From mboxrd@z Thu Jan 1 00:00:00 1970 From: viresh.kumar@st.com (viresh kumar) Date: Tue, 19 Apr 2011 12:16:43 +0530 Subject: [PATCH 2/6] dmaengine/dw_dmac: Enable resubmission from callback routine. In-Reply-To: <438BB0150E931F4B9CE701519A44630108519E1C1D@bgsmsx502.gar.corp.intel.com> References: <438BB0150E931F4B9CE701519A44630108519E1BE6@bgsmsx502.gar.corp.intel.com> <4DAD2A61.4020406@st.com> <438BB0150E931F4B9CE701519A44630108519E1C1D@bgsmsx502.gar.corp.intel.com> Message-ID: <4DAD2FD3.5010103@st.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 04/19/2011 12:03 PM, Koul, Vinod wrote: > On 04/19/2011 11:53 AM, viresh kumar wrote: >> On 04/19/2011 11:42 AM, Koul, Vinod wrote: >>> On 04/18/2011 04:20 PM, viresh kumar wrote: >> Vinod, >> >> I initiated this discussion few days back on a separate thread: >> >> http://www.spinics.net/lists/arm-kernel/msg121139.html >> >> And so implemented this. > Ahhh, on the road, catching up on email... > > This was a dma engine requirement, not absolute for slave-dma. > But would be good idea to resist this. What is your reason for this? Mainline drivers using dmaengine are already submitting new transfers from callback routines. For example: amba-pl011. And i have to use pl011 with DMA, so faced this issue in dw_dmac. -- viresh