From cdef1b62edf30b7a9035bb8af4b129830ae2f73b Mon Sep 17 00:00:00 2001
From: Mel Gorman <mgorman@suse.de>
Date: Thu, 19 Oct 2017 13:56:53 +0100
Subject: [PATCH] mm: remove __GFP_COLD
References: bsc#1060256
Patch-mainline: v4.15-rc1
Git-commit: 453f85d43fa9ee243f0fc3ac4e1be45615301e3f
As the page free path makes no distinction between cache hot and cold
pages, there is no real useful ordering of pages in the free list that
allocation requests can take advantage of. Juding from the users of
__GFP_COLD, it is likely that a number of them are the result of copying
other sites instead of actually measuring the impact. Remove the
__GFP_COLD parameter which simplifies a number of paths in the page
allocator.
This is potentially controversial but bear in mind that the size of the
per-cpu pagelists versus modern cache sizes means that the whole per-cpu
list can often fit in the L3 cache. Hence, there is only a potential
benefit for microbenchmarks that alloc/free pages in a tight loop. It's
even worse when THP is taken into account which has little or no chance of
getting a cache-hot page as the per-cpu list is bypassed and the zeroing
of multiple pages will thrash the cache anyway.
The truncate microbenchmarks are not shown as this patch affects the
allocation path and not the free path. A page fault microbenchmark was
tested but it showed no sigificant difference which is not surprising
given that the __GFP_COLD branches are a miniscule percentage of the fault
path.
Link: http://lkml.kernel.org/r/20171018075952.10627-9-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
drivers/net/ethernet/amazon/ena/ena_netdev.c | 2 -
drivers/net/ethernet/aquantia/atlantic/aq_ring.c | 3 --
drivers/net/ethernet/cavium/liquidio/octeon_network.h | 2 -
drivers/net/ethernet/mellanox/mlx4/en_rx.c | 5 +---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 2 -
drivers/net/ethernet/qlogic/qlge/qlge_main.c | 3 --
drivers/net/ethernet/sfc/falcon/rx.c | 2 -
drivers/net/ethernet/sfc/rx.c | 2 -
drivers/net/ethernet/synopsys/dwc-xlgmac-desc.c | 2 -
drivers/net/ethernet/ti/netcp_core.c | 2 -
drivers/net/virtio_net.c | 1
drivers/staging/lustre/lustre/mdc/mdc_request.c | 2 -
fs/cachefiles/rdwr.c | 6 +----
fs/ubifs/file.c | 2 -
include/linux/gfp.h | 5 ----
include/linux/pagemap.h | 8 -------
include/linux/skbuff.h | 2 -
include/linux/slab.h | 3 --
include/trace/events/mmflags.h | 1
kernel/power/snapshot.c | 4 +--
mm/filemap.c | 6 ++---
mm/page_alloc.c | 20 +++++-------------
mm/percpu-vm.c | 2 -
net/core/skbuff.c | 4 +--
tools/perf/builtin-kmem.c | 1
26 files changed, 32 insertions(+), 62 deletions(-)
--- a/drivers/net/ethernet/amazon/ena/ena_netdev.c
+++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c
@@ -517,7 +517,7 @@ static int ena_refill_rx_bufs(struct ena
rc = ena_alloc_rx_page(rx_ring, rx_info,
- __GFP_COLD | GFP_ATOMIC | __GFP_COMP);
+ GFP_ATOMIC | __GFP_COMP);
if (unlikely(rc < 0)) {
netif_warn(rx_ring->adapter, rx_err, rx_ring->netdev,
"failed to alloc buffer for rx queue %d\n",
--- a/drivers/net/ethernet/aquantia/atlantic/aq_ring.c
+++ b/drivers/net/ethernet/aquantia/atlantic/aq_ring.c
@@ -265,8 +265,7 @@ int aq_ring_rx_fill(struct aq_ring_s *se
buff->flags = 0U;
buff->len = AQ_CFG_RX_FRAME_MAX;
- buff->page = alloc_pages(GFP_ATOMIC | __GFP_COLD |
- __GFP_COMP, pages_order);
+ buff->page = alloc_pages(GFP_ATOMIC | __GFP_COMP, pages_order);
if (!buff->page) {
err = -ENOMEM;
goto err_exit;
--- a/drivers/net/ethernet/cavium/liquidio/octeon_network.h
+++ b/drivers/net/ethernet/cavium/liquidio/octeon_network.h
@@ -195,7 +195,7 @@ static inline void
struct sk_buff *skb;
struct octeon_skb_page_info *skb_pg_info;
- page = alloc_page(GFP_ATOMIC | __GFP_COLD);
+ page = alloc_page(GFP_ATOMIC);
if (unlikely(!page))
return NULL;
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -193,7 +193,7 @@ static int mlx4_en_fill_rx_buffers(struc
if (mlx4_en_prepare_rx_desc(priv, ring,
ring->actual_size,
- GFP_KERNEL | __GFP_COLD)) {
+ GFP_KERNEL)) {
if (ring->actual_size < MLX4_EN_MIN_RX_SIZE) {
en_err(priv, "Failed to allocate enough rx buffers\n");
return -ENOMEM;
@@ -552,8 +552,7 @@ static void mlx4_en_refill_rx_buffers(st
do {
if (mlx4_en_prepare_rx_desc(priv, ring,
ring->prod & ring->size_mask,
- GFP_ATOMIC | __GFP_COLD |
- __GFP_MEMALLOC))
+ GFP_ATOMIC | __GFP_MEMALLOC))
break;
ring->prod++;
} while (likely(--missing));
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -1185,7 +1185,7 @@ static void *nfp_net_rx_alloc_one(struct
} else {
struct page *page;
- page = alloc_page(GFP_KERNEL | __GFP_COLD);
+ page = alloc_page(GFP_KERNEL);
frag = page ? page_address(page) : NULL;
}
if (!frag) {
--- a/drivers/net/ethernet/qlogic/qlge/qlge_main.c
+++ b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
@@ -1092,8 +1092,7 @@ static int ql_get_next_chunk(struct ql_a
{
if (!rx_ring->pg_chunk.page) {
u64 map;
- rx_ring->pg_chunk.page = alloc_pages(__GFP_COLD | __GFP_COMP |
- GFP_ATOMIC,
+ rx_ring->pg_chunk.page = alloc_pages(__GFP_COMP | GFP_ATOMIC,
qdev->lbq_buf_order);
if (unlikely(!rx_ring->pg_chunk.page)) {
netif_err(qdev, drv, qdev->ndev,
--- a/drivers/net/ethernet/sfc/falcon/rx.c
+++ b/drivers/net/ethernet/sfc/falcon/rx.c
@@ -163,7 +163,7 @@ static int ef4_init_rx_buffers(struct ef
do {
page = ef4_reuse_page(rx_queue);
if (page == NULL) {
- page = alloc_pages(__GFP_COLD | __GFP_COMP |
+ page = alloc_pages(__GFP_COMP |
(atomic ? GFP_ATOMIC : GFP_KERNEL),
efx->rx_buffer_order);
if (unlikely(page == NULL))
--- a/drivers/net/ethernet/sfc/rx.c
+++ b/drivers/net/ethernet/sfc/rx.c
@@ -163,7 +163,7 @@ static int efx_init_rx_buffers(struct ef
do {
page = efx_reuse_page(rx_queue);
if (page == NULL) {
- page = alloc_pages(__GFP_COLD | __GFP_COMP |
+ page = alloc_pages(__GFP_COMP |
(atomic ? GFP_ATOMIC : GFP_KERNEL),
efx->rx_buffer_order);
if (unlikely(page == NULL))
--- a/drivers/net/ethernet/synopsys/dwc-xlgmac-desc.c
+++ b/drivers/net/ethernet/synopsys/dwc-xlgmac-desc.c
@@ -335,7 +335,7 @@ static int xlgmac_alloc_pages(struct xlg
dma_addr_t pages_dma;
/* Try to obtain pages, decreasing order if necessary */
- gfp |= __GFP_COLD | __GFP_COMP | __GFP_NOWARN;
+ gfp |= __GFP_COMP | __GFP_NOWARN;
while (order >= 0) {
pages = alloc_pages(gfp, order);
if (pages)
--- a/drivers/net/ethernet/ti/netcp_core.c
+++ b/drivers/net/ethernet/ti/netcp_core.c
@@ -906,7 +906,7 @@ static int netcp_allocate_rx_buf(struct
sw_data[0] = (u32)bufptr;
} else {
/* Allocate a secondary receive queue entry */
- page = alloc_page(GFP_ATOMIC | GFP_DMA | __GFP_COLD);
+ page = alloc_page(GFP_ATOMIC | GFP_DMA);
if (unlikely(!page)) {
dev_warn_ratelimited(netcp->ndev_dev, "Secondary page alloc failed\n");
goto fail;
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1031,7 +1031,6 @@ static bool try_fill_recv(struct virtnet
int err;
bool oom;
- gfp |= __GFP_COLD;
do {
if (vi->mergeable_rx_bufs)
err = add_recvbuf_mergeable(vi, rq, gfp);
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1150,7 +1150,7 @@ static int mdc_read_page_remote(void *da
}
for (npages = 1; npages < max_pages; npages++) {
- page = page_cache_alloc_cold(inode->i_mapping);
+ page = page_cache_alloc(inode->i_mapping);
if (!page)
break;
page_pool[npages] = page;
--- a/fs/cachefiles/rdwr.c
+++ b/fs/cachefiles/rdwr.c
@@ -256,8 +256,7 @@ static int cachefiles_read_backing_file_
goto backing_page_already_present;
if (!newpage) {
- newpage = __page_cache_alloc(cachefiles_gfp |
- __GFP_COLD);
+ newpage = __page_cache_alloc(cachefiles_gfp);
if (!newpage)
goto nomem_monitor;
}
@@ -493,8 +492,7 @@ static int cachefiles_read_backing_file(
goto backing_page_already_present;
if (!newpage) {
- newpage = __page_cache_alloc(cachefiles_gfp |
- __GFP_COLD);
+ newpage = __page_cache_alloc(cachefiles_gfp);
if (!newpage)
goto nomem;
}
--- a/fs/ubifs/file.c
+++ b/fs/ubifs/file.c
@@ -797,7 +797,7 @@ static int ubifs_do_bulk_read(struct ubi
if (page_offset > end_index)
break;
page = find_or_create_page(mapping, page_offset,
- GFP_NOFS | __GFP_COLD);
+ GFP_NOFS);
if (!page)
break;
if (!PageUptodate(page))
--- a/include/linux/gfp.h
+++ b/include/linux/gfp.h
@@ -23,7 +23,6 @@ struct vm_area_struct;
#define ___GFP_HIGH 0x20u
#define ___GFP_IO 0x40u
#define ___GFP_FS 0x80u
-#define ___GFP_COLD 0x100u
#define ___GFP_NOWARN 0x200u
#define ___GFP_REPEAT 0x400u
#define ___GFP_NOFAIL 0x800u
@@ -162,9 +161,6 @@ struct vm_area_struct;
/*
* Action modifiers
*
- * __GFP_COLD indicates that the caller does not expect to be used in the near
- * future. Where possible, a cache-cold page will be returned.
- *
* __GFP_NOWARN suppresses allocation failure reports.
*
* __GFP_COMP address compound page metadata.
@@ -177,7 +173,6 @@ struct vm_area_struct;
* distinguishing in the source between false positives and allocations that
* cannot be supported (e.g. page tables).
*/
-#define __GFP_COLD ((__force gfp_t)___GFP_COLD)
#define __GFP_NOWARN ((__force gfp_t)___GFP_NOWARN)
#define __GFP_COMP ((__force gfp_t)___GFP_COMP)
#define __GFP_ZERO ((__force gfp_t)___GFP_ZERO)
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -233,15 +233,9 @@ static inline struct page *page_cache_al
return __page_cache_alloc(mapping_gfp_mask(x));
}
-static inline struct page *page_cache_alloc_cold(struct address_space *x)
-{
- return __page_cache_alloc(mapping_gfp_mask(x)|__GFP_COLD);
-}
-
static inline gfp_t readahead_gfp_mask(struct address_space *x)
{
- return mapping_gfp_mask(x) |
- __GFP_COLD | __GFP_NORETRY | __GFP_NOWARN;
+ return mapping_gfp_mask(x) | __GFP_NORETRY | __GFP_NOWARN;
}
typedef int filler_t(void *, struct page *);
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -2530,7 +2530,7 @@ static inline struct page *__dev_alloc_p
* 4. __GFP_MEMALLOC is ignored if __GFP_NOMEMALLOC is set due to
* code in gfp_to_alloc_flags that should be enforcing this.
*/
- gfp_mask |= __GFP_COLD | __GFP_COMP | __GFP_MEMALLOC;
+ gfp_mask |= __GFP_COMP | __GFP_MEMALLOC;
return alloc_pages_node(NUMA_NO_NODE, gfp_mask, order);
}
--- a/include/linux/slab.h
+++ b/include/linux/slab.h
@@ -458,9 +458,6 @@ static __always_inline void *kmalloc_lar
* Also it is possible to set different flags by OR'ing
* in one or more of the following additional @flags:
*
- * %__GFP_COLD - Request cache-cold pages instead of
- * trying to return cache-warm pages.
- *
* %__GFP_HIGH - This allocation has high priority and may use emergency pools.
*
* %__GFP_NOFAIL - Indicate that this allocation is in no way allowed to fail
--- a/include/trace/events/mmflags.h
+++ b/include/trace/events/mmflags.h
@@ -32,7 +32,6 @@
{(unsigned long)__GFP_ATOMIC, "__GFP_ATOMIC"}, \
{(unsigned long)__GFP_IO, "__GFP_IO"}, \
{(unsigned long)__GFP_FS, "__GFP_FS"}, \
- {(unsigned long)__GFP_COLD, "__GFP_COLD"}, \
{(unsigned long)__GFP_NOWARN, "__GFP_NOWARN"}, \
{(unsigned long)__GFP_REPEAT, "__GFP_REPEAT"}, \
{(unsigned long)__GFP_NOFAIL, "__GFP_NOFAIL"}, \
--- a/kernel/power/snapshot.c
+++ b/kernel/power/snapshot.c
@@ -1884,7 +1884,7 @@ static int enough_free_mem(unsigned int
*/
static inline int get_highmem_buffer(int safe_needed)
{
- buffer = get_image_page(GFP_ATOMIC | __GFP_COLD, safe_needed);
+ buffer = get_image_page(GFP_ATOMIC, safe_needed);
return buffer ? 0 : -ENOMEM;
}
@@ -1946,7 +1946,7 @@ static int swsusp_alloc(struct memory_bi
while (nr_pages-- > 0) {
struct page *page;
- page = alloc_image_page(GFP_ATOMIC | __GFP_COLD);
+ page = alloc_image_page(GFP_ATOMIC);
if (!page)
goto err_out;
memory_bm_set_bit(copy_bm, page_to_pfn(page));
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2225,7 +2225,7 @@ no_cached_page:
* Ok, it wasn't cached, so we need to create a new
* page..
*/
- page = page_cache_alloc_cold(mapping);
+ page = page_cache_alloc(mapping);
if (!page) {
error = -ENOMEM;
goto out;
@@ -2335,7 +2335,7 @@ static int page_cache_read(struct file *
int ret;
do {
- page = __page_cache_alloc(gfp_mask|__GFP_COLD);
+ page = __page_cache_alloc(gfp_mask);
if (!page)
return -ENOMEM;
@@ -2739,7 +2739,7 @@ static struct page *do_read_cache_page(s
repeat:
page = find_get_page(mapping, index);
if (!page) {
- page = __page_cache_alloc(gfp | __GFP_COLD);
+ page = __page_cache_alloc(gfp);
if (!page)
return ERR_PTR(-ENOMEM);
err = add_to_page_cache_lru(page, mapping, index, gfp);
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2276,7 +2276,7 @@ retry:
*/
static int rmqueue_bulk(struct zone *zone, unsigned int order,
unsigned long count, struct list_head *list,
- int migratetype, bool cold)
+ int migratetype)
{
int i, alloced = 0;
@@ -2298,10 +2298,7 @@ static int rmqueue_bulk(struct zone *zon
* merge IO requests if the physical pages are ordered
* properly.
*/
- if (likely(!cold))
- list_add(&page->lru, list);
- else
- list_add_tail(&page->lru, list);
+ list_add(&page->lru, list);
list = &page->lru;
alloced++;
if (is_migrate_cma(get_pcppage_migratetype(page)))
@@ -2744,7 +2741,7 @@ static inline void zone_statistics(struc
/* Remove page from the per-cpu list, caller must protect the list */
static struct page *__rmqueue_pcplist(struct zone *zone, int migratetype,
- bool cold, struct per_cpu_pages *pcp,
+ struct per_cpu_pages *pcp,
struct list_head *list)
{
struct page *page;
@@ -2753,16 +2750,12 @@ static struct page *__rmqueue_pcplist(st
if (list_empty(list)) {
pcp->count += rmqueue_bulk(zone, 0,
pcp->batch, list,
- migratetype, cold);
+ migratetype);
if (unlikely(list_empty(list)))
return NULL;
}
- if (cold)
- page = list_last_entry(list, struct page, lru);
- else
- page = list_first_entry(list, struct page, lru);
-
+ page = list_first_entry(list, struct page, lru);
list_del(&page->lru);
pcp->count--;
} while (check_new_pcp(page));
@@ -2777,14 +2770,13 @@ static struct page *rmqueue_pcplist(stru
{
struct per_cpu_pages *pcp;
struct list_head *list;
- bool cold = ((gfp_flags & __GFP_COLD) != 0);
struct page *page;
unsigned long flags;
local_irq_save(flags);
pcp = &this_cpu_ptr(zone->pageset)->pcp;
list = &pcp->lists[migratetype];
- page = __rmqueue_pcplist(zone, migratetype, cold, pcp, list);
+ page = __rmqueue_pcplist(zone, migratetype, pcp, list);
if (page) {
__count_zid_vm_events(PGALLOC, page_zonenum(page), 1 << order);
zone_statistics(preferred_zone, zone);
--- a/mm/percpu-vm.c
+++ b/mm/percpu-vm.c
@@ -81,7 +81,7 @@ static void pcpu_free_pages(struct pcpu_
static int pcpu_alloc_pages(struct pcpu_chunk *chunk,
struct page **pages, int page_start, int page_end)
{
- const gfp_t gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_COLD;
+ const gfp_t gfp = GFP_KERNEL | __GFP_HIGHMEM;
unsigned int cpu, tcpu;
int i;
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -382,7 +382,7 @@ static void *__netdev_alloc_frag(unsigne
*/
void *netdev_alloc_frag(unsigned int fragsz)
{
- return __netdev_alloc_frag(fragsz, GFP_ATOMIC | __GFP_COLD);
+ return __netdev_alloc_frag(fragsz, GFP_ATOMIC);
}
EXPORT_SYMBOL(netdev_alloc_frag);
@@ -395,7 +395,7 @@ static void *__napi_alloc_frag(unsigned
void *napi_alloc_frag(unsigned int fragsz)
{
- return __napi_alloc_frag(fragsz, GFP_ATOMIC | __GFP_COLD);
+ return __napi_alloc_frag(fragsz, GFP_ATOMIC);
}
EXPORT_SYMBOL(napi_alloc_frag);
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -641,7 +641,6 @@ static const struct {
{ "__GFP_ATOMIC", "_A" },
{ "__GFP_IO", "I" },
{ "__GFP_FS", "F" },
- { "__GFP_COLD", "CO" },
{ "__GFP_NOWARN", "NWR" },
{ "__GFP_REPEAT", "R" },
{ "__GFP_NOFAIL", "NF" },