From d8be75663cec0069b85f80191abd2682ce4a512f Mon Sep 17 00:00:00 2001 From: "Levin, Alexander (Sasha Levin)" Date: Wed, 15 Nov 2017 17:35:58 -0800 Subject: kmemcheck: remove whats left of NOTRACK flags Now that kmemcheck is gone, we don't need the NOTRACK flags. Link: http://lkml.kernel.org/r/20171007030159.22241-5-alexander.levin@verizon.com Signed-off-by: Sasha Levin Cc: Alexander Potapenko Cc: Eric W. Biederman Cc: Michal Hocko Cc: Pekka Enberg Cc: Steven Rostedt Cc: Tim Hansen Cc: Vegard Nossum Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- tools/perf/builtin-kmem.c | 1 - 1 file changed, 1 deletion(-) (limited to 'tools/perf') diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 557d391f564a..cbf70738ef5f 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -655,7 +655,6 @@ static const struct { { "__GFP_RECLAIMABLE", "RC" }, { "__GFP_MOVABLE", "M" }, { "__GFP_ACCOUNT", "AC" }, - { "__GFP_NOTRACK", "NT" }, { "__GFP_WRITE", "WR" }, { "__GFP_RECLAIM", "R" }, { "__GFP_DIRECT_RECLAIM", "DR" }, -- cgit v1.2.3 From 453f85d43fa9ee243f0fc3ac4e1be45615301e3f Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 15 Nov 2017 17:38:03 -0800 Subject: 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 Acked-by: Vlastimil Babka Cc: Andi Kleen Cc: Dave Chinner Cc: Dave Hansen Cc: Jan Kara Cc: Johannes Weiner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/net/ethernet/amazon/ena/ena_netdev.c | 2 +- drivers/net/ethernet/amd/xgbe/xgbe-desc.c | 2 +- drivers/net/ethernet/aquantia/atlantic/aq_ring.c | 3 +-- .../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 | 4 ++-- 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 ++---- 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 - 25 files changed, 32 insertions(+), 62 deletions(-) (limited to 'tools/perf') diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c index c6bd5e24005d..fbbbd8b3eb45 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 45d92304068e..cc1e4f820e64 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 0654e0c76bc2..519ca6534b85 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 9e36319cead6..57853eead4b5 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 b97a55c827eb..ffead38cf5da 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 e118b5f23996..ffb12dc13a5a 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 29fea74bff2e..7b97a9969046 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 6a8406dc0c2b..91097aea6c41 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 42443f434569..0004c50d3c83 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 e9672b1f9968..031cf9c3435a 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 437d36289786..50d2b76771b5 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 511f8339fa96..5eec09d63fc0 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 9e538a59f09d..03e55bca4ada 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; diff --git a/fs/cachefiles/rdwr.c b/fs/cachefiles/rdwr.c index 23097cca2674..883bc7bb12c5 100644 --- a/fs/cachefiles/rdwr.c +++ b/fs/cachefiles/rdwr.c @@ -256,8 +256,7 @@ static int cachefiles_read_backing_file_one(struct cachefiles_object *object, 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(struct cachefiles_object *object, 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; } diff --git a/include/linux/gfp.h b/include/linux/gfp.h index f7e62d9096fe..1a4582b44d32 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -24,7 +24,6 @@ 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_RETRY_MAYFAIL 0x400u #define ___GFP_NOFAIL 0x800u @@ -192,16 +191,12 @@ 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. * * __GFP_ZERO returns a zeroed page on success. */ -#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) diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index 4c6790bb7afb..34ce3ebf97d5 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -234,15 +234,9 @@ static inline struct page *page_cache_alloc(struct address_space *x) 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 *); diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index aa1341474916..7c46fd0b8b64 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -2672,7 +2672,7 @@ static inline struct page *__dev_alloc_pages(gfp_t gfp_mask, * 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); } diff --git a/include/linux/slab.h b/include/linux/slab.h index 79f6532f8a0b..50697a1d6621 100644 --- a/include/linux/slab.h +++ b/include/linux/slab.h @@ -467,9 +467,6 @@ static __always_inline void *kmalloc_large(size_t size, gfp_t flags) * 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 diff --git a/include/trace/events/mmflags.h b/include/trace/events/mmflags.h index 72162f3a03fa..dbe1bb058c09 100644 --- 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_RETRY_MAYFAIL, "__GFP_RETRY_MAYFAIL"}, \ {(unsigned long)__GFP_NOFAIL, "__GFP_NOFAIL"}, \ diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index a917a301e201..bce0464524d8 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -1884,7 +1884,7 @@ static int enough_free_mem(unsigned int nr_pages, unsigned int nr_highmem) */ 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; } @@ -1945,7 +1945,7 @@ static int swsusp_alloc(struct memory_bitmap *copy_bm, 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)); diff --git a/mm/filemap.c b/mm/filemap.c index 90a9f261f85f..923fc2ebd74a 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2272,7 +2272,7 @@ 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; @@ -2384,7 +2384,7 @@ static int page_cache_read(struct file *file, pgoff_t offset, gfp_t gfp_mask) int ret; do { - page = __page_cache_alloc(gfp_mask|__GFP_COLD); + page = __page_cache_alloc(gfp_mask); if (!page) return -ENOMEM; @@ -2788,7 +2788,7 @@ static struct page *do_read_cache_page(struct address_space *mapping, 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); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index f265d37b3152..370b64d03e3f 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -2336,7 +2336,7 @@ 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; @@ -2358,10 +2358,7 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order, * 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))) @@ -2795,7 +2792,7 @@ static inline void zone_statistics(struct zone *preferred_zone, struct zone *z) /* 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; @@ -2804,16 +2801,12 @@ static struct page *__rmqueue_pcplist(struct zone *zone, int migratetype, 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)); @@ -2828,14 +2821,13 @@ static struct page *rmqueue_pcplist(struct zone *preferred_zone, { 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); diff --git a/mm/percpu-vm.c b/mm/percpu-vm.c index 15dab691ea70..9158e5a81391 100644 --- a/mm/percpu-vm.c +++ b/mm/percpu-vm.c @@ -81,7 +81,7 @@ static void pcpu_free_pages(struct pcpu_chunk *chunk, 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; diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 6cd057b41f34..9c68555bb906 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -353,7 +353,7 @@ static void *__netdev_alloc_frag(unsigned int fragsz, gfp_t gfp_mask) */ 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); @@ -366,7 +366,7 @@ static void *__napi_alloc_frag(unsigned int fragsz, gfp_t gfp_mask) 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); diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index cbf70738ef5f..ae11e4c3516a 100644 --- 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_RETRY_MAYFAIL", "R" }, { "__GFP_NOFAIL", "NF" }, -- cgit v1.2.3 From a3f22d505f56bf3efd8eecd9b7adcc40a40a4a8e Mon Sep 17 00:00:00 2001 From: Pu Hou Date: Thu, 1 Sep 2016 10:48:22 +0200 Subject: s390/perf: add callback to perf to enable using AUX buffer Perf tool need implement a callback to enable using AUX buffer. Perf will do another mmap() to trigger the setup of AUX buffer in kernel if there is such callback. The default size of the AUX buffer is set properly according to the sampling frequency to avoid overflow. It could also be manually set by -m option of perf. The interface of perf is not changed. Diagnostic mode sampling could be started by `perf record -e rBD000` like before. Signed-off-by: Pu Hou Reviewed-by: Hendrik Brueckner Signed-off-by: Martin Schwidefsky --- tools/perf/arch/s390/util/Build | 2 + tools/perf/arch/s390/util/auxtrace.c | 118 +++++++++++++++++++++++++++++++++++ 2 files changed, 120 insertions(+) create mode 100644 tools/perf/arch/s390/util/auxtrace.c (limited to 'tools/perf') diff --git a/tools/perf/arch/s390/util/Build b/tools/perf/arch/s390/util/Build index 5bd7b9260cc0..397084382b23 100644 --- a/tools/perf/arch/s390/util/Build +++ b/tools/perf/arch/s390/util/Build @@ -4,3 +4,5 @@ libperf-y += kvm-stat.o libperf-$(CONFIG_DWARF) += dwarf-regs.o libperf-y += machine.o + +libperf-$(CONFIG_AUXTRACE) += auxtrace.o diff --git a/tools/perf/arch/s390/util/auxtrace.c b/tools/perf/arch/s390/util/auxtrace.c new file mode 100644 index 000000000000..6cb48e4cffd9 --- /dev/null +++ b/tools/perf/arch/s390/util/auxtrace.c @@ -0,0 +1,118 @@ +#include +#include +#include +#include +#include + +#include "../../util/evlist.h" +#include "../../util/auxtrace.h" +#include "../../util/evsel.h" + +#define PERF_EVENT_CPUM_SF 0xB0000 /* Event: Basic-sampling */ +#define PERF_EVENT_CPUM_SF_DIAG 0xBD000 /* Event: Combined-sampling */ +#define DEFAULT_AUX_PAGES 128 +#define DEFAULT_FREQ 4000 + +static void cpumsf_free(struct auxtrace_record *itr) +{ + free(itr); +} + +static size_t cpumsf_info_priv_size(struct auxtrace_record *itr __maybe_unused, + struct perf_evlist *evlist __maybe_unused) +{ + return 0; +} + +static int +cpumsf_info_fill(struct auxtrace_record *itr __maybe_unused, + struct perf_session *session __maybe_unused, + struct auxtrace_info_event *auxtrace_info __maybe_unused, + size_t priv_size __maybe_unused) +{ + return 0; +} + +static unsigned long +cpumsf_reference(struct auxtrace_record *itr __maybe_unused) +{ + return 0; +} + +static int +cpumsf_recording_options(struct auxtrace_record *ar __maybe_unused, + struct perf_evlist *evlist __maybe_unused, + struct record_opts *opts) +{ + unsigned int factor = 1; + unsigned int pages; + + opts->full_auxtrace = true; + + /* + * The AUX buffer size should be set properly to avoid + * overflow of samples if it is not set explicitly. + * DEFAULT_AUX_PAGES is an proper size when sampling frequency + * is DEFAULT_FREQ. It is expected to hold about 1/2 second + * of sampling data. The size used for AUX buffer will scale + * according to the specified frequency and DEFAULT_FREQ. + */ + if (!opts->auxtrace_mmap_pages) { + if (opts->user_freq != UINT_MAX) + factor = (opts->user_freq + DEFAULT_FREQ + - 1) / DEFAULT_FREQ; + pages = DEFAULT_AUX_PAGES * factor; + opts->auxtrace_mmap_pages = roundup_pow_of_two(pages); + } + + return 0; +} + +static int +cpumsf_parse_snapshot_options(struct auxtrace_record *itr __maybe_unused, + struct record_opts *opts __maybe_unused, + const char *str __maybe_unused) +{ + return 0; +} + +/* + * auxtrace_record__init is called when perf record + * check if the event really need auxtrace + */ +struct auxtrace_record *auxtrace_record__init(struct perf_evlist *evlist, + int *err) +{ + struct auxtrace_record *aux; + struct perf_evsel *pos; + int diagnose = 0; + + if (evlist->nr_entries == 0) + return NULL; + + evlist__for_each_entry(evlist, pos) { + if (pos->attr.config == PERF_EVENT_CPUM_SF_DIAG) { + diagnose = 1; + break; + } + } + + if (!diagnose) + return NULL; + + /* sampling in diagnose mode. alloc aux buffer */ + aux = zalloc(sizeof(*aux)); + if (aux == NULL) { + *err = -ENOMEM; + return NULL; + } + + aux->parse_snapshot_options = cpumsf_parse_snapshot_options; + aux->recording_options = cpumsf_recording_options; + aux->info_priv_size = cpumsf_info_priv_size; + aux->info_fill = cpumsf_info_fill; + aux->free = cpumsf_free; + aux->reference = cpumsf_reference; + + return aux; +} -- cgit v1.2.3 From f704ef44602fbf403e6722c7ed13f62d17e8cb20 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 19 Jan 2016 11:23:38 +0100 Subject: s390/perf: add support for perf_regs and libdw With support for perf_regs and libdw, you can record and report call graphs for user space programs. Simply invoke perf with the --call-graph=dwarf command line option. Signed-off-by: Heiko Carstens [brueckner: added dwfl_thread_state_register_pc() call] Signed-off-by: Hendrik Brueckner Reviewed-and-tested-by: Thomas Richter Signed-off-by: Martin Schwidefsky --- tools/perf/Makefile.config | 6 ++- tools/perf/arch/s390/include/perf_regs.h | 63 ++++++++++++++++++++++++++++++++ tools/perf/arch/s390/util/Build | 1 + tools/perf/arch/s390/util/dwarf-regs.c | 4 ++ tools/perf/arch/s390/util/unwind-libdw.c | 40 ++++++++++++++++++++ 5 files changed, 113 insertions(+), 1 deletion(-) create mode 100644 tools/perf/arch/s390/include/perf_regs.h create mode 100644 tools/perf/arch/s390/util/unwind-libdw.c (limited to 'tools/perf') diff --git a/tools/perf/Makefile.config b/tools/perf/Makefile.config index 63f534a0902f..ed65e82f034e 100644 --- a/tools/perf/Makefile.config +++ b/tools/perf/Makefile.config @@ -53,6 +53,10 @@ ifeq ($(SRCARCH),arm64) LIBUNWIND_LIBS = -lunwind -lunwind-aarch64 endif +ifeq ($(ARCH),s390) + NO_PERF_REGS := 0 +endif + ifeq ($(NO_PERF_REGS),0) $(call detected,CONFIG_PERF_REGS) endif @@ -61,7 +65,7 @@ endif # Disable it on all other architectures in case libdw unwind # support is detected in system. Add supported architectures # to the check. -ifneq ($(SRCARCH),$(filter $(SRCARCH),x86 arm powerpc)) +ifneq ($(SRCARCH),$(filter $(SRCARCH),x86 arm powerpc s390)) NO_LIBDW_DWARF_UNWIND := 1 endif diff --git a/tools/perf/arch/s390/include/perf_regs.h b/tools/perf/arch/s390/include/perf_regs.h new file mode 100644 index 000000000000..ac4b8c8c49e2 --- /dev/null +++ b/tools/perf/arch/s390/include/perf_regs.h @@ -0,0 +1,63 @@ +#ifndef ARCH_PERF_REGS_H +#define ARCH_PERF_REGS_H + +#include +#include +#include <../../../../arch/s390/include/uapi/asm/perf_regs.h> + +void perf_regs_load(u64 *regs); + +#define PERF_REGS_MASK ((1ULL << PERF_REG_S390_MAX) - 1) +#define PERF_REGS_MAX PERF_REG_S390_MAX +#define PERF_SAMPLE_REGS_ABI PERF_SAMPLE_REGS_ABI_64 + +#define PERF_REG_IP PERF_REG_S390_PC +#define PERF_REG_SP PERF_REG_S390_R15 + +static inline const char *perf_reg_name(int id) +{ + switch (id) { + case PERF_REG_S390_R0: + return "R0"; + case PERF_REG_S390_R1: + return "R1"; + case PERF_REG_S390_R2: + return "R2"; + case PERF_REG_S390_R3: + return "R3"; + case PERF_REG_S390_R4: + return "R4"; + case PERF_REG_S390_R5: + return "R5"; + case PERF_REG_S390_R6: + return "R6"; + case PERF_REG_S390_R7: + return "R7"; + case PERF_REG_S390_R8: + return "R8"; + case PERF_REG_S390_R9: + return "R9"; + case PERF_REG_S390_R10: + return "R10"; + case PERF_REG_S390_R11: + return "R11"; + case PERF_REG_S390_R12: + return "R12"; + case PERF_REG_S390_R13: + return "R13"; + case PERF_REG_S390_R14: + return "R14"; + case PERF_REG_S390_R15: + return "R15"; + case PERF_REG_S390_MASK: + return "MASK"; + case PERF_REG_S390_PC: + return "PC"; + default: + return NULL; + } + + return NULL; +} + +#endif /* ARCH_PERF_REGS_H */ diff --git a/tools/perf/arch/s390/util/Build b/tools/perf/arch/s390/util/Build index 397084382b23..4a233683c684 100644 --- a/tools/perf/arch/s390/util/Build +++ b/tools/perf/arch/s390/util/Build @@ -2,6 +2,7 @@ libperf-y += header.o libperf-y += kvm-stat.o libperf-$(CONFIG_DWARF) += dwarf-regs.o +libperf-$(CONFIG_LIBDW_DWARF_UNWIND) += unwind-libdw.o libperf-y += machine.o diff --git a/tools/perf/arch/s390/util/dwarf-regs.c b/tools/perf/arch/s390/util/dwarf-regs.c index 0dff5b2ed1e5..0cd7cba5d6ee 100644 --- a/tools/perf/arch/s390/util/dwarf-regs.c +++ b/tools/perf/arch/s390/util/dwarf-regs.c @@ -19,5 +19,9 @@ static const char *gpr_names[NUM_GPRS] = { const char *get_arch_regstr(unsigned int n) { + if (n == 64) + return "mask"; + if (n == 65) + return "pc"; return (n >= NUM_GPRS) ? NULL : gpr_names[n]; } diff --git a/tools/perf/arch/s390/util/unwind-libdw.c b/tools/perf/arch/s390/util/unwind-libdw.c new file mode 100644 index 000000000000..281bbb82402a --- /dev/null +++ b/tools/perf/arch/s390/util/unwind-libdw.c @@ -0,0 +1,40 @@ +#include +#include "../../util/unwind-libdw.h" +#include "../../util/perf_regs.h" +#include "../../util/event.h" + + +bool libdw__arch_set_initial_registers(Dwfl_Thread *thread, void *arg) +{ + struct unwind_info *ui = arg; + struct regs_dump *user_regs = &ui->sample->user_regs; + Dwarf_Word dwarf_regs[PERF_REG_S390_MAX]; + +#define REG(r) ({ \ + Dwarf_Word val = 0; \ + perf_reg_value(&val, user_regs, PERF_REG_S390_##r); \ + val; \ +}) + + dwarf_regs[0] = REG(R0); + dwarf_regs[1] = REG(R1); + dwarf_regs[2] = REG(R2); + dwarf_regs[3] = REG(R3); + dwarf_regs[4] = REG(R4); + dwarf_regs[5] = REG(R5); + dwarf_regs[6] = REG(R6); + dwarf_regs[7] = REG(R7); + dwarf_regs[8] = REG(R8); + dwarf_regs[9] = REG(R9); + dwarf_regs[10] = REG(R10); + dwarf_regs[11] = REG(R11); + dwarf_regs[12] = REG(R12); + dwarf_regs[13] = REG(R13); + dwarf_regs[14] = REG(R14); + dwarf_regs[15] = REG(R15); + dwarf_regs[16] = REG(MASK); + dwarf_regs[17] = REG(PC); + + dwfl_thread_state_register_pc(thread, dwarf_regs[17]); + return dwfl_thread_state_registers(thread, 0, 16, dwarf_regs); +} -- cgit v1.2.3 From a9fc2db0a8abf51c81122799c5ae4808f2324b6d Mon Sep 17 00:00:00 2001 From: Hendrik Brueckner Date: Wed, 8 Nov 2017 09:01:12 +0100 Subject: s390/perf: define common DWARF register string table Instead of defining DWARF register to string table in dwarf-regs-table.h and dwarf-regs.c, use a common table in dwarf-regs-table.h. Ensure that the DWARF register table is up-to-date with http://refspecs.linuxfoundation.org/ELF/zSeries/lzsabi0_s390/x1542.html. For unwinding with libdw, also ensure to correctly setup the DWARF register frame according to the register mappings. Currently, libdw supports up to 32 registers only. Suggested-by: Thomas Richter Signed-off-by: Hendrik Brueckner Reviewed-and-tested-by: Thomas Richter Signed-off-by: Martin Schwidefsky --- tools/perf/arch/s390/include/dwarf-regs-table.h | 71 +++++++++++++++++++++++-- tools/perf/arch/s390/util/dwarf-regs.c | 15 ++---- tools/perf/arch/s390/util/unwind-libdw.c | 15 ++++-- 3 files changed, 80 insertions(+), 21 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/arch/s390/include/dwarf-regs-table.h b/tools/perf/arch/s390/include/dwarf-regs-table.h index 792d4c277225..671553525f41 100644 --- a/tools/perf/arch/s390/include/dwarf-regs-table.h +++ b/tools/perf/arch/s390/include/dwarf-regs-table.h @@ -1,9 +1,72 @@ /* SPDX-License-Identifier: GPL-2.0 */ -#ifdef DEFINE_DWARF_REGSTR_TABLE -/* This is included in perf/util/dwarf-regs.c */ +#ifndef S390_DWARF_REGS_TABLE_H +#define S390_DWARF_REGS_TABLE_H -static const char * const s390_regstr_tbl[] = { +#define REG_DWARFNUM_NAME(reg, idx) [idx] = "%" #reg + +/* + * For reference, see DWARF register mapping: + * http://refspecs.linuxfoundation.org/ELF/zSeries/lzsabi0_s390/x1542.html + */ +static const char * const s390_dwarf_regs[] = { "%r0", "%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14", "%r15", + REG_DWARFNUM_NAME(f0, 16), + REG_DWARFNUM_NAME(f1, 20), + REG_DWARFNUM_NAME(f2, 17), + REG_DWARFNUM_NAME(f3, 21), + REG_DWARFNUM_NAME(f4, 18), + REG_DWARFNUM_NAME(f5, 22), + REG_DWARFNUM_NAME(f6, 19), + REG_DWARFNUM_NAME(f7, 23), + REG_DWARFNUM_NAME(f8, 24), + REG_DWARFNUM_NAME(f9, 28), + REG_DWARFNUM_NAME(f10, 25), + REG_DWARFNUM_NAME(f11, 29), + REG_DWARFNUM_NAME(f12, 26), + REG_DWARFNUM_NAME(f13, 30), + REG_DWARFNUM_NAME(f14, 27), + REG_DWARFNUM_NAME(f15, 31), + REG_DWARFNUM_NAME(c0, 32), + REG_DWARFNUM_NAME(c1, 33), + REG_DWARFNUM_NAME(c2, 34), + REG_DWARFNUM_NAME(c3, 35), + REG_DWARFNUM_NAME(c4, 36), + REG_DWARFNUM_NAME(c5, 37), + REG_DWARFNUM_NAME(c6, 38), + REG_DWARFNUM_NAME(c7, 39), + REG_DWARFNUM_NAME(c8, 40), + REG_DWARFNUM_NAME(c9, 41), + REG_DWARFNUM_NAME(c10, 42), + REG_DWARFNUM_NAME(c11, 43), + REG_DWARFNUM_NAME(c12, 44), + REG_DWARFNUM_NAME(c13, 45), + REG_DWARFNUM_NAME(c14, 46), + REG_DWARFNUM_NAME(c15, 47), + REG_DWARFNUM_NAME(a0, 48), + REG_DWARFNUM_NAME(a1, 49), + REG_DWARFNUM_NAME(a2, 50), + REG_DWARFNUM_NAME(a3, 51), + REG_DWARFNUM_NAME(a4, 52), + REG_DWARFNUM_NAME(a5, 53), + REG_DWARFNUM_NAME(a6, 54), + REG_DWARFNUM_NAME(a7, 55), + REG_DWARFNUM_NAME(a8, 56), + REG_DWARFNUM_NAME(a9, 57), + REG_DWARFNUM_NAME(a10, 58), + REG_DWARFNUM_NAME(a11, 59), + REG_DWARFNUM_NAME(a12, 60), + REG_DWARFNUM_NAME(a13, 61), + REG_DWARFNUM_NAME(a14, 62), + REG_DWARFNUM_NAME(a15, 63), + REG_DWARFNUM_NAME(pswm, 64), + REG_DWARFNUM_NAME(pswa, 65), }; -#endif + +#ifdef DEFINE_DWARF_REGSTR_TABLE +/* This is included in perf/util/dwarf-regs.c */ + +#define s390_regstr_tbl s390_dwarf_regs + +#endif /* DEFINE_DWARF_REGSTR_TABLE */ +#endif /* S390_DWARF_REGS_TABLE_H */ diff --git a/tools/perf/arch/s390/util/dwarf-regs.c b/tools/perf/arch/s390/util/dwarf-regs.c index 0cd7cba5d6ee..f47576ce13ea 100644 --- a/tools/perf/arch/s390/util/dwarf-regs.c +++ b/tools/perf/arch/s390/util/dwarf-regs.c @@ -9,19 +9,10 @@ #include #include - -#define NUM_GPRS 16 - -static const char *gpr_names[NUM_GPRS] = { - "%r0", "%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", - "%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14", "%r15", -}; +#include +#include "dwarf-regs-table.h" const char *get_arch_regstr(unsigned int n) { - if (n == 64) - return "mask"; - if (n == 65) - return "pc"; - return (n >= NUM_GPRS) ? NULL : gpr_names[n]; + return (n >= ARRAY_SIZE(s390_dwarf_regs)) ? NULL : s390_dwarf_regs[n]; } diff --git a/tools/perf/arch/s390/util/unwind-libdw.c b/tools/perf/arch/s390/util/unwind-libdw.c index 281bbb82402a..ba8d98b8b09e 100644 --- a/tools/perf/arch/s390/util/unwind-libdw.c +++ b/tools/perf/arch/s390/util/unwind-libdw.c @@ -1,21 +1,26 @@ +#include #include #include "../../util/unwind-libdw.h" #include "../../util/perf_regs.h" #include "../../util/event.h" +#include "dwarf-regs-table.h" bool libdw__arch_set_initial_registers(Dwfl_Thread *thread, void *arg) { struct unwind_info *ui = arg; struct regs_dump *user_regs = &ui->sample->user_regs; - Dwarf_Word dwarf_regs[PERF_REG_S390_MAX]; + Dwarf_Word dwarf_regs[ARRAY_SIZE(s390_dwarf_regs)]; #define REG(r) ({ \ Dwarf_Word val = 0; \ perf_reg_value(&val, user_regs, PERF_REG_S390_##r); \ val; \ }) - + /* + * For DWARF register mapping details, + * see also perf/arch/s390/include/dwarf-regs-table.h + */ dwarf_regs[0] = REG(R0); dwarf_regs[1] = REG(R1); dwarf_regs[2] = REG(R2); @@ -32,9 +37,9 @@ bool libdw__arch_set_initial_registers(Dwfl_Thread *thread, void *arg) dwarf_regs[13] = REG(R13); dwarf_regs[14] = REG(R14); dwarf_regs[15] = REG(R15); - dwarf_regs[16] = REG(MASK); - dwarf_regs[17] = REG(PC); + dwarf_regs[64] = REG(MASK); + dwarf_regs[65] = REG(PC); - dwfl_thread_state_register_pc(thread, dwarf_regs[17]); + dwfl_thread_state_register_pc(thread, dwarf_regs[65]); return dwfl_thread_state_registers(thread, 0, 16, dwarf_regs); } -- cgit v1.2.3 From de9954b75e992d26a67d83990e6219dcb6edc3bf Mon Sep 17 00:00:00 2001 From: Hendrik Brueckner Date: Wed, 8 Nov 2017 09:17:38 +0100 Subject: s390/perf: add perf register support for floating-point registers For correct unwinding of user space processes, the floating-point register contents are required. For example, leaf functions might use fp registers to temporarily store the return address. Signed-off-by: Hendrik Brueckner Reviewed-and-tested-by: Thomas Richter Signed-off-by: Martin Schwidefsky --- tools/perf/arch/s390/include/perf_regs.h | 32 ++++++++++++++++++++++++++++++++ tools/perf/arch/s390/util/unwind-libdw.c | 20 +++++++++++++++++++- 2 files changed, 51 insertions(+), 1 deletion(-) (limited to 'tools/perf') diff --git a/tools/perf/arch/s390/include/perf_regs.h b/tools/perf/arch/s390/include/perf_regs.h index ac4b8c8c49e2..d2df54a6bc5a 100644 --- a/tools/perf/arch/s390/include/perf_regs.h +++ b/tools/perf/arch/s390/include/perf_regs.h @@ -49,6 +49,38 @@ static inline const char *perf_reg_name(int id) return "R14"; case PERF_REG_S390_R15: return "R15"; + case PERF_REG_S390_FP0: + return "FP0"; + case PERF_REG_S390_FP1: + return "FP1"; + case PERF_REG_S390_FP2: + return "FP2"; + case PERF_REG_S390_FP3: + return "FP3"; + case PERF_REG_S390_FP4: + return "FP4"; + case PERF_REG_S390_FP5: + return "FP5"; + case PERF_REG_S390_FP6: + return "FP6"; + case PERF_REG_S390_FP7: + return "FP7"; + case PERF_REG_S390_FP8: + return "FP8"; + case PERF_REG_S390_FP9: + return "FP9"; + case PERF_REG_S390_FP10: + return "FP10"; + case PERF_REG_S390_FP11: + return "FP11"; + case PERF_REG_S390_FP12: + return "FP12"; + case PERF_REG_S390_FP13: + return "FP13"; + case PERF_REG_S390_FP14: + return "FP14"; + case PERF_REG_S390_FP15: + return "FP15"; case PERF_REG_S390_MASK: return "MASK"; case PERF_REG_S390_PC: diff --git a/tools/perf/arch/s390/util/unwind-libdw.c b/tools/perf/arch/s390/util/unwind-libdw.c index ba8d98b8b09e..387c698cdd1b 100644 --- a/tools/perf/arch/s390/util/unwind-libdw.c +++ b/tools/perf/arch/s390/util/unwind-libdw.c @@ -37,9 +37,27 @@ bool libdw__arch_set_initial_registers(Dwfl_Thread *thread, void *arg) dwarf_regs[13] = REG(R13); dwarf_regs[14] = REG(R14); dwarf_regs[15] = REG(R15); + + dwarf_regs[16] = REG(FP0); + dwarf_regs[17] = REG(FP2); + dwarf_regs[18] = REG(FP4); + dwarf_regs[19] = REG(FP6); + dwarf_regs[20] = REG(FP1); + dwarf_regs[21] = REG(FP3); + dwarf_regs[22] = REG(FP5); + dwarf_regs[23] = REG(FP7); + dwarf_regs[24] = REG(FP8); + dwarf_regs[25] = REG(FP10); + dwarf_regs[26] = REG(FP12); + dwarf_regs[27] = REG(FP14); + dwarf_regs[28] = REG(FP9); + dwarf_regs[29] = REG(FP11); + dwarf_regs[30] = REG(FP13); + dwarf_regs[31] = REG(FP15); + dwarf_regs[64] = REG(MASK); dwarf_regs[65] = REG(PC); dwfl_thread_state_register_pc(thread, dwarf_regs[65]); - return dwfl_thread_state_registers(thread, 0, 16, dwarf_regs); + return dwfl_thread_state_registers(thread, 0, 32, dwarf_regs); } -- cgit v1.2.3