Greg KH wrote: > On Mon, May 13, 2002 at 11:17:30AM -0500, David Stroupe wrote: > >>I have noticed that if the passed stuff (DMA) size approaches a page >>(0x1000) in size I start to see the problem....is there a size limit >>involved here? I do see references to page faults in any oops that may >>occur if this call crashes. >> > > Yes, I think you are limited to a page in size here. > > If you want to send that much data across the user<->kernel boundry, > there are better ways of doing it than this (I think you use vmalloc, > but the mm people know this much better than I do.) > > Or just do a write() on the /dev node with the data. Or my personal > favorite, create a filesystem, and for every different ioctl that you > would have created, create a different file entry. > > Thanks for the suggestions....this is a port from win32 and since I never need 512 pages of memeory (the windows #define) I can reduce the struct size to 400 bytes or so...man I wish I had know about that limit cos I sure wasted a lot of hours trying to debug it:) Thanks very much for helping make sense from chaos:) Best regards, David -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
