mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Liam R. Howlett" <Liam.Howlett@oracle.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: maple-tree@lists.infradead.org, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org,
	"Liam R. Howlett" <Liam.Howlett@oracle.com>
Subject: [PATCH 13/14] maple_tree: Refine mas_preallocate() node calculations
Date: Wed, 31 May 2023 22:16:04 -0400	[thread overview]
Message-ID: <20230601021605.2823123-14-Liam.Howlett@oracle.com> (raw)
In-Reply-To: <20230601021605.2823123-1-Liam.Howlett@oracle.com>

Calculate the number of nodes based on the pending write action instead
of assuming the worst case.

This addresses a performance regression introduced in platforms that
have longer allocation timing.

Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
---
 lib/maple_tree.c | 44 +++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 43 insertions(+), 1 deletion(-)

diff --git a/lib/maple_tree.c b/lib/maple_tree.c
index adf662bc413e..5ea211c3f186 100644
--- a/lib/maple_tree.c
+++ b/lib/maple_tree.c
@@ -5541,9 +5541,51 @@ EXPORT_SYMBOL_GPL(mas_store_prealloc);
  */
 int mas_preallocate(struct ma_state *mas, void *entry, gfp_t gfp)
 {
+	MA_WR_STATE(wr_mas, mas, entry);
+	unsigned char node_size;
+	int request = 1;
 	int ret;
 
-	mas_node_count_gfp(mas, 1 + mas_mt_height(mas) * 3, gfp);
+
+	if (unlikely(!mas->index && mas->last == ULONG_MAX))
+		goto ask_now;
+
+	mas_wr_store_setup(&wr_mas);
+	wr_mas.content = mas_start(mas);
+	/* Root expand */
+	if (unlikely(mas_is_none(mas) || mas_is_ptr(mas)))
+		goto ask_now;
+
+	if (unlikely(!mas_wr_walk(&wr_mas))) {
+		/* Spanning store, use worst case for now */
+		request = 1 + mas_mt_height(mas) * 3;
+		goto ask_now;
+	}
+
+	/* At this point, we are at the leaf node that needs to be altered. */
+	/* Exact fit, no nodes needed. */
+	if (wr_mas.r_min == mas->index && wr_mas.r_max == mas->last)
+		return 0;
+
+	mas_wr_end_piv(&wr_mas);
+	node_size = mas_wr_new_end(&wr_mas);
+	if (node_size >= mt_slots[wr_mas.type]) {
+		/* Split, worst case for now. */
+		request = 1 + mas_mt_height(mas) * 2;
+		goto ask_now;
+	}
+
+	/* New root needs a singe node */
+	if (unlikely(mte_is_root(mas->node)))
+		goto ask_now;
+
+	/* Potential spanning rebalance collapsing a node, use worst-case */
+	if (node_size  - 1 <= mt_min_slots[wr_mas.type])
+		request = mas_mt_height(mas) * 2 - 1;
+
+	/* node store, slot store needs one node */
+ask_now:
+	mas_node_count_gfp(mas, request, gfp);
 	mas->mas_flags |= MA_STATE_PREALLOC;
 	if (likely(!mas_is_err(mas)))
 		return 0;
-- 
2.39.2


  parent reply	other threads:[~2023-06-01  2:18 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-06-01  2:15 [PATCH 00/14] Reduce preallocations for maple tree Liam R. Howlett
2023-06-01  2:15 ` [PATCH 01/14] maple_tree: Add benchmarking for mas_for_each Liam R. Howlett
2023-06-01  2:15 ` [PATCH 02/14] maple_tree: Add benchmarking for mas_prev() Liam R. Howlett
2023-06-01  2:15 ` [PATCH 03/14] mm: Move unmap_vmas() declaration to internal header Liam R. Howlett
2023-06-01  2:15 ` [PATCH 04/14] mm: Change do_vmi_align_munmap() side tree index Liam R. Howlett
2023-06-01  2:15 ` [PATCH 05/14] mm: Remove prev check from do_vmi_align_munmap() Liam R. Howlett
2023-06-01  2:15 ` [PATCH 06/14] maple_tree: Introduce __mas_set_range() Liam R. Howlett
2023-06-01  2:15 ` [PATCH 07/14] mm: Remove re-walk from mmap_region() Liam R. Howlett
2023-06-01  2:15 ` [PATCH 08/14] maple_tree: Re-introduce entry to mas_preallocate() arguments Liam R. Howlett
2023-06-01  2:16 ` [PATCH 09/14] mm: Use vma_iter_clear_gfp() in nommu Liam R. Howlett
2023-06-01  2:16 ` [PATCH 10/14] mm: Set up vma iterator for vma_iter_prealloc() calls Liam R. Howlett
2023-06-01  2:16 ` [PATCH 11/14] maple_tree: Move mas_wr_end_piv() below mas_wr_extend_null() Liam R. Howlett
2023-06-01  2:16 ` [PATCH 12/14] maple_tree: Update mas_preallocate() testing Liam R. Howlett
2023-06-01  2:16 ` Liam R. Howlett [this message]
2023-06-02 10:37   ` [PATCH 13/14] maple_tree: Refine mas_preallocate() node calculations Peng Zhang
2023-06-02 18:53     ` Liam R. Howlett
2023-06-01  2:16 ` [PATCH 14/14] mm/mmap: Change vma iteration order in do_vmi_align_munmap() Liam R. Howlett
2023-06-02  8:10 ` [PATCH 00/14] Reduce preallocations for maple tree Yin, Fengwei
2023-06-02 18:55   ` Liam R. Howlett
2023-06-04 12:10     ` Yin, Fengwei
2023-06-05  3:28   ` Peng Zhang
2023-06-05  4:41     ` Yin Fengwei
2023-06-05  6:18       ` Yin, Fengwei
2023-06-05  7:59         ` Peng Zhang
2023-06-05 14:03           ` Liam R. Howlett
2023-06-05 14:27             ` Peng Zhang
2023-06-05 14:44               ` Liam R. Howlett
2023-06-05 15:06                 ` Peng Zhang
2023-06-06  1:34             ` Yin Fengwei
2023-06-06  2:41             ` Hugh Dickins
2023-06-06  2:55               ` Yin, Fengwei
2023-06-06  3:08                 ` Hugh Dickins
2023-06-06  3:11                   ` Yin, Fengwei
2023-06-06 16:07                     ` Liam R. Howlett

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230601021605.2823123-14-Liam.Howlett@oracle.com \
    --to=liam.howlett@oracle.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=maple-tree@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®