Dear Werner, I'm sorry to trouble you again.
>> Is s3cmci.c in drivers/mmc/host/ in kernel 2.6.29 a good reference? > > It's quite readable, yes. What's missing in s3cmci is handling of > the SDIO interrupt, so we fall back to polling the interrupt in > software, which is a bit inefficient and probably wastes power. I generally went throught the code of s3cmci in andy-tracking branch. About why s3cmci doesn't handle the SDIO interrupt , i did a lot of google , and found something you talked before, it seems that s3c24xx platform(include s3c2442b) have some issues on handling the interrupt signal of SDIO, is that right? As you said, you choose to fall back to polling the interrupt in software, however, i cound not find the relevant code on polling and handling the sdio interrupt, can you point out them to me ? BTW, it seems that in s3cmci.c , the DMA enable indicator , dodma, is alway set as 0 durint the whole driver process. Does s3cmci in andy-tracking use DMA operation to do data transfer? Thanks a lot. Regards Dongas
