LKML Archive on lore.kernel.org
 help / Atom feed
* [PATCH v2] mm: hugetlb: don't zero 1GiB bootmem pages.
@ 2018-07-11 21:33 Cannon Matthews
  2018-07-11 21:56 ` Mike Kravetz
  2018-07-12  7:48 ` Michal Hocko
  0 siblings, 2 replies; 3+ messages in thread
From: Cannon Matthews @ 2018-07-11 21:33 UTC (permalink / raw)
  To: Andrew Morton, Mike Kravetz, Nadia Yvette Chambers
  Cc: linux-mm, linux-kernel, andreslc, pfeiner, dmatlack, gthelen,
	mhocko, Cannon Matthews

When using 1GiB pages during early boot, use the new
memblock_virt_alloc_try_nid_raw() function to allocate memory without
zeroing it.  Zeroing out hundreds or thousands of GiB in a single core
memset() call is very slow, and can make early boot last upwards of
20-30 minutes on multi TiB machines.

The memory does not need to be zero'd as the hugetlb pages are always
zero'd on page fault.

Tested: Booted with ~3800 1G pages, and it booted successfully in
roughly the same amount of time as with 0, as opposed to the 25+
minutes it would take before.

Signed-off-by: Cannon Matthews <cannonmatthews@google.com>
---
v2: removed the memset of the huge_bootmem_page area and added
INIT_LIST_HEAD instead.

 mm/hugetlb.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 3612fbb32e9d..488330f23f04 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -2101,7 +2101,7 @@ int __alloc_bootmem_huge_page(struct hstate *h)
 	for_each_node_mask_to_alloc(h, nr_nodes, node, &node_states[N_MEMORY]) {
 		void *addr;

-		addr = memblock_virt_alloc_try_nid_nopanic(
+		addr = memblock_virt_alloc_try_nid_raw(
 				huge_page_size(h), huge_page_size(h),
 				0, BOOTMEM_ALLOC_ACCESSIBLE, node);
 		if (addr) {
@@ -2119,6 +2119,7 @@ int __alloc_bootmem_huge_page(struct hstate *h)
 found:
 	BUG_ON(!IS_ALIGNED(virt_to_phys(m), huge_page_size(h)));
 	/* Put them into a private list first because mem_map is not up yet */
+	INIT_LIST_HEAD(&m->list);
 	list_add(&m->list, &huge_boot_pages);
 	m->hstate = h;
 	return 1;
--
2.18.0.203.gfac676dfb9-goog


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v2] mm: hugetlb: don't zero 1GiB bootmem pages.
  2018-07-11 21:33 [PATCH v2] mm: hugetlb: don't zero 1GiB bootmem pages Cannon Matthews
@ 2018-07-11 21:56 ` Mike Kravetz
  2018-07-12  7:48 ` Michal Hocko
  1 sibling, 0 replies; 3+ messages in thread
From: Mike Kravetz @ 2018-07-11 21:56 UTC (permalink / raw)
  To: Cannon Matthews, Andrew Morton, Nadia Yvette Chambers
  Cc: linux-mm, linux-kernel, andreslc, pfeiner, dmatlack, gthelen, mhocko

On 07/11/2018 02:33 PM, Cannon Matthews wrote:
> When using 1GiB pages during early boot, use the new
> memblock_virt_alloc_try_nid_raw() function to allocate memory without
> zeroing it.  Zeroing out hundreds or thousands of GiB in a single core
> memset() call is very slow, and can make early boot last upwards of
> 20-30 minutes on multi TiB machines.
> 
> The memory does not need to be zero'd as the hugetlb pages are always
> zero'd on page fault.
> 
> Tested: Booted with ~3800 1G pages, and it booted successfully in
> roughly the same amount of time as with 0, as opposed to the 25+
> minutes it would take before.
> 
> Signed-off-by: Cannon Matthews <cannonmatthews@google.com>

Thanks,

Acked-by: Mike Kravetz <mike.kravetz@oracle.com>

-- 
Mike Kravetz

> ---
> v2: removed the memset of the huge_bootmem_page area and added
> INIT_LIST_HEAD instead.
> 
>  mm/hugetlb.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> index 3612fbb32e9d..488330f23f04 100644
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -2101,7 +2101,7 @@ int __alloc_bootmem_huge_page(struct hstate *h)
>  	for_each_node_mask_to_alloc(h, nr_nodes, node, &node_states[N_MEMORY]) {
>  		void *addr;
> 
> -		addr = memblock_virt_alloc_try_nid_nopanic(
> +		addr = memblock_virt_alloc_try_nid_raw(
>  				huge_page_size(h), huge_page_size(h),
>  				0, BOOTMEM_ALLOC_ACCESSIBLE, node);
>  		if (addr) {
> @@ -2119,6 +2119,7 @@ int __alloc_bootmem_huge_page(struct hstate *h)
>  found:
>  	BUG_ON(!IS_ALIGNED(virt_to_phys(m), huge_page_size(h)));
>  	/* Put them into a private list first because mem_map is not up yet */
> +	INIT_LIST_HEAD(&m->list);
>  	list_add(&m->list, &huge_boot_pages);
>  	m->hstate = h;
>  	return 1;
> --
> 2.18.0.203.gfac676dfb9-goog
> 

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v2] mm: hugetlb: don't zero 1GiB bootmem pages.
  2018-07-11 21:33 [PATCH v2] mm: hugetlb: don't zero 1GiB bootmem pages Cannon Matthews
  2018-07-11 21:56 ` Mike Kravetz
@ 2018-07-12  7:48 ` Michal Hocko
  1 sibling, 0 replies; 3+ messages in thread
From: Michal Hocko @ 2018-07-12  7:48 UTC (permalink / raw)
  To: Cannon Matthews
  Cc: Andrew Morton, Mike Kravetz, Nadia Yvette Chambers, linux-mm,
	linux-kernel, andreslc, pfeiner, dmatlack, gthelen

On Wed 11-07-18 14:33:13, Cannon Matthews wrote:
> When using 1GiB pages during early boot, use the new
> memblock_virt_alloc_try_nid_raw() function to allocate memory without
> zeroing it.  Zeroing out hundreds or thousands of GiB in a single core
> memset() call is very slow, and can make early boot last upwards of
> 20-30 minutes on multi TiB machines.
> 
> The memory does not need to be zero'd as the hugetlb pages are always
> zero'd on page fault.
> 
> Tested: Booted with ~3800 1G pages, and it booted successfully in
> roughly the same amount of time as with 0, as opposed to the 25+
> minutes it would take before.
> 
> Signed-off-by: Cannon Matthews <cannonmatthews@google.com>

Thanks for the updated version.

Acked-by: Michal Hocko <mhocko@suse.com>

> ---
> v2: removed the memset of the huge_bootmem_page area and added
> INIT_LIST_HEAD instead.
> 
>  mm/hugetlb.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> index 3612fbb32e9d..488330f23f04 100644
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -2101,7 +2101,7 @@ int __alloc_bootmem_huge_page(struct hstate *h)
>  	for_each_node_mask_to_alloc(h, nr_nodes, node, &node_states[N_MEMORY]) {
>  		void *addr;
> 
> -		addr = memblock_virt_alloc_try_nid_nopanic(
> +		addr = memblock_virt_alloc_try_nid_raw(
>  				huge_page_size(h), huge_page_size(h),
>  				0, BOOTMEM_ALLOC_ACCESSIBLE, node);
>  		if (addr) {
> @@ -2119,6 +2119,7 @@ int __alloc_bootmem_huge_page(struct hstate *h)
>  found:
>  	BUG_ON(!IS_ALIGNED(virt_to_phys(m), huge_page_size(h)));
>  	/* Put them into a private list first because mem_map is not up yet */
> +	INIT_LIST_HEAD(&m->list);
>  	list_add(&m->list, &huge_boot_pages);
>  	m->hstate = h;
>  	return 1;
> --
> 2.18.0.203.gfac676dfb9-goog

-- 
Michal Hocko
SUSE Labs

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, back to index

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-07-11 21:33 [PATCH v2] mm: hugetlb: don't zero 1GiB bootmem pages Cannon Matthews
2018-07-11 21:56 ` Mike Kravetz
2018-07-12  7:48 ` Michal Hocko

LKML Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/lkml/0 lkml/git/0.git
	git clone --mirror https://lore.kernel.org/lkml/1 lkml/git/1.git
	git clone --mirror https://lore.kernel.org/lkml/2 lkml/git/2.git
	git clone --mirror https://lore.kernel.org/lkml/3 lkml/git/3.git
	git clone --mirror https://lore.kernel.org/lkml/4 lkml/git/4.git
	git clone --mirror https://lore.kernel.org/lkml/5 lkml/git/5.git
	git clone --mirror https://lore.kernel.org/lkml/6 lkml/git/6.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 lkml lkml/ https://lore.kernel.org/lkml \
		linux-kernel@vger.kernel.org linux-kernel@archiver.kernel.org
	public-inbox-index lkml


Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-kernel


AGPL code for this site: git clone https://public-inbox.org/ public-inbox