On 7/4/2018 11:16 AM, Andy Shevchenko wrote: > On Wed, Jul 4, 2018 at 8:13 PM, Logan Gunthorpe <logang@xxxxxxxxxxxx> wrote: >> On 7/4/2018 11:10 AM, Andy Shevchenko wrote: >>> We have an iDMA 32-bit hardware (see drivers/dma/dw/) which has an >>> extension 64-bit registers where only one of them has a specific bit >>> to "commit" the changes written to all of them. And by some very >>> unknown reason that bit is in lo part which automatically means we >>> must to write it last. >> >> And it supports both BE and LE? And in both cases it's the lo part? > > It's only LE for now. So the main question is if they were to add BE support, would they leave the trigger in the same address or swap it to the other address so that it's always the LO part that triggers? Otherwise it's hard to say what we really want for the BE variants of the non-atomic hi-lo operations. In the end, the driver author is free to use specifically which ever function is necessary in any given situation (lo-hi vs hi-lo) and they can read the definitions and no one is using them yet. So either way is probably just as valid and it's probably not really worth fussing too much about. Logan
