Commit 8d521d94 authored Mar 07, 2019 by Linus Torvalds

percpu

Pull percpu updates from Dennis Zhou:
 "There are 2 minor changes to the percpu allocator this merge window:

   - for loop condition that could be out of bounds on multi-socket UP

   - cosmetic removal of pcpu_group_offsets[0] in UP code as it is 0

  There has been an interest in having better alignment with percpu
  allocations. This has caused a performance regression in at least one
  reported workload. I have a series out which adds scan hints to the
  allocator as well as some other performance oriented changes. I hope
  to have this queued for v5.2 soon"

* 'for-5.1' of git://git.kernel.org/pub/scm/linux/kernel/git/dennis/percpu:
  percpu: km: no need to consider pcpu_group_offsets[0]
  percpu: use nr_groups as check condition

parents bdfa15f1 1b046b44

mm/percpu-km.c

+1 −1

Original line number	Diff line number	Diff line
		@@ -67,7 +67,7 @@ static struct pcpu_chunk *pcpu_create_chunk(gfp_t gfp)
		pcpu_set_page_chunk(nth_page(pages, i), chunk);

		chunk->data = pages;
		chunk->base_addr = page_address(pages) - pcpu_group_offsets[0];
		chunk->base_addr = page_address(pages);

		spin_lock_irqsave(&pcpu_lock, flags);
		pcpu_chunk_populated(chunk, 0, nr_pages, false);

mm/percpu.c

+1 −1

Original line number	Diff line number	Diff line
		@@ -2384,7 +2384,7 @@ static struct pcpu_alloc_info * __init pcpu_build_alloc_info(
		ai->atom_size = atom_size;
		ai->alloc_size = alloc_size;

		for (group = 0, unit = 0; group_cnt[group]; group++) {
		for (group = 0, unit = 0; group < nr_groups; group++) {
		struct pcpu_group_info *gi = &ai->groups[group];

		/*

Admin message