Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1525708
| From | Matthew Wilcox <mawilcox@microsoft.com> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | RE: [PATCH 20/29] radix tree: Improve multiorder iterators |
| Date | 2016-11-18 22:00 +0100 |
| Message-ID | <sEYql-3F7-5@gated-at.bofh.it> (permalink) |
| References | <sEgSl-rg-1@gated-at.bofh.it> <sEh22-uC-39@gated-at.bofh.it> <sEPQ5-6yx-5@gated-at.bofh.it> <sEUmJ-18a-21@gated-at.bofh.it> <sEVCa-1Ow-7@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
From: Konstantin Khlebnikov [mailto:koct9i@gmail.com]
> On Fri, Nov 18, 2016 at 7:31 PM, Matthew Wilcox <mawilcox@microsoft.com>
> wrote:
> > I think what you're suggesting is that we introduce a new API:
> >
> > slot = radix_tree_iter_save(&iter, order);
> >
> > where the caller tells us the order of the entry it just consumed. Or maybe
> you're suggesting
> >
> > slot = radix_tree_iter_advance(&iter, newindex)
>
> Yes, someting like that.
>
> >
> > which would allow us to skip to any index. Although ... isn't that just
> radix_tree_iter_init()?
>
> Iterator could keep pointer to current node and reuse it for next
> iteration if possible.
The point of this API is that it's never possible, because we're about to drop the lock and allow other users to modify the tree. Actually, it is different from radix_tree_iter_init() because it has to set ->tags to 0 and ->index == ->next_index in order to get through a call to radix_tree_next_slot().
> > It does push a bit of complexity onto the callers. We have 7 callers of
> > radix_tree_iter_next() in my current tree (after applying this patch set, so
> > range_tag_if_tagged and locate_item have been pushed into their callers):
> > btrfs, kugepaged, page-writeback and shmem. btrfs knows its objects occupy
> > one slot. khugepaged knows that its page is order 0 at the time it calls
> > radix_tree_iter_next(). Page-writeback has a struct page and can simply use
> > compound_order(). It's shmem where things get sticky, although it's all
> > solvable with some temporary variables.
>
> Users who work only with single slot enties don't get any complications,
> all other already manage these multiorder entries somehow.
If you read the patch below, you'll see that most callers don't need to be concerned with the size of the entry they're looking at. I'll trim away the trivial ones so it's easier to see my point.
It's not a huge amount of code in each caller, but is this a burden we really want to push onto the callers when we could handle it behind the interface?
diff --git a/mm/shmem.c b/mm/shmem.c
index 8f9c9aa..90dd18d9 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -658,7 +658,10 @@ unsigned long shmem_partial_swap_usage(struct address_space *mapping,
swapped++;
if (need_resched()) {
- slot = radix_tree_iter_next(slot, &iter);
+ unsigned int order = 0;
+ if (!radix_tree_exceptional_entry(page))
+ order = compound_order(page);
+ slot = radix_tree_iter_save(&iter, order);
cond_resched_rcu();
}
}
@@ -2450,6 +2453,7 @@ static void shmem_tag_pins(struct address_space *mapping)
slot = radix_tree_iter_retry(&iter);
continue;
}
+ page = NULL;
} else if (page_count(page) - page_mapcount(page) > 1) {
spin_lock_irq(&mapping->tree_lock);
radix_tree_tag_set(&mapping->page_tree, iter.index,
@@ -2458,7 +2462,8 @@ static void shmem_tag_pins(struct address_space *mapping)
}
if (need_resched()) {
- slot = radix_tree_iter_next(slot, &iter);
+ unsigned int order = page ? compound_order(page) : 0;
+ slot = radix_tree_iter_save(&iter, order);
cond_resched_rcu();
}
}
@@ -2528,7 +2533,10 @@ static int shmem_wait_for_pins(struct address_space *mapping)
spin_unlock_irq(&mapping->tree_lock);
continue_resched:
if (need_resched()) {
- slot = radix_tree_iter_next(slot, &iter);
+ unsigned int order = 0;
+ if (page)
+ order = compound_order(page);
+ slot = radix_tree_iter_save(&iter, order);
cond_resched_rcu();
}
}
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH 00/29] Improve radix tree for 4.10 Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 06/29] radix tree test suite: benchmark for iterator Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 07/29] radix tree test suite: Use rcu_barrier Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 24/29] tpm: Use idr_find(), not idr_find_slowpath() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 01/29] tools: Add WARN_ON_ONCE Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 06/29] radix tree test suite: benchmark for iterator Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 20/29] radix tree: Improve multiorder iterators Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 12/29] radix-tree: Add radix_tree_split_preload() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 19/29] radix tree test suite: iteration test misuses RCU Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 18/29] btrfs: Fix race in btrfs_free_dummy_fs_info() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 10/29] radix-tree: Add radix_tree_join Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 23/29] idr: Add ida_is_empty Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 22/29] radix-tree: Delete radix_tree_range_tag_if_tagged() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 25/29] rxrpc: Abstract away knowledge of IDR internals Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 05/29] radix tree test suite: Make runs more reproducible Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 14/29] radix-tree: Move rcu_head into a union with private_list Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 15/29] radix-tree: Create node_tag_set() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 27/29] radix tree test suite: Add some more functionality Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:30 +0100
[PATCH 08/29] tools: Add more bitmap functions Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 21/29] radix-tree: Delete radix_tree_locate_item() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
Re: [PATCH 21/29] radix-tree: Delete radix_tree_locate_item() Konstantin Khlebnikov <koct9i@gmail.com> - 2016-11-18 13:00 +0100
RE: [PATCH 21/29] radix-tree: Delete radix_tree_locate_item() Matthew Wilcox <mawilcox@microsoft.com> - 2016-11-18 17:50 +0100
[PATCH 11/29] radix-tree: Add radix_tree_split Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 08/29] tools: Add more bitmap functions Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 07/29] radix tree test suite: Use rcu_barrier Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 28/29] radix-tree: Create all_tag_set Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 10/29] radix-tree: Add radix_tree_join Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 17/29] radix-tree: Improve dump output Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 14/29] radix-tree: Move rcu_head into a union with private_list Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 02/29] radix tree test suite: Allow GFP_ATOMIC allocations to fail Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 25/29] rxrpc: Abstract away knowledge of IDR internals Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 02/29] radix tree test suite: Allow GFP_ATOMIC allocations to fail Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 18/29] btrfs: Fix race in btrfs_free_dummy_fs_info() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 23/29] idr: Add ida_is_empty Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
Re: [PATCH 23/29] idr: Add ida_is_empty Konstantin Khlebnikov <koct9i@gmail.com> - 2016-11-18 13:00 +0100
[PATCH 16/29] radix-tree: Make radix_tree_find_next_bit more useful Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 20/29] radix tree: Improve multiorder iterators Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
Re: [PATCH 20/29] radix tree: Improve multiorder iterators Konstantin Khlebnikov <koct9i@gmail.com> - 2016-11-18 12:50 +0100
RE: [PATCH 20/29] radix tree: Improve multiorder iterators Matthew Wilcox <mawilcox@microsoft.com> - 2016-11-18 17:40 +0100
Re: [PATCH 20/29] radix tree: Improve multiorder iterators Konstantin Khlebnikov <koct9i@gmail.com> - 2016-11-18 19:00 +0100
RE: [PATCH 20/29] radix tree: Improve multiorder iterators Matthew Wilcox <mawilcox@microsoft.com> - 2016-11-18 22:00 +0100
[PATCH 17/29] radix-tree: Improve dump output Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 04/29] radix tree test suite: Free preallocated nodes Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 09/29] radix tree test suite: Use common find-bit code Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 22/29] radix-tree: Delete radix_tree_range_tag_if_tagged() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 13/29] radix-tree: Fix typo Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 24/29] tpm: Use idr_find(), not idr_find_slowpath() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 04/29] radix tree test suite: Free preallocated nodes Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 21/29] radix-tree: Delete radix_tree_locate_item() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 26/29] idr: Reduce the number of bits per level from 8 to 6 Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 26/29] idr: Reduce the number of bits per level from 8 to 6 Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:40 +0100
[PATCH 28/29] radix-tree: Create all_tag_set Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 16/29] radix-tree: Make radix_tree_find_next_bit more useful Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 05/29] radix tree test suite: Make runs more reproducible Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 03/29] radix tree test suite: Track preempt_count Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 09/29] radix tree test suite: Use common find-bit code Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 27/29] radix tree test suite: Add some more functionality Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 12/29] radix-tree: Add radix_tree_split_preload() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 01/29] tools: Add WARN_ON_ONCE Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 15/29] radix-tree: Create node_tag_set() Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 11/29] radix-tree: Add radix_tree_split Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
[PATCH 00/29] Improve radix tree for 4.10 Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
RE: [PATCH 00/29] Improve radix tree for 4.10 Matthew Wilcox <mawilcox@microsoft.com> - 2016-11-17 21:00 +0100
Re: [PATCH 00/29] Improve radix tree for 4.10 Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-11-17 23:20 +0100
RE: [PATCH 00/29] Improve radix tree for 4.10 Matthew Wilcox <mawilcox@microsoft.com> - 2016-11-18 05:40 +0100
[PATCH 19/29] radix tree test suite: iteration test misuses RCU Matthew Wilcox <mawilcox@linuxonhyperv.com> - 2016-11-16 23:50 +0100
csiph-web