From: Ned F. <nfo...@wh...> - 2008-02-27 13:25:13
|
Harry J Mason wrote: > On Tue, 26 Feb 2008, Ned Forrester wrote: > >> Unfortunately for me, the pxa2xx-spi driver already has a significant update >> in 2.6.25, so I will have to wait a while longer to base my version on the >> mainline version. > > Do you know if any of the 2.6.25 fixes would affect a DMA underrun I have > been experiencing (interfacing to an MMC card)? No. The large recent changes are structural changes that adapt the driver to changes going on elsewhere in the kernel. There are no recent functional changes other than a small bug fix relating to the sequence of setting modes and chip select. How do you suppose you are generating a TX FIFO underrun when using DMA? The entire transmit buffer must exist in memory before starting the DMA, so this could only happen if the DMA does not service the TX FIFO in time. I expect that you may not have set the burst size, and thresholds optimally. If you post your clock rate, burst, threshold and other parameters, I will see if I see anything wrong. I assume you are using the version of pxa2xx_spi.c that is distributed with 2.6.20 or later. Earlier versions certainly might have this problem. -- Ned Forrester nfo...@wh... Oceanographic Systems Lab 508-289-2226 Applied Ocean Physics and Engineering Dept. Woods Hole Oceanographic Institution Woods Hole, MA 02543, USA http://www.whoi.edu/sbl/liteSite.do?litesiteid=7212 http://www.whoi.edu/hpb/Site.do?id=1532 http://www.whoi.edu/page.do?pid=10079 |