[RESEND PATCH v4 0/7] hugetlb: parallelize hugetlb page init on boot

From: Gang Li
Date: Thu Jan 18 2024 - 07:39:33 EST


Hi all, hugetlb init parallelization has now been updated to v4.

This version is tested on next-20240112.

Update Summary:
- Make padata_do_multithreaded dispatch all jobs with a global iterator
- Revise commit message
- Rename some functions
- Collect Tested-by and Reviewed-by

# Introduction
Hugetlb initialization during boot takes up a considerable amount of time.
For instance, on a 2TB system, initializing 1,800 1GB huge pages takes 1-2
seconds out of 10 seconds. Initializing 11,776 1GB pages on a 12TB Intel
host takes more than 1 minute[1]. This is a noteworthy figure.

Inspired by [2] and [3], hugetlb initialization can also be accelerated
through parallelization. Kernel already has infrastructure like
padata_do_multithreaded, this patch uses it to achieve effective results
by minimal modifications.

[1] https://lore.kernel.org/all/783f8bac-55b8-5b95-eb6a-11a583675000@xxxxxxxxxx/
[2] https://lore.kernel.org/all/20200527173608.2885243-1-daniel.m.jordan@xxxxxxxxxx/
[3] https://lore.kernel.org/all/20230906112605.2286994-1-usama.arif@xxxxxxxxxxxxx/
[4] https://lore.kernel.org/all/76becfc1-e609-e3e8-2966-4053143170b6@xxxxxxxxxx/

# Test result
test no patch(ms) patched(ms) saved
------------------- -------------- ------------- --------
256c2t(4 node) 1G 4745 2024 57.34%
128c1t(2 node) 1G 3358 1712 49.02%
12t 1G 77000 18300 76.23%

256c2t(4 node) 2M 3336 1051 68.52%
128c1t(2 node) 2M 1943 716 63.15%

# Change log
Changes in v4:
- Make padata_do_multithreaded dispatch all jobs with a global iterator
- Revise commit message
- Rename some functions
- Collect Tested-by and Reviewed-by

Changes in v3:
- https://lore.kernel.org/all/20240102131249.76622-1-gang.li@xxxxxxxxx/
- Select CONFIG_PADATA as we use padata_do_multithreaded
- Fix a race condition in h->next_nid_to_alloc
- Fix local variable initialization issues
- Remove RFC tag

Changes in v2:
- https://lore.kernel.org/all/20231208025240.4744-1-gang.li@xxxxxxxxx/
- Reduce complexity with `padata_do_multithreaded`
- Support 1G hugetlb

v1:
- https://lore.kernel.org/all/20231123133036.68540-1-gang.li@xxxxxxxxx/
- parallelize 2M hugetlb initialization with workqueue

Gang Li (7):
hugetlb: code clean for hugetlb_hstate_alloc_pages
hugetlb: split hugetlb_hstate_alloc_pages
padata: dispatch works on different nodes
hugetlb: pass *next_nid_to_alloc directly to
for_each_node_mask_to_alloc
hugetlb: have CONFIG_HUGETLBFS select CONFIG_PADATA
hugetlb: parallelize 2M hugetlb allocation and initialization
hugetlb: parallelize 1G hugetlb initialization

fs/Kconfig | 1 +
include/linux/hugetlb.h | 2 +-
include/linux/padata.h | 3 +
kernel/padata.c | 14 ++-
mm/hugetlb.c | 227 ++++++++++++++++++++++++++--------------
mm/mm_init.c | 1 +
6 files changed, 168 insertions(+), 80 deletions(-)

--
2.20.1