[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6bc0cf7ea58fae2e6a3765f673a777788d465984.camel@maquefel.me>
Date: Fri, 08 Dec 2023 14:24:52 +0300
From: Nikita Shubin <nikita.shubin@...uefel.me>
To: Andy Shevchenko <andriy.shevchenko@...el.com>
Cc: Miquel Raynal <miquel.raynal@...tlin.com>,
Richard Weinberger <richard@....at>,
Vignesh Raghavendra <vigneshr@...com>,
Arnd Bergmann <arnd@...db.de>,
Linus Walleij <linus.walleij@...aro.org>,
Neil Armstrong <neil.armstrong@...aro.org>,
Daniel Golle <daniel@...rotopia.org>,
Jean Delvare <jdelvare@...e.de>,
Christophe Kerello <christophe.kerello@...s.st.com>,
Chris Packham <chris.packham@...iedtelesis.co.nz>,
Johan Jonker <jbx6244@...il.com>, linux-kernel@...r.kernel.org,
linux-mtd@...ts.infradead.org,
Alexander Sverdlin <alexander.sverdlin@...il.com>
Subject: Re: [PATCH v5 19/39] mtd: rawnand: add support for ts72xx
Hello Andy!
On Wed, 2023-11-22 at 14:24 +0200, Andy Shevchenko wrote:
> On Wed, Nov 22, 2023 at 11:59:57AM +0300, Nikita Shubin wrote:
> > Technologic Systems has it's own nand controller implementation in
> > CPLD.
>
> ...
>
> > +static int ts72xx_nand_attach_chip(struct nand_chip *chip)
> > +{
> > + switch (chip->ecc.engine_type) {
> > + case NAND_ECC_ENGINE_TYPE_SOFT:
> > + if (chip->ecc.algo == NAND_ECC_ALGO_UNKNOWN)
> > + chip->ecc.algo = NAND_ECC_ALGO_HAMMING;
> > + chip->ecc.algo = NAND_ECC_ALGO_HAMMING;
> > + break;
> > + case NAND_ECC_ENGINE_TYPE_ON_HOST:
> > + return -EINVAL;
>
> > + default:
> > + break;
> > + }
> > +
> > + return 0;
>
> Move this to default.
>
> > +}
>
> ...
>
> > + for (i = 0; i < instr->ctx.addr.naddrs; i++)
> > + iowrite8(instr->ctx.addr.addrs[i], data-
> > >base);
>
> iowrite8_rep() ?
>
> > + case NAND_OP_DATA_IN_INSTR:
> > + ioread8_rep(data->base, instr->ctx.data.buf.in,
> > instr->ctx.data.len);
>
> Hehe, you are even using it...
>
> ...
>
> > + if (instr->delay_ns)
>
> What will happen if you drop this check?
>
> > + ndelay(instr->delay_ns);
No idea! I was asked to keep it by Miquel:
https://lore.kernel.org/lkml/8bbe66a23eb5c8a2404b72d754b1bcb6f4d23867.camel@maquefel.me/T/
But it looks some one can at least still do some calculation like:
https://elixir.bootlin.com/linux/latest/source/include/linux/delay.h#L50
At worst try to divide something.
Do you think ndelay(instr->delay_ns) is safe enough ?
>
> ...
>
> > + int ret;
> > +
> > + ret = mtd_device_unregister(nand_to_mtd(chip));
> > + WARN_ON(ret);
>
> Is this a requirement by MTD to have return value being checked?
>
Aswered by Miquel.
Powered by blists - more mailing lists