Re: [PATCH 01/23] mtd: spi-nor: Stop prefixing generic functions with a manufacturer name

From: <Tudor.Ambarus@microchip.com>
To: <linux-mtd@lists.infradead.org>
Cc: <boris.brezillon@collabora.com>, <kstewart@linuxfoundation.org>,
	<alexandre.belloni@bootlin.com>, <vigneshr@ti.com>,
	<linux-aspeed@lists.ozlabs.org>, <thor.thayer@linux.intel.com>,
	<jethro@fortanix.com>, <rfontana@redhat.com>,
	<miquel.raynal@bootlin.com>, <opensource@jilayne.com>,
	<richard@nod.at>, <allison@lohutok.net>,
	<michal.simek@xilinx.com>, <Ludovic.Desroches@microchip.com>,
	<joel@jms.id.au>, <nishkadg.linux@gmail.com>,
	<john.garry@huawei.com>, <vz@mleia.com>,
	<alexander.sverdlin@nokia.com>, <matthias.bgg@gmail.com>,
	<tglx@linutronix.de>, <swboyd@chromium.org>,
	<mika.westerberg@linux.intel.com>, <ludovic.barre@st.com>,
	<linux-arm-kernel@lists.infradead.org>, <bbrezillon@kernel.org>,
	<andrew@aj.id.au>, <Nicolas.Ferre@microchip.com>,
	<linux-kernel@vger.kernel.org>, <dinguyen@kernel.org>,
	<michael@walle.cc>, <linux-mediatek@lists.infradead.org>,
	<info@metux.net>
Subject: Re: [PATCH 01/23] mtd: spi-nor: Stop prefixing generic functions with a manufacturer name
Date: Fri, 13 Mar 2020 15:50:10 +0000	[thread overview]
Message-ID: <16008638.kvcHKcsGDv@192.168.1.3> (raw)
In-Reply-To: <2838624.3XVpXx8FI0@192.168.1.3>

On Friday, March 13, 2020 4:30:48 PM EET Tudor.Ambarus@microchip.com wrote:
> > > > - * macronix_set_4byte() - Set 4-byte address mode for Macronix
> > > > flashes.
> > > > + * spi_nor_en4_ex4_set_4byte() - Enter/Exit 4-byte mode for Macronix
> > > > like
> > > > + * flashes.
> > > > 
> > > > * @nor:   pointer to 'struct spi_nor'.
> > > > * @enable:        true to enter the 4-byte address mode, false to exit
> > > > the 4-byte *         address mode.
> > > > *
> > > > * Return: 0 on success, -errno otherwise.
> > > > */
> > > > 
> > > > -static int macronix_set_4byte(struct spi_nor *nor, bool enable)
> > > > +static int spi_nor_en4_ex4_set_4byte(struct spi_nor *nor, bool
> > > > enable)
> > > 
> > > Sounds a bit weird, how about simplifying this to:
> > > spi_nor_set_4byte_addr_mode()
> > > 
> > > Or if you want to be specific:
> > > spi_nor_en_ex_4byte_addr_mode()
> > 
> > You're right. Maybe we can simplify things by having a single function
> > that does optional steps based on new flags
> > 
> > SPI_NOR_EN_EX_4B_NEEDS_WEN
> > SPI_NOR_CLEAR_EAR_ON_4B_EXIT
> > 
> > This should probably be done in a separate patch though, so ack on the
> > spi_nor_en_ex_4byte_addr_mode() rename, assuming we also change the
> > bool argument name to enter.
> 
> A single big function will be ugly to handle because of the
> spansion_set_4byte() -> it uses a different opcode.
> 
> I like the nor->params>set_4byte hook.
> 
> I think that spi_nor_en4_ex4_set_4byte() can be renamed to
> spi_nor_set_4byte() and be the only set_4byte() method exposed to
> manufacturers.
> spansion_set_4byte() will be static in core.c and the rest will be private
> in the manufacturer drivers.
> 
> Here's how the manufacturers enter and exit the 4 byte mode:
> -> eon, gidadevice, issi, macronix, xmc use EN4B/EX4B
> -> micron-st needs WEN -> private method as they are the only ones that
> require this
> -> newer spansion have a 4BAM opcode (new, public command), older spansion
> flashes use BRWR (legacy in core.c, spansion_set_4byte())
> -> winbond set_4byte method is hackish and may be reason for just a flash
> fixup hook -> private method

I'll drop the set_4byte changes from this patch. I'm adding a dedicated patch 
immediately after this, to implement what I said above.

Cheers,
ta