Re: [EXT] [PATCH v2 3/3] mtd: rawnand: micron: Address the shallow erase issue

[Boris Brezillon <boris.brezillon@xxxxxxxxxxxxx>]

On Thu, 7 May 2020 09:28:29 +0000
"Bean Huo (beanhuo)" <beanhuo@xxxxxxxxxx> wrote:

> Hi, Boris, Richard 
> 
> > >
> > > A missing VID header plus good payload will cause UBI to stop
> > > attaching since it violates the IO model.  
> > 
> > Sure, and that's not what we want to do anyway. We basically have 2 choices
> > here:
> > 
> > 1/ overwrite all pages starting from page 0 and ending at page 15. This will lead
> > to ECC errors on already written pages, and 0-filled pages for others (unless we
> > go for a raw write, in which case it might actually lead to ECC errors depending
> > on the engine).  
> 
> No need for overwriting all pages. I overwrite EC and VID page just for prevent of 
> Erase power loss issue. But you hate this since FS-specific approach.
> According to Richard's emails, we don't need to overwrite page0,  overwrite page1
> Is enough.

It's still UBI specific, so that's still a no-go for me.

> 
> > 2/ track already written pages (by reading them first), and only writes those that
> > have not been written yet. That means no ECC error in that case, and no
> > corrupted EC/VID header as well.  
> 
> This is similar with my patch approach, but corrupted EC and VID headers. 
> 
> I have two proposals:
> 1.  rebase my patch, and copy one idea from Miquel's patch which records the programmed pages.
> But page 1 should be overwrote considering the UBIFS re-erase mechanism.

See above, and remember we do all that because Micron broke one of the
base assumptions we have about NANDs => erase should reset all bits to
1 or return an error if that fails.

> 
> 2.  add a new padding structure in the MTD,  which is used to filling empty page in the MTD.
> And once FS layer detects this padding data, just ignore or trigger re-erase this block. 

Not sure what you mean by padding structure, but yes, filling empty
pages with 0s and expecting the FS/wear-leveling layers to be happy
with that would be ideal. But, as for the "erase only 2-page because
this is where UBI puts its VID header" thing, that's a NACK on my side
if you want to introduce a new pattern that's only understood by
UBI/UBIFS.

I really wish FS/wear-leveling layers were immune to corrupted/invalid
LEBs they should no longer reference, but according to Richard, that's
not that simple.

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/