Re: VFS 64-bit clean?

Albert D. Cahalan (acahalan@cs.uml.edu)
Tue, 17 Feb 1998 11:21:58 -0500 (EST)

>>>> So you claim that:
>>>>
>>>> 1. you have more than a terabyte in a single file
>>>> 2. that file is on a 32-bit machine
>>>> 3. that terabyte was cheap compared to the staff
>>>>
>>>> Really, I find it hard to believe that you satisfy all 3 conditions.
>>>> If so, I'd suggest an Alpha and some cheaper staff.
>>>>
>>>> The "limit" is 72 bits on 64-bit hardware, which we will all
>>>> have by the time a terabyte is limiting. Even Intel will have
>>>> 64-bit hardware in a few years.
>>>
>>> To test mkXXfs I do the following:
>>>
>>> % seekwrite [filesize] > testfile
>>> % ls -ls testfile
>>> % mkXXfs testfile
>>>
>>> Currently I'm limited to 2G tests, while the mkXXfs program claims to
>>> be able to go much further. (I just realized that I can emulate my
>>> seekwrite program using "dd":
>>> dd if=/dev/zero of=testfile bs=1024 count=1 seek=[filesize_in_k]
>>
>> Yes? That is certainly less than a terabyte, which is the "limit"
>> that I proposed. I don't see any problem for you.
>
> Well, I currently cannot go beyond 2G. That's a REAL limit nowadays,
> I'd want to be able to test upto a terabyte NOW.
>
> Harddisks become cheaper/larger at about a factor of two per two years
> (faster the last two years or so, but this figure is generally stable
> for longer periods). That means that you consume half a bit per year.
> A friend of mine has a 12G disk. 34 bits. If you increase the limit to
> 40 bits, we'll hit the new limit in at most 12 years. That's the
> short-sightedness that makes DOS such an awfully awkward operating
> system.

It is not short-sightedness. The limit disappears on 64-bit hardware,
which will happen long before those 12 years are past. In 12 years,
the $300 PC will be 64-bit.

Performance _today_ really matters. Since terabyte-size files are
large enough and can be fast, signed sector numbers are good.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu