> Please let me summarize... Thanks for the summary, Geert! > Solutions under consideration: > 1. Wolfram posted a patch to make i2c-sh_mobile fall back to PIO, > and retry DMA initialization in every request, so it will switch to > DMA when it becomes available. But this is suboptimal, as it adds > overhead to every request (and DMA may never become available in > case 2). Still, I'd think I should repost my patches with your comments addressed. It does add a bit of overhead IF the dmaengine core is compiled in AND the driver for the DMA hardware is not. Well, yeah. On the other hand, it fixes the regression that the driver is not even loaded in that case (because it currently will be deferred endlessly). What do you think? All the best, Wolfram
Attachment:
signature.asc
Description: Digital signature
