Hi Gary, On Sat, Jan 28, 2012 at 21:38:34, Gary Thomas wrote: > On 2012-01-28 07:53, Sergei Shtylyov wrote: > > On 28.01.2012 15:16, Gary Thomas wrote: > > > >> The use of pdev->id as a channel select is problematic, at least > >> on the OMAP-L13x when using 4 bit hardware assisted ECC. On this > >> hardware, only EMiF channels (chip select) 2 through 5 are available. > >> Specifying platform device id 1 when using chip select 3 is obscure > >> at best. > > > >> If the core_chipsel value used by the driver is incorrect, then > >> the ECC hardware is not set up properly and ends up always reading > >> ECC values of zero. This is totally broken as no ECC errors will > >> be detected on reads and incorrect ECC values will be stored on writes. > > > >> The best way to fix this is to pass the actual chip select in > >> the platform data, rather than have it be inferred from the > >> platform device id. The attached patch provides for this. Also, > >> this patch is a no-op for target platforms that do not specify > >> the core_chipsel value in the platform data. This patch was > >> generated against the mainline 3.0 kernel, but applies to 3.2 > >> as well. > > > >> Note: there does not seem to be an identified maintainer of this > >> code, so this list was my first guess at where to post. > > > > Pelase send your patches inline, not as attachments. > >
As Sergei suggested, please modify your patch to convert spaces to tabs and remove trailing white-spaces. You can run your patch through the scripts/checkpatch.pl script which will detect such errors. Once done, please resubmit your patch inline. Thanks, Sudhakar _______________________________________________ Davinci-linux-open-source mailing list Davinci-linux-open-source@linux.davincidsp.com http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
