From: marex@denx.de (Marek Vasut) Date: Mon, 24 Aug 2015 15:15:44 +0200 Subject: [PATCH linux-next v4 5/5] mtd: atmel-quadspi: add driver for Atmel QSPI controller In-Reply-To: <20150824124924.GZ7557@n2100.arm.linux.org.uk> References: <201508241303.52066.marex@denx.de> <20150824124924.GZ7557@n2100.arm.linux.org.uk> Message-ID: <201508241515.45036.marex@denx.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Monday, August 24, 2015 at 02:49:24 PM, Russell King - ARM Linux wrote: Hi Russell, > On Mon, Aug 24, 2015 at 01:03:51PM +0200, Marek Vasut wrote: > > These are functions, not macros :) > > > > btw is there any reason for these ? I'd say, just put the read*() and > > write*() functions directly into the code and be done with it, it is > > much less confusing. > > > > Also, why do you use the _relaxed() versions of the functions ? > > Now that the _relaxed() accessors are available throughout the kernel, > everyone should be using the _relaxed() versions unless they need the > properties of the non-relaxed versions. You mean the memory barrier, right ? > Remember that the non-relaxed > versions are rather expensive on ARM due to the need to go all the way > out to the L2 cache - it at least doubles the number of accesses for > every read*/write*(). I think in case of this driver, we don't need the non-relaxed version anywhere, right ? Thanks for the educational writeup :) btw. is [1] still the current study material on the I/O accessor best practices please ? [1] http://permalink.gmane.org/gmane.linux.ports.arm.kernel/117644 Best regards, Marek Vasut