Re: [PATCH v2 4/4] media: pxa_camera: conversion to dmaengine

2015-07-26 Thread Robert Jarzmik
Guennadi Liakhovetski writes: > Hi Robert, > > On Sun, 19 Jul 2015, Robert Jarzmik wrote: > In principle, yes, it doesn't look all that horrible to me. You first > split the global SG list into up to 3 per-channel ones, then you > initialise your channels, what's wrong with that? Just have to a

Re: [PATCH v2 4/4] media: pxa_camera: conversion to dmaengine

2015-07-26 Thread Guennadi Liakhovetski
Hi Robert, On Sun, 19 Jul 2015, Robert Jarzmik wrote: > Robert Jarzmik writes: > > > Guennadi Liakhovetski writes: > > > >>> /* init DMA for Y channel */ > >> > >> How about taking the loop over the sg list out of pxa_init_dma_channel() > >> to avoid having to iterate it from the be

Re: [PATCH v2 4/4] media: pxa_camera: conversion to dmaengine

2015-07-18 Thread Robert Jarzmik
Robert Jarzmik writes: > Guennadi Liakhovetski writes: > >>> /* init DMA for Y channel */ >> >> How about taking the loop over the sg list out of pxa_init_dma_channel() >> to avoid having to iterate it from the beginning each time? Then you would >> be able to split it into channel

Re: [PATCH v2 4/4] media: pxa_camera: conversion to dmaengine

2015-07-12 Thread Robert Jarzmik
Guennadi Liakhovetski writes: >> /* init DMA for Y channel */ > > How about taking the loop over the sg list out of pxa_init_dma_channel() > to avoid having to iterate it from the beginning each time? Then you would > be able to split it into channels inside that global loop? Would

Re: [PATCH v2 4/4] media: pxa_camera: conversion to dmaengine

2015-07-12 Thread Guennadi Liakhovetski
Hi Robert, On Sun, 5 Jul 2015, Robert Jarzmik wrote: > Convert pxa_camera to dmaengine. This removes all DMA registers > manipulation in favor of the more generic dmaengine API. > > The functional level should be the same as before. The biggest change is > in the videobuf_sg_splice() function, w