mm: remove __GFP_COLD

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@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
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>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

13 files changed, 14 insertions, 18 deletions

diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c
index c6bd5e2..fbbbd8b 100644
--- 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_ring *rx_ring, u32 num)
 
 
 		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",
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-desc.c b/drivers/net/ethernet/amd/xgbe/xgbe-desc.c
index 45d9230..cc1e4f8 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe-desc.c
+++ b/drivers/net/ethernet/amd/xgbe/xgbe-desc.c
@@ -295,7 +295,7 @@ again:
 	order = alloc_order;
 
 	/* Try to obtain pages, decreasing order if necessary */
-	gfp = GFP_ATOMIC | __GFP_COLD | __GFP_COMP | __GFP_NOWARN;
+	gfp = GFP_ATOMIC | __GFP_COMP | __GFP_NOWARN;
 	while (order >= 0) {
 		pages = alloc_pages_node(node, gfp, order);
 		if (pages)
diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_ring.c b/drivers/net/ethernet/aquantia/atlantic/aq_ring.c
index 0654e0c..519ca65 100644
--- a/drivers/net/ethernet/aquantia/atlantic/aq_ring.c
+++ b/drivers/net/ethernet/aquantia/atlantic/aq_ring.c
@@ -304,8 +304,7 @@ int aq_ring_rx_fill(struct aq_ring_s *self)
 		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;
diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_network.h b/drivers/net/ethernet/cavium/liquidio/octeon_network.h
index 9e36319..57853ee 100644
--- 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;
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index b97a55c8..ffead38 100644
--- 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(struct mlx4_en_priv *priv)
 
 			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(struct mlx4_en_priv *priv,
 	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));
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index e118b5f..ffb12dc 100644
--- 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 nfp_net_dp *dp, dma_addr_t *dma_addr)
 	} else {
 		struct page *page;
 
-		page = alloc_page(GFP_KERNEL | __GFP_COLD);
+		page = alloc_page(GFP_KERNEL);
 		frag = page ? page_address(page) : NULL;
 	}
 	if (!frag) {
@@ -1212,7 +1212,7 @@ static void *nfp_net_napi_alloc_one(struct nfp_net_dp *dp, dma_addr_t *dma_addr)
 	} else {
 		struct page *page;
 
-		page = alloc_page(GFP_ATOMIC | __GFP_COLD);
+		page = alloc_page(GFP_ATOMIC);
 		frag = page ? page_address(page) : NULL;
 	}
 	if (!frag) {
diff --git a/drivers/net/ethernet/qlogic/qlge/qlge_main.c b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
index 29fea74..7b97a99 100644
--- 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_adapter *qdev, struct rx_ring *rx_ring,
 {
 	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,
diff --git a/drivers/net/ethernet/sfc/falcon/rx.c b/drivers/net/ethernet/sfc/falcon/rx.c
index 6a8406d..91097ae 100644
--- 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 ef4_rx_queue *rx_queue, bool atomic)
 	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))
diff --git a/drivers/net/ethernet/sfc/rx.c b/drivers/net/ethernet/sfc/rx.c
index 42443f4..0004c50 100644
--- 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 efx_rx_queue *rx_queue, bool atomic)
 	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))
diff --git a/drivers/net/ethernet/synopsys/dwc-xlgmac-desc.c b/drivers/net/ethernet/synopsys/dwc-xlgmac-desc.c
index e9672b1..031cf9c 100644
--- 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 xlgmac_pdata *pdata,
 	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)
diff --git a/drivers/net/ethernet/ti/netcp_core.c b/drivers/net/ethernet/ti/netcp_core.c
index 437d362..50d2b76 100644
--- 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 netcp_intf *netcp, int fdq)
 		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;
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 511f833..5eec09d 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -988,7 +988,6 @@ static bool try_fill_recv(struct virtnet_info *vi, struct receive_queue *rq,
 	int err;
 	bool oom;
 
-	gfp |= __GFP_COLD;
 	do {
 		if (vi->mergeable_rx_bufs)
 			err = add_recvbuf_mergeable(vi, rq, gfp);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 9e538a5..03e55bc 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1152,7 +1152,7 @@ static int mdc_read_page_remote(void *data, struct page *page0)
 	}
 
 	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;