Re: [PATCH resend] shdma: fixup sh_dmae_get_partial() calculation error

[Simon Horman <horms@verge.net.au>]

On Tue, Jul 30, 2013 at 11:22:31AM +0200, Guennadi Liakhovetski wrote:
> Hi Simon
> 
> On Tue, 30 Jul 2013, Simon Horman wrote:
> 
> > [ Cc Vinod ]
> > 
> > On Tue, Jul 23, 2013 at 11:12:41PM -0700, Kuninori Morimoto wrote:
> > > sh_desc->hw.tcr is controlling real data size,
> > > and, register TCR is controlling data transfer count
> > > which was xmit_shifted value of hw.tcr.
> > > Current sh_dmae_get_partial() is calculating in different unit.
> > > This patch fixes it.
> > > 
> > > Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
> > 
> > Can I confirm that this was a regression introduced by
> > 4f46f8ac80416b0e8fd3aba6a0d842205fb29140                   
> > ("dmaengine: shdma: restore partial transfer calculation") ?
> 
> No, don't think so. That patch only restores what was accidentally deleted 
> from the driver before. So, you cannot say, that before that patch the 
> calculation was correct. And even the way it restores it was already wrong 
> before. The error goes back to the original implementation:
> 
> commit c014906a870ce70e009def0c9d170ccabeb0be63
> Author: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
> Date:   Thu Feb 18 16:30:02 2010 +0000
> 
>     dmaengine: shdma: extend .device_terminate_all() to record partial transfer

Thanks, I will add the following to the changelog:

This bug has been present since c014906a870ce70e009def0c9d170ccabeb0be63
("dmaengine: shdma: extend .device_terminate_all() to record partial
transfer"), which was added in 2.6.34-rc1.