On Thu, May 17, 2012 at 8:55 PM, H. Peter Anvin <hpa@xxxxxxxxxx> wrote: > On 05/17/2012 08:49 PM, Linus Torvalds wrote: >> On Thu, May 17, 2012 at 8:39 PM, H.J. Lu <hjl.tools@xxxxxxxxx> wrote: >>> >>> That will be wrong. __BITS_PER_LONG defines # bits of long >>> as seen by kernel. We don't use it in user space. >> >> Yes you do. Exactly in that structure that Peter points to. *Exactly* >> because that structure uses "long" instead of some fixed size. Which >> will be different in user mode than in kernel mode. >> >> And if user mode doesn't use these headers at all, then we should stop >> playing the insane games. >> > > User mode can, and should, be able to use the exported headers. David > Howells have been doing even more work to distill out the actual > exported ABIs from the kernel and remove remaining chaff. > > That being said it seems kind of loopy to expect something called > __BITS_PER_LONG to be anything other than (CHAR_BIT*sizeof(long)), > especially since one of the main uses of it seems to be sizing > bitvectors (which has its own issues on bigendian machines because I > think we do littleendian bit numbering even on bigendian iron). > But __BITS_PER_LONG used in kernel header files really means "long" as seen by kernel, not by user space. 64-bit kernel can have 32-bit and 64-bit longs in use space. __BITS_PER_LONG is a bad name for user space. -- H.J. -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
