From mboxrd@z Thu Jan 1 00:00:00 1970 From: linux@arm.linux.org.uk (Russell King - ARM Linux) Date: Sun, 31 Jul 2011 10:06:19 +0100 Subject: [PATCH 04/18] dmaengine/amba-pl08x: Remove redundant comment and rewrite original In-Reply-To: References: <1311941355.1536.529.camel@vkoul-udesk3> <20110730225708.GB17570@n2100.arm.linux.org.uk> Message-ID: <20110731090619.GA30888@n2100.arm.linux.org.uk> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Sun, Jul 31, 2011 at 01:37:46AM +0200, Linus Walleij wrote: > 2011/7/31 Russell King - ARM Linux : > > On Sun, Jul 31, 2011 at 12:32:55AM +0200, Linus Walleij wrote: > >> The PL08x has two bus mastering interfaces. They can be > >> connected to two separate busses and two adress spaces > >> even, usually that is not the case though. > > > > This is not what the function is about... that's done by my > > pl08x_select_bus() function. > > To be clear, I was explaining what I percieved as a > question from Viresh about what the comment in this hunk of > the patch was about: Yes, and the comment and function name are wrong for what it's doing. I repeat: it has nothing to do with selecting an AHB bus. > As for the LLI code, all of it's weird complexity comes from the fact > that the DMAC cannot let a single element pass across 1KB (0x400) > boundaries, which is one more than a little insane hardware > restriction, and I have no clue where that limitation actually comes > from, likely strange VHDL code. Viresh's discussion with ARM Ltd indicates that this is not the case, and that restriction comes from misunderstanding of the documentation.