Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 13 Apr 2023 10:22:00 +0200
From:   Miquel Raynal <miquel.raynal@bootlin.com>
To:     Arseniy Krasnov <avkrasnov@sberdevices.ru>
Cc:     Liang Yang <liang.yang@amlogic.com>,
        Richard Weinberger <richard@nod.at>,
        Vignesh Raghavendra <vigneshr@ti.com>,
        Neil Armstrong <neil.armstrong@linaro.org>,
        Kevin Hilman <khilman@baylibre.com>,
        Jerome Brunet <jbrunet@baylibre.com>,
        Martin Blumenstingl <martin.blumenstingl@googlemail.com>,
        Jianxin Pan <jianxin.pan@amlogic.com>,
        Yixun Lan <yixun.lan@amlogic.com>, <oxffffaa@gmail.com>,
        <kernel@sberdevices.ru>, <linux-mtd@lists.infradead.org>,
        <linux-arm-kernel@lists.infradead.org>,
        <linux-amlogic@lists.infradead.org>,
        <linux-kernel@vger.kernel.org>,
        "yonghui.yu" <yonghui.yu@amlogic.com>
Subject: Re: [PATCH v1 4/5] mtd: rawnand: meson: clear OOB buffer before
 read
Message-ID: <20230413102200.309fbe9c@xps-13>
In-Reply-To: <7520e512-8a19-ff04-364c-b5be0a579ef0@sberdevices.ru>
References: <20230412061700.1492474-1-AVKrasnov@sberdevices.ru>
        <20230412061700.1492474-5-AVKrasnov@sberdevices.ru>
        <20230412094400.3c82f631@xps-13>
        <ac4b66da-6a76-c2ec-7e21-31632f3448d5@sberdevices.ru>
        <20230412113654.183350d0@xps-13>
        <4eace0a0-f6af-7d99-a52f-7913a2139330@sberdevices.ru>
        <20230412141824.755b2bca@xps-13>
        <eedaaed9-0a41-2c18-9eb2-792613566986@sberdevices.ru>
        <20230412145715.58c2be4a@xps-13>
        <7c996832-258f-001c-56bd-87bbdf23eeaa@amlogic.com>
        <20230412163214.0c764bb3@xps-13>
        <0c61eaae-053e-5768-a533-70b2ff0cc95d@amlogic.com>
        <fbb8cb2b-0996-3029-b368-f67087d73487@amlogic.com>
        <7520e512-8a19-ff04-364c-b5be0a579ef0@sberdevices.ru>
Organization: Bootlin
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi Arseniy,

avkrasnov@sberdevices.ru wrote on Thu, 13 Apr 2023 10:00:24 +0300:

> On 13.04.2023 09:11, Liang Yang wrote:
> >=20
> > On 2023/4/13 13:32, Liang Yang wrote: =20
> >> Hi Miquel,
> >>
> >> On 2023/4/12 22:32, Miquel Raynal wrote: =20
> >>> [ EXTERNAL EMAIL ]
> >>>
> >>> Hello,
> >>>
> >>> liang.yang@amlogic.com wrote on Wed, 12 Apr 2023 22:04:28 +0800:
> >>> =20
> >>>> Hi Miquel and Arseniy,
> >>>>
> >>>> On 2023/4/12 20:57, Miquel Raynal wrote: =20
> >>>>> [ EXTERNAL EMAIL ]
> >>>>>
> >>>>> Hi Arseniy,
> >>>>>
> >>>>> avkrasnov@sberdevices.ru wrote on Wed, 12 Apr 2023 15:22:26 +0300: =
=20
> >>>>>> On 12.04.2023 15:18, Miquel Raynal wrote: =20
> >>>>>>> Hi Arseniy,
> >>>>>>>
> >>>>>>> avkrasnov@sberdevices.ru wrote on Wed, 12 Apr 2023 13:14:52 +0300:
> >>>>>>> =C2=A0=C2=A0=C2=A0 >>>> On 12.04.2023 12:36, Miquel Raynal wrote:=
 =20
> >>>>>>>>> Hi Arseniy,
> >>>>>>>>>
> >>>>>>>>> avkrasnov@sberdevices.ru wrote on Wed, 12 Apr 2023 12:20:55 +03=
00:
> >>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 >>>>>> On 12.04.2023 10:44, Miqu=
el Raynal wrote: =20
> >>>>>>>>>>> Hi Arseniy,
> >>>>>>>>>>>
> >>>>>>>>>>> AVKrasnov@sberdevices.ru wrote on Wed, 12 Apr 2023 09:16:58 +=
0300:
> >>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 >>>>>>>> This NAND=
 reads only few user's bytes in ECC mode (not full OOB), so
> >>>>>>>>>>>
> >>>>>>>>>>> "This NAND reads" does not look right, do you mean "Subpage r=
eads do
> >>>>>>>>>>> not retrieve all the OOB bytes,"?
> >>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 >>>>>>>> fill OOB =
buffer with zeroes to not return garbage from previous reads =20
> >>>>>>>>>>>> to user.
> >>>>>>>>>>>> Otherwise 'nanddump' utility prints something like this for =
just erased
> >>>>>>>>>>>> page:
> >>>>>>>>>>>>
> >>>>>>>>>>>> ...
> >>>>>>>>>>>> 0x000007f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> >>>>>>>>>>>> =C2=A0=C2=A0=C2=A0 OOB Data: ff ff ff ff 00 00 ff ff 80 cf 2=
2 99 cb ad d3 be
> >>>>>>>>>>>> =C2=A0=C2=A0=C2=A0 OOB Data: 63 27 ae 06 16 0a 2f eb bb dd 4=
6 74 41 8e 88 6e
> >>>>>>>>>>>> =C2=A0=C2=A0=C2=A0 OOB Data: 38 a1 2d e6 77 d4 05 06 f2 a5 7=
e 25 eb 34 7c ff
> >>>>>>>>>>>> =C2=A0=C2=A0=C2=A0 OOB Data: 38 ea de 14 10 de 9b 40 33 16 6=
a cc 9d aa 2f 5e
> >>>>>>>>>>>>
> >>>>>>>>>>>> Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
> >>>>>>>>>>>> ---
> >>>>>>>>>>>> =C2=A0=C2=A0 drivers/mtd/nand/raw/meson_nand.c | 5 +++++
> >>>>>>>>>>>> =C2=A0=C2=A0 1 file changed, 5 insertions(+)
> >>>>>>>>>>>>
> >>>>>>>>>>>> diff --git a/drivers/mtd/nand/raw/meson_nand.c b/drivers/mtd=
/nand/raw/meson_nand.c
> >>>>>>>>>>>> index f84a10238e4d..f2f2472cb511 100644
> >>>>>>>>>>>> --- a/drivers/mtd/nand/raw/meson_nand.c
> >>>>>>>>>>>> +++ b/drivers/mtd/nand/raw/meson_nand.c
> >>>>>>>>>>>> @@ -858,9 +858,12 @@ static int meson_nfc_read_page_sub(stru=
ct nand_chip *nand,
> >>>>>>>>>>>> =C2=A0=C2=A0 static int meson_nfc_read_page_raw(struct nand_=
chip *nand, u8 *buf,
> >>>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 int oob_=
required, int page)
> >>>>>>>>>>>> =C2=A0=C2=A0 {
> >>>>>>>>>>>> +=C2=A0=C2=A0=C2=A0 struct mtd_info *mtd =3D nand_to_mtd(nan=
d);
> >>>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 u8 *oob_buf =3D nand->o=
ob_poi;
> >>>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 int ret;
> >>>>>>>>>>>> =C2=A0=C2=A0 >>>>>>>> +=C2=A0=C2=A0=C2=A0 memset(oob_buf, 0,=
 mtd->oobsize); =20
> >>>>>>>>>>>
> >>>>>>>>>>> I'm surprised raw reads do not read the entire OOB? =20
> >>>>>>>>>>
> >>>>>>>>>> Yes! Seems in case of raw access (what i see in this driver) n=
umber of OOB bytes read
> >>>>>>>>>> still depends on ECC parameters: for each portion of data cove=
red with ECC code we can
> >>>>>>>>>> read it's ECC code and "user bytes" from OOB - it is what i se=
e by dumping DMA buffer by
> >>>>>>>>>> printk(). For example I'm working with 2K NAND pages, each pag=
e has 2 x 1K ECC blocks.
> >>>>>>>>>> For each ECC block I have 16 OOB bytes which I can access by r=
ead/write. Each 16 bytes
> >>>>>>>>>> contains 2 bytes of user's data and 14 bytes ECC codes. So whe=
n I read page in raw mode
> >>>>>>>>>> controller returns 32 bytes (2 x (2 + 14)) of OOB. While OOB i=
s reported as 64 bytes. =20
> >>>>>>>>>
> >>>>>>>>> In all modes, when you read OOB, you should get the full OOB. T=
he fact
> >>>>>>>>> that ECC correction is enabled or disabled does not matter. If =
the NAND
> >>>>>>>>> features OOB sections of 64 bytes, you should get the 64 bytes.
> >>>>>>>>>
> >>>>>>>>> What happens sometimes, is that some of the bytes are not prote=
cted
> >>>>>>>>> against bitflips, but the policy is to return the full buffer. =
=20
> >>>>>>>>
> >>>>>>>> Ok, so to clarify case for this NAND controller:
> >>>>>>>> 1) In both ECC and raw modes i need to return the same raw OOB d=
ata (e.g. user bytes
> >>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0 + ECC codes)? =20
> >>>>>>>
> >>>>>>> Well, you need to cover the same amount of data, yes. But in the =
ECC
> >>>>>>> case the data won't be raw (at least not all of it). =20
> >>>>>>
> >>>>>> So "same amount of data", in ECC mode current implementation retur=
ns only user OOB bytes (e.g.
> >>>>>> OOB data excluding ECC codes), in raw it returns user bytes + ECC =
codes. IIUC correct
> >>>>>> behaviour is to always return user bytes + ECC codes as OOB data e=
ven in ECC mode ? =20
> >>>>>
> >>>>> If the page are 2k+64B you should read 2k+64B when OOB are requeste=
d.
> >>>>>
> >>>>> If the controller only returns 2k+32B, then perform a random read to
> >>>>> just move the read pointer to mtd->size + mtd->oobsize - 32 and
> >>>>> retrieve the missing 32 bytes? =20
> >>>>
> >>>> 1) raw read can read out the whole page data 2k+64B, decided by the =
len in the controller raw read command:
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0cmd =3D (len & GENMASK(5, 0)) | scrambler | =
DMA_DIR(dir);
> >>>> after that, the missing oob bytes(not used) can be copied from meson=
_chip->data_buf. so the implementation of meson_nfc_read_page_raw() is like=
 this if need.
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0{
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ......
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 meson_nfc_read_page_sub(n=
and, page, 1);
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 meson_nfc_get_data_oob(na=
nd, buf, oob_buf);
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 oob_len =3D (nand->ecc.by=
tes + 2) * nand->ecc.steps;
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 memcpy(oob_buf + oob_len,=
 meson_chip->data_buf + oob_len, mtd->oobsize - oob_len);
> >>>>
> >>>> =C2=A0=C2=A0=C2=A0=C2=A0}
> >>>> 2) In ECC mode, the controller can't bring back the missing OOB byte=
s. it can read out the user bytes and ecc bytes per meson_ooblayout_ops def=
ine. =20
> >>>
> >>> And then (if oob_required) you can bring the missing bytes with
> >>> something along:
> >>> nand_change_read_column_op(chip, mtd->writesize + oob_len,
> >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=
=A0=C2=A0=C2=A0 oob_buf + oob_len,
> >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=
=A0=C2=A0=C2=A0 mtd->oobsize - oob_len,
> >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=
=A0=C2=A0=C2=A0 false);
> >>> Should not be a huge performance hit. =20
> >>
> >> After finishing ECC mode reading, the column address internal in NAND =
device should be the right pos; it doesn't need to change the column again.=
 so adding controller raw read for the missing bytes after ECC reading may =
works.
> >> =20
> > use raw read for the missing bytes, but they are not protected by host =
ECC. to the NAND type of storage, is it ok or missing bytes better to be fi=
lled with 0xff? =20
>=20
> IIUC Miqu=C3=A8l's reply, valid behaviour is to return full OOB data in b=
oth modes. So in:
> ECC mode it is user bytes(corrected by ECC, read from info buffer) + ECC =
+ missing bytes. ECC and missing bytes read in RAW mode.

I believe the ECC bytes you'll get will be corrected.
You can check this by using nandflipbits in mtd-utils.

> RAW mode it is user bytes(not corrected by ECC) + ECC + missing bytes
>=20
>=20
> Also @Liang, is this valid code (drivers/mtd/nand/raw/meson_nand.c)?
>=20
> 	ret =3D nand_check_erased_ecc_chunk(data, ecc->size,
>                                           oob, ecc->bytes + 2,
>                                           NULL, 0,     =20
>                                           ecc->strength);
>=20
> It confused me, because 'oob' buffer contains both user bytes and ECC cod=
e,
> 'ecc->bytes + 2' is 16. May be it should be:
>=20
> 	ret =3D nand_check_erased_ecc_chunk(data, ecc->size,
>                                           oob + 2, ecc->bytes,
>                                           NULL, 0,     =20
>                                           ecc->strength);

When you check for an erased chunk you should probably check the whole
OOB area.

>=20
> For example let's look on Tegra's driver (drivers/mtd/nand/raw/tegra_nand=
.c):
>=20
>                         u8 *oob =3D chip->oob_poi + nand->ecc.offset +   =
=20
>                                   (chip->ecc.bytes * bit);             =20
>                                                                        =20
>                         ret =3D nand_check_erased_ecc_chunk(data, chip->e=
cc.size,
>                                                           oob, chip->ecc.=
bytes,
>                                                           NULL, 0,     =20
>                                                           chip->ecc.stren=
gth);
>=20
> 'oob' contains 'nand->ecc.offset', and ECC length does not account user b=
ytes length
> (e.g. 2) - it is just 'chip->ecc.bytes'

I haven't looked carefully, but be aware there are two user bytes
reserved at the beginning of the OOB area for marking bad blocks. There
*may* be a confusion somewhere. I am not saying there is one, just a
hint if you can't find an explanation.


Thanks,
Miqu=C3=A8l