OMAP baseline test results for v3.7-rc3

Hiremath, Vaibhav hvaibhav at ti.com
Tue Nov 6 03:07:47 EST 2012


On Tue, Nov 06, 2012 at 13:28:00, Balbi, Felipe wrote:
> Hi,
> 
> On Tue, Nov 06, 2012 at 07:17:19AM +0100, Hiremath, Vaibhav wrote:
> > On Wed, Oct 31, 2012 at 00:21:02, Balbi, Felipe wrote:
> > > Hi,
> > > 
> > > On Tue, Oct 30, 2012 at 10:58:59AM -0700, Tony Lindgren wrote:
> > > > * Felipe Balbi <balbi at ti.com> [121030 10:34]:
> > > > > Hi,
> > > > > 
> > > > > On Tue, Oct 30, 2012 at 09:27:28AM -0700, Tony Lindgren wrote:
> > > > > > * Vaibhav Hiremath <hvaibhav at ti.com> [121030 07:50]:
> > > > > > > > 
> > > > > > > > MMC is dependent on EDMA-DMA conversion patches from Matt, which he has 
> > > > > > > > already submitted to the list recently. So MMC support will come along with
> > > > > > > > EDMA support. DMA-EDMA patches are targeted for v3.8, lets see how it goes.
> > > > > > 
> > > > > > This is a bogus dependency, the MMC driver needs to also work
> > > > > > without DMA.
> > > > > 
> > > > > heh, too bad driver errors out when it doesn't find DMA channels :-)
> > > > 
> > > > It should just print a warning instead and continue.
> > > >  
> > > > > 1869         host->rx_chan = dma_request_channel(mask, omap_dma_filter_fn, &rx_req);
> > > > > 1870         if (!host->rx_chan) {
> > > > > 1871                 dev_err(mmc_dev(host->mmc), "unable to obtain RX DMA engine channel %u\n", rx_req);
> > > > > 1872                 ret = -ENXIO;
> > > > > 1873                 goto err_irq;
> > > > > 1874         }
> > > > > 1875 
> > > > > 1876         host->tx_chan = dma_request_channel(mask, omap_dma_filter_fn, &tx_req);
> > > > > 1877         if (!host->tx_chan) {
> > > > > 1878                 dev_err(mmc_dev(host->mmc), "unable to obtain TX DMA engine channel %u\n", tx_req);
> > > > > 1879                 ret = -ENXIO;
> > > > > 1880                 goto err_irq;
> > > > > 1881         }
> > > > > 
> > > > > in fact, if DMAENGINE isn't enabled, this won't even compile due to
> > > > > omap_dma_filter_fn() right ?
> > > > 
> > > > It should, CONFIG_DMADEVICES is optional. If it does not compile,
> > > > then there's a bug somewhere.
> > > 
> > > you're right, there's a static inline nop for when we don't have omap
> > > dma engine driver compiled.
> > > 
> > > nevermind.
> > > 
> > 
> > Did anybody tried polling mode MMC support on OMAP devices in the
> > past?
> 
> why are you trying out polling when we have a working interrupt mode ?
> 
The thread started with need of MMC support without EDMA.
So when I say polling, I would like to try MMC without DMA support, not to 
put dependency on EDMA-DMA engine migration patches.

Thanks,
Vaibhav

> -- 
> balbi
> 




More information about the linux-arm-kernel mailing list