Messages in this thread | | | From | David Brownell <> | Subject | Re: [patch 2.6.26] /dev/hpet - fixes and cleanup | Date | Fri, 25 Jul 2008 12:55:35 -0700 |
| |
On Thursday 24 July 2008, Clemens Ladisch wrote: > David Brownell wrote:
> > > This seems to be the only part of the userspace interface that is > > > used in practice. Because of the availability of POSIX timers, it might > > > make sense to deprecate the HPET ioctl interface. > > > > I'll leave that part up to someone else. If POSIX timers > > are a sufficient userspace interface, great ... then that > > mmap son't really be needed either! > > The idea is to be able to get a high-precision timer value without doing > a syscall. (Whether the syscall overhead actually matters in a specific > application is another question.)
On x86_64 the vsyscall stuff should kick in for gettimeofday() when HPET is in use, eliminating the need for /dev/hpet mmapping.
I don't really know vsyscalls ... but if that could be done on i386 systems, the last argument for /dev/hpet would seem gone...
| |