Re: UBI Create or Update Volume

Re: UBI Create or Update Volume

[BRIJESH SINGH]

 [Sorry. I will try to take care of it in future emails.]
Hi,
On Wed, Jun 17, 2009 at 3:26 PM, Artem Bityutskiy <dedekind@infradead.org> wrote:
     On Wed, 2009-06-17 at 09:33 +0000, BRIJESH SINGH wrote:
      >> Hi Artem,
      >>      Just came across the ubi_change_vtbl_record function. If first copy is written successfully and error occurs while writing second copy, this function still returns error. And hence the creation or change in Volume becomes unsuccessful. So applications consider it failure.
      >>
      >>   But if I unmount now, (clean or unclean) the first copy will become valid and second will be called corrupted (during mount time). It's not correct.
      >>    So should ubi_change_vtbl_record  call flush wl and  return success? Because 2nd copy(old) is unmapped till now.And should UBI go to read-only mode?
> Yes, sounds right. Would you send a patch?
Yes, I will send a patch once I test it properly.

---
Brijesh

Re: UBI Create or Update Volume

[Artem Bityutskiy]

[Very difficult to read your email due to
looooooooooooonnnnnnnnnnnnnnnnnnngggggggggggggggggg
lines]

Hi,

On Wed, 2009-06-17 at 09:33 +0000, BRIJESH SINGH wrote:
> Hi Artem, 
>      Just came across the ubi_change_vtbl_record function. If first copy is written successfully and error occurs while writing second copy, this function still returns error. And hence the creation or change in Volume becomes unsuccessful. So applications consider it failure.
> 
>    But if I unmount now, (clean or unclean) the first copy will become valid and second will be called corrupted (during mount time). It's not correct.
>    So should ubi_change_vtbl_record  call flush wl and  return success? Because 2nd copy(old) is unmapped till now.And should UBI go to read-only mode?

Yes, sounds right. Would you send a patch?

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

UBI Create or Update Volume

[BRIJESH SINGH]

Hi Artem, 
     Just came across the ubi_change_vtbl_record function. If first copy is written successfully and error occurs while writing second copy, this function still returns error. And hence the creation or change in Volume becomes unsuccessful. So applications consider it failure.

   But if I unmount now, (clean or unclean) the first copy will become valid and second will be called corrupted (during mount time). It's not correct.
   So should ubi_change_vtbl_record  call flush wl and  return success? Because 2nd copy(old) is unmapped till now.And should UBI go to read-only mode?


Code Snip:
Funct ubi_change_vtbl_record(){
...
for (i = 0; i < UBI_LAYOUT_VOLUME_EBS; i++) {
                 err = ubi_eba_unmap_leb(ubi, layout_vol, i);
                 if (err)
                        return err;
                      err = ubi_eba_write_leb(ubi, layout_vol, i, ubi->vtbl, 0,
                                                                 ubi->vtbl_size, UBI_LONGTERM);
                        if (err)
                          return err;
 }
...
}

Thanks
Brijesh