Alex Buell writes:
> On Sat, 19 Feb 2000, Matti Aarnio wrote:
>
>> The VFS layer carries only time_t object for files, so even
>> if FS supplies finer granularity, you can't get it thru the VFS..
>
> Hmm. That seems to be a problem, as part of our work, we get a lot of
> files through ftp. We then need to sort them in exact time order, which
> means accuracy down to the microsecond is required as the ftp transfer
> runs so quickly that files gets saved with the same time_t value. Doesn't
> matter, I'll do an extra sort on the filenames anyway for the same time
> range.
>
> We could do with a much more precise time_t structure for the future,
> methinks.
The code will be ugly. It is better to push bloat out into libc
than to have it in the kernel.
How about 64-bit nanoseconds? This fixes the year-2038 problem too.
(this is for the VFS and new filesystems)
For ext2, maybe milliseconds or 1/1024 second units would be good.
That fits the 3 normal time stamps into one __u32, of which we have
very few to spare.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Feb 23 2000 - 21:00:25 EST