Daniel Phillips <[EMAIL PROTECTED]> writes:

> On Sunday 19 October 2008 22:09, OGAWA Hirofumi wrote:
>> >   - Buffer operations need to be adapted to work with page cache and
>> >     bio instead of userspace synchronous Posix IO.
>> 
>> For now, I can't image what we want... BTW, do we support the blocksize
>> smaller than PAGE_CACHE_SIZE?
>
> Of course, and that will require some extra fiddling with the page cache
> access code.
>
> By coincidence, Tux3 University has lately been concentrating on exactly
> the code we will be working with here.

Oh, is it talking about design too? Can I join to?

>> >   - Spinlocks for VFS ops
>> 
>> Um.. what does this mean?
>
> Spinlocks are needed all over the place to serialize VFs operations
> that would otherwise be racy with respect to Tux3 data structures.

Yes, e.g. the spinlock/mutex for btree? I thought it may be replacements
of inode->i_mutex or something.

>> In future, how are we going to use the userland? I worry we are bothered
>> to sync userland <-> kernel...
>
> This gives us a very good start on a userspace based fsck and repair
> tool.  There is no reason why we cannot include full access to the Tux3
> filesystem from our repair/analysis tool.
>
> I don't know if the fuse interface will have long term value, but it
> certainly has value now and uses largely the same code we need for
> userspace support tools, so it will not be much additional work to
> continue to support it.

I see.
-- 
OGAWA Hirofumi <[EMAIL PROTECTED]>

_______________________________________________
Tux3 mailing list
[email protected]
http://tux3.org/cgi-bin/mailman/listinfo/tux3

Reply via email to