One of my reasons for absorbing ratcache into my current stuff is
just that - to give it a serious beating. The fact that I found a
hitherto-undiscovered BUG() and a deadlock in the first 30 minutes
just shows what a mean beat I have :)
I haven't yet even looked at lib/rat.c, but based on testing, I
believe radix-tree pagecache is ready for 2.5. It would be good
if the other Christoph could check over the shmem.c changes.
As far as I know, the sole remaining "issue" is that block_flushpage()
is being called under spinlock. Well, there's nothing new here.
The kernel is *already* calling block_flushpage() under spinlock
it at least three places. But it just so happens that there are
never (?) any locked buffers against the page from those call sites.
So I don't see the block_flushpage() thing as a blocker for this
patch - it's just general ickiness which needs sorting out separately.
I looked at block_flushpage() a month or so back. I ended up
concluding that we should just create block_flushpage_atomic() and
make it go BUG() if the page has locked buffers. Then call the atomic
version from under spinlock, and leave it at that.
-
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/