On Tue, Jan 20, 2004 at 06:52:40PM +0100, Dimitri Torfs wrote: > On Tue, Jan 20, 2004 at 09:19:22AM -0800, Greg KH wrote: > > On Tue, Jan 20, 2004 at 11:10:29AM +0100, Dimitri Torfs wrote: > > > I ask this question because there are places in the code where this > > > requirement clearly is not met. Is it because people are not aware > > > of this requirement or because they expect that cache-effects will > > > not come into play ... (or because they don't run on a > > > cache-incoherent platform) ? > > > > Where are those places? And yes, it's probably because we all run on > > systems where this is not noticed :) > > > > The first one I encountered is in the usb_new_device() function in usb.c > where the first 8 bytes of the descriptor are fetched: > > err = usb_get_descriptor(dev, USB_DT_DEVICE, 0, &dev->descriptor, 8);
But isn't that data (descriptor) created with a call to kmalloc()? As long as it isn't on the stack it should be ok, correct? > I'm using some usb net devices and I'm pretty sure they also "violate" > the alignment requirement, will check. Please do, and let us know if you find any problems. thanks, greg k-h ------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
