write should be the size of data pointed by buffer. So, filesystem_filewrite("TestPol1",&polo,sizeof(polo)) should be better. The same for reading. Maybe this wrong call of write has corrupted the file, I don't know.

[KRekha3]

Hi @lsilvaalmeida , @safirxps even i am facing similar issues, i have declared prog, read, lookahead size as 256bytes, if i try to write more than 256 it is corrupting mounted filesystem. what might be the mistake? your earliest reply would help me a lot

[geky]

Hi @safirxps, thanks for creating an issue.

Are there any errors returned by lfs_file_close? The data usually not committed until lfs_file_close. You can commit earlier (and check for errors) with lfs_file_sync if you need to.

If there's no error, are you able to post a dump of the filesystem? I'm not sure what could be wrong.

Also what chip are you using? Does it support 4KiB erase blocks? I've seen a number of MCU internal flash require much larger erase blocks.

[KRekha3]

Hi I have resolved the issue.. in my case issue was with erase function which I was using was incorrect. It was reading more than expected area.. Thanks for your reply too. It would help in future

On Thu, Jul 18, 2019, 12:53 PM geniusgogo notifications@github.com wrote:

but i switch littlefs v1.7 no error。

[I/SF] SPI Flash JEDEC-ID Detection(ID:c84017) [I/SF] SPI Flash Capacity:(8MB) [I/INIT] File System initialized! msh />

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ARMmbed/littlefs/issues/206?email_source=notifications&email_token=AMDC2CCDGSF4JCQLRP452JTQAAK5ZA5CNFSM4HYIPNUKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2HSK5Q#issuecomment-512697718, or mute the thread https://github.com/notifications/unsubscribe-auth/AMDC2CGNL6JAE4H26GGPVGLQAAK5ZANCNFSM4HYIPNUA .