Re: [PATCH v2] mtd: nand: brcmnand: fix mtd corrected bits stat

[Florian Fainelli]

On 6/18/25 02:07, Miquel Raynal wrote:
> Hello David,
>
> > I'm not familiar with nandbiterrs but here's the results from
> > mtd_nandbiterrs.ko on my NAND set to BCH8:
> >
> > # insmod mtd_nandbiterrs.ko dev=0
> > [  676.097190]
> > [  676.098760] ==================================================
> > [  676.104609] mtd_nandbiterrs: MTD device: 0
> > [  676.108732] mtd_nandbiterrs: MTD device size 2097152,
> > eraseblock=262144, page=4096, oob=216
> > [  676.117089] mtd_nandbiterrs: Device uses 1 subpages of 4096 bytes
> > [  676.123188] mtd_nandbiterrs: Using page=0, offset=0, eraseblock=0
> > [  676.130863] mtd_nandbiterrs: incremental biterrors test
> > [  676.136154] mtd_nandbiterrs: write_page
> > [  676.140761] mtd_nandbiterrs: rewrite page
> > [  676.145473] mtd_nandbiterrs: read_page
> > [  676.149621] mtd_nandbiterrs: verify_page
> > [  676.153625] mtd_nandbiterrs: Successfully corrected 0 bit errors per subpage
> > [  676.160678] mtd_nandbiterrs: Inserted biterror @ 0/5
> > [  676.165647] mtd_nandbiterrs: rewrite page
> > [  676.170363] mtd_nandbiterrs: read_page
> > [  676.174508] mtd_nandbiterrs: Read reported 1 corrected bit errors
> > [  676.180606] mtd_nandbiterrs: verify_page
> > [  676.184609] mtd_nandbiterrs: Successfully corrected 1 bit errors per subpage
> > [  676.191662] mtd_nandbiterrs: Inserted biterror @ 0/2
> > [  676.196631] mtd_nandbiterrs: rewrite page
> > [  676.201342] mtd_nandbiterrs: read_page
> > [  676.205487] mtd_nandbiterrs: Read reported 2 corrected bit errors
> > [  676.211586] mtd_nandbiterrs: verify_page
> > [  676.215588] mtd_nandbiterrs: Successfully corrected 2 bit errors per subpage
> > [  676.222641] mtd_nandbiterrs: Inserted biterror @ 0/0
> > [  676.227608] mtd_nandbiterrs: rewrite page
> > [  676.228356] mtd_nandbiterrs: read_page
> > [  676.228749] mtd_nandbiterrs: Read reported 3 corrected bit errors
> > [  676.228751] mtd_nandbiterrs: verify_page
> > [  676.228829] mtd_nandbiterrs: Successfully corrected 3 bit errors per subpage
> > [  676.228831] mtd_nandbiterrs: Inserted biterror @ 1/7
> > [  676.228833] mtd_nandbiterrs: rewrite page
> > [  676.229530] mtd_nandbiterrs: read_page
> > [  676.229922] mtd_nandbiterrs: Read reported 4 corrected bit errors
> > [  676.229924] mtd_nandbiterrs: verify_page
> > [  676.230001] mtd_nandbiterrs: Successfully corrected 4 bit errors per subpage
> > [  676.230003] mtd_nandbiterrs: Inserted biterror @ 1/5
> > [  676.230005] mtd_nandbiterrs: rewrite page
> > [  676.294177] mtd_nandbiterrs: read_page
> > [  676.298337] mtd_nandbiterrs: Read reported 5 corrected bit errors
> > [  676.304436] mtd_nandbiterrs: verify_page
> > [  676.308441] mtd_nandbiterrs: Successfully corrected 5 bit errors per subpage
> > [  676.315494] mtd_nandbiterrs: Inserted biterror @ 1/2
> > [  676.320464] mtd_nandbiterrs: rewrite page
> > [  676.325174] mtd_nandbiterrs: read_page
> > [  676.329327] mtd_nandbiterrs: Read reported 6 corrected bit errors
> > [  676.335426] mtd_nandbiterrs: verify_page
> > [  676.339429] mtd_nandbiterrs: Successfully corrected 6 bit errors per subpage
> > [  676.346483] mtd_nandbiterrs: Inserted biterror @ 1/0
> > [  676.351452] mtd_nandbiterrs: rewrite page
> > [  676.356162] mtd_nandbiterrs: read_page
> > [  676.360308] mtd_nandbiterrs: Read reported 7 corrected bit errors
> > [  676.366407] mtd_nandbiterrs: verify_page
> > [  676.370409] mtd_nandbiterrs: Successfully corrected 7 bit errors per subpage
> > [  676.377462] mtd_nandbiterrs: Inserted biterror @ 2/6
> > [  676.382432] mtd_nandbiterrs: rewrite page
> > [  676.387142] mtd_nandbiterrs: read_page
> > [  676.391287] mtd_nandbiterrs: Read reported 8 corrected bit errors
> > [  676.397385] mtd_nandbiterrs: verify_page
> > [  676.401388] mtd_nandbiterrs: Successfully corrected 8 bit errors
> > per subpage
>
> So far the reporting looks good (and the nandflipbits output looks
> correct as well).
>
> > [  676.408441] mtd_nandbiterrs: Inserted biterror @ 2/5
> > [  676.413411] mtd_nandbiterrs: rewrite page
> > [  676.418122] mtd_nandbiterrs: read_page
> > [  676.422267] mtd_nandbiterrs: verify_page
> > [  676.426194] mtd_nandbiterrs: Error: page offset 0, expected 25, got 00
> > [  676.432727] mtd_nandbiterrs: Error: page offset 1, expected a5, got 00
> > [  676.439260] mtd_nandbiterrs: Error: page offset 2, expected 65, got 05
> > [  676.445868] mtd_nandbiterrs: ECC failure, read data is incorrect
> > despite read success
>
> Here however there is something wrong. We do expect a read failure,
> instead of returning wrong data. There is still a problem, I do not know
> if this problem was there before though, but this must be fixed.
>
> Hello Florian, if you have time, I'd welcome you opinion on this patch :)

I don't think I am entitled any opinion on this patch other than 
confirming the register offset locations. Both William and David have 
far more experience with this NAND controller and devices that I do.
--
Florian