From f8b8db77b0cc36670ef4ed6bc31e64537ffa197e Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Mon, 7 Nov 2005 01:01:39 -0800 Subject: [PATCH] unexport nr_swap_pages I didn't find any possible modular usage in the kernel. Signed-off-by: Adrian Bunk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 1 - 1 file changed, 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 2dbdd98426fd..ff81b5c65511 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -64,7 +64,6 @@ long nr_swap_pages; int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { 256, 32 }; EXPORT_SYMBOL(totalram_pages); -EXPORT_SYMBOL(nr_swap_pages); /* * Used by page_zone() to look up the address of the struct zone whose -- cgit v1.2.3 From 6b482c6779daaa893b277fc9b70767a7c2e7c5eb Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Thu, 10 Nov 2005 15:45:56 -0500 Subject: [PATCH] Don't print per-cpu vm stats for offline cpus. I just hit a page allocation error on a kernel configured to support 64 CPUs. It spewed 60 completely useless unnecessary lines of info. Signed-off-by: Dave Jones Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index ff81b5c65511..987225bdd661 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1330,7 +1330,7 @@ void show_free_areas(void) } else printk("\n"); - for_each_cpu(cpu) { + for_each_online_cpu(cpu) { struct per_cpu_pageset *pageset; pageset = zone_pcp(zone, cpu); -- cgit v1.2.3 From 885036d32f5d3c427c3e2b385b5a5503805e3e52 Mon Sep 17 00:00:00 2001 From: Kirill Korotaev Date: Sun, 13 Nov 2005 16:06:41 -0800 Subject: [PATCH] mm: __GFP_NOFAIL fix In __alloc_pages(): if ((p->flags & (PF_MEMALLOC | PF_MEMDIE)) && !in_interrupt()) { /* go through the zonelist yet again, ignoring mins */ for (i = 0; zones[i] != NULL; i++) { struct zone *z = zones[i]; page = buffered_rmqueue(z, order, gfp_mask); if (page) { zone_statistics(zonelist, z); goto got_pg; } } goto nopage; <<<< HERE!!! FAIL... } kswapd (which has PF_MEMALLOC flag) can fail to allocate memory even when it allocates it with __GFP_NOFAIL flag. Signed-Off-By: Pavel Emelianov Signed-Off-By: Denis Lunev Signed-Off-By: Kirill Korotaev Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 987225bdd661..b37dc0f78d07 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -895,6 +895,7 @@ zone_reclaim_retry: if (((p->flags & PF_MEMALLOC) || unlikely(test_thread_flag(TIF_MEMDIE))) && !in_interrupt()) { if (!(gfp_mask & __GFP_NOMEMALLOC)) { +nofail_alloc: /* go through the zonelist yet again, ignoring mins */ for (i = 0; (z = zones[i]) != NULL; i++) { if (!cpuset_zone_allowed(z, gfp_mask)) @@ -903,6 +904,10 @@ zone_reclaim_retry: if (page) goto got_pg; } + if (gfp_mask & __GFP_NOFAIL) { + blk_congestion_wait(WRITE, HZ/50); + goto nofail_alloc; + } } goto nopage; } -- cgit v1.2.3 From 7fb1d9fca5c6e3b06773b69165a73f3fb786b8ee Mon Sep 17 00:00:00 2001 From: Rohit Seth Date: Sun, 13 Nov 2005 16:06:43 -0800 Subject: [PATCH] mm: __alloc_pages cleanup Clean up of __alloc_pages. Restoration of previous behaviour, plus further cleanups by introducing an 'alloc_flags', removing the last of should_reclaim_zone. Signed-off-by: Rohit Seth Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mmzone.h | 2 +- mm/page_alloc.c | 195 +++++++++++++++++++++---------------------------- mm/vmscan.c | 6 +- 3 files changed, 89 insertions(+), 114 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index f5fa3082fd6a..6cfb114a0c34 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -329,7 +329,7 @@ void get_zone_counts(unsigned long *active, unsigned long *inactive, void build_all_zonelists(void); void wakeup_kswapd(struct zone *zone, int order); int zone_watermark_ok(struct zone *z, int order, unsigned long mark, - int alloc_type, int can_try_harder, gfp_t gfp_high); + int classzone_idx, int alloc_flags); #ifdef CONFIG_HAVE_MEMORY_PRESENT void memory_present(int nid, unsigned long start, unsigned long end); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index b37dc0f78d07..845b91749a42 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -732,9 +732,7 @@ buffered_rmqueue(struct zone *zone, int order, gfp_t gfp_flags) } local_irq_restore(flags); put_cpu(); - } - - if (page == NULL) { + } else { spin_lock_irqsave(&zone->lock, flags); page = __rmqueue(zone, order); spin_unlock_irqrestore(&zone->lock, flags); @@ -754,20 +752,25 @@ buffered_rmqueue(struct zone *zone, int order, gfp_t gfp_flags) return page; } +#define ALLOC_NO_WATERMARKS 0x01 /* don't check watermarks at all */ +#define ALLOC_HARDER 0x02 /* try to alloc harder */ +#define ALLOC_HIGH 0x04 /* __GFP_HIGH set */ +#define ALLOC_CPUSET 0x08 /* check for correct cpuset */ + /* * Return 1 if free pages are above 'mark'. This takes into account the order * of the allocation. */ int zone_watermark_ok(struct zone *z, int order, unsigned long mark, - int classzone_idx, int can_try_harder, gfp_t gfp_high) + int classzone_idx, int alloc_flags) { /* free_pages my go negative - that's OK */ long min = mark, free_pages = z->free_pages - (1 << order) + 1; int o; - if (gfp_high) + if (alloc_flags & ALLOC_HIGH) min -= min / 2; - if (can_try_harder) + if (alloc_flags & ALLOC_HARDER) min -= min / 4; if (free_pages <= min + z->lowmem_reserve[classzone_idx]) @@ -785,14 +788,40 @@ int zone_watermark_ok(struct zone *z, int order, unsigned long mark, return 1; } -static inline int -should_reclaim_zone(struct zone *z, gfp_t gfp_mask) +/* + * get_page_from_freeliest goes through the zonelist trying to allocate + * a page. + */ +static struct page * +get_page_from_freelist(gfp_t gfp_mask, unsigned int order, + struct zonelist *zonelist, int alloc_flags) { - if (!z->reclaim_pages) - return 0; - if (gfp_mask & __GFP_NORECLAIM) - return 0; - return 1; + struct zone **z = zonelist->zones; + struct page *page = NULL; + int classzone_idx = zone_idx(*z); + + /* + * Go through the zonelist once, looking for a zone with enough free. + * See also cpuset_zone_allowed() comment in kernel/cpuset.c. + */ + do { + if ((alloc_flags & ALLOC_CPUSET) && + !cpuset_zone_allowed(*z, gfp_mask)) + continue; + + if (!(alloc_flags & ALLOC_NO_WATERMARKS)) { + if (!zone_watermark_ok(*z, order, (*z)->pages_low, + classzone_idx, alloc_flags)) + continue; + } + + page = buffered_rmqueue(*z, order, gfp_mask); + if (page) { + zone_statistics(zonelist, *z); + break; + } + } while (*(++z) != NULL); + return page; } /* @@ -803,92 +832,60 @@ __alloc_pages(gfp_t gfp_mask, unsigned int order, struct zonelist *zonelist) { const gfp_t wait = gfp_mask & __GFP_WAIT; - struct zone **zones, *z; + struct zone **z; struct page *page; struct reclaim_state reclaim_state; struct task_struct *p = current; - int i; - int classzone_idx; int do_retry; - int can_try_harder; + int alloc_flags; int did_some_progress; might_sleep_if(wait); - /* - * The caller may dip into page reserves a bit more if the caller - * cannot run direct reclaim, or is the caller has realtime scheduling - * policy - */ - can_try_harder = (unlikely(rt_task(p)) && !in_interrupt()) || !wait; - - zones = zonelist->zones; /* the list of zones suitable for gfp_mask */ + z = zonelist->zones; /* the list of zones suitable for gfp_mask */ - if (unlikely(zones[0] == NULL)) { + if (unlikely(*z == NULL)) { /* Should this ever happen?? */ return NULL; } +restart: + page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, + zonelist, ALLOC_CPUSET); + if (page) + goto got_pg; - classzone_idx = zone_idx(zones[0]); + do + wakeup_kswapd(*z, order); + while (*(++z)); -restart: /* - * Go through the zonelist once, looking for a zone with enough free. - * See also cpuset_zone_allowed() comment in kernel/cpuset.c. + * OK, we're below the kswapd watermark and have kicked background + * reclaim. Now things get more complex, so set up alloc_flags according + * to how we want to proceed. + * + * The caller may dip into page reserves a bit more if the caller + * cannot run direct reclaim, or if the caller has realtime scheduling + * policy. */ - for (i = 0; (z = zones[i]) != NULL; i++) { - int do_reclaim = should_reclaim_zone(z, gfp_mask); - - if (!cpuset_zone_allowed(z, __GFP_HARDWALL)) - continue; - - /* - * If the zone is to attempt early page reclaim then this loop - * will try to reclaim pages and check the watermark a second - * time before giving up and falling back to the next zone. - */ -zone_reclaim_retry: - if (!zone_watermark_ok(z, order, z->pages_low, - classzone_idx, 0, 0)) { - if (!do_reclaim) - continue; - else { - zone_reclaim(z, gfp_mask, order); - /* Only try reclaim once */ - do_reclaim = 0; - goto zone_reclaim_retry; - } - } - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } - - for (i = 0; (z = zones[i]) != NULL; i++) - wakeup_kswapd(z, order); + alloc_flags = 0; + if ((unlikely(rt_task(p)) && !in_interrupt()) || !wait) + alloc_flags |= ALLOC_HARDER; + if (gfp_mask & __GFP_HIGH) + alloc_flags |= ALLOC_HIGH; + if (wait) + alloc_flags |= ALLOC_CPUSET; /* * Go through the zonelist again. Let __GFP_HIGH and allocations - * coming from realtime tasks to go deeper into reserves + * coming from realtime tasks go deeper into reserves. * * This is the last chance, in general, before the goto nopage. * Ignore cpuset if GFP_ATOMIC (!wait) rather than fail alloc. * See also cpuset_zone_allowed() comment in kernel/cpuset.c. */ - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!zone_watermark_ok(z, order, z->pages_min, - classzone_idx, can_try_harder, - gfp_mask & __GFP_HIGH)) - continue; - - if (wait && !cpuset_zone_allowed(z, gfp_mask)) - continue; - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask, order, zonelist, alloc_flags); + if (page) + goto got_pg; /* This allocation should allow future memory freeing. */ @@ -897,13 +894,10 @@ zone_reclaim_retry: if (!(gfp_mask & __GFP_NOMEMALLOC)) { nofail_alloc: /* go through the zonelist yet again, ignoring mins */ - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!cpuset_zone_allowed(z, gfp_mask)) - continue; - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask, order, + zonelist, ALLOC_NO_WATERMARKS|ALLOC_CPUSET); + if (page) + goto got_pg; if (gfp_mask & __GFP_NOFAIL) { blk_congestion_wait(WRITE, HZ/50); goto nofail_alloc; @@ -924,7 +918,7 @@ rebalance: reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; - did_some_progress = try_to_free_pages(zones, gfp_mask); + did_some_progress = try_to_free_pages(zonelist->zones, gfp_mask); p->reclaim_state = NULL; p->flags &= ~PF_MEMALLOC; @@ -932,19 +926,10 @@ rebalance: cond_resched(); if (likely(did_some_progress)) { - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!zone_watermark_ok(z, order, z->pages_min, - classzone_idx, can_try_harder, - gfp_mask & __GFP_HIGH)) - continue; - - if (!cpuset_zone_allowed(z, gfp_mask)) - continue; - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask, order, + zonelist, alloc_flags); + if (page) + goto got_pg; } else if ((gfp_mask & __GFP_FS) && !(gfp_mask & __GFP_NORETRY)) { /* * Go through the zonelist yet one more time, keep @@ -952,18 +937,10 @@ rebalance: * a parallel oom killing, we must fail if we're still * under heavy pressure. */ - for (i = 0; (z = zones[i]) != NULL; i++) { - if (!zone_watermark_ok(z, order, z->pages_high, - classzone_idx, 0, 0)) - continue; - - if (!cpuset_zone_allowed(z, __GFP_HARDWALL)) - continue; - - page = buffered_rmqueue(z, order, gfp_mask); - if (page) - goto got_pg; - } + page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, + zonelist, ALLOC_CPUSET); + if (page) + goto got_pg; out_of_memory(gfp_mask, order); goto restart; @@ -996,9 +973,7 @@ nopage: dump_stack(); show_mem(); } - return NULL; got_pg: - zone_statistics(zonelist, z); return page; } diff --git a/mm/vmscan.c b/mm/vmscan.c index 135bf8ca96ee..28130541270f 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -1074,7 +1074,7 @@ loop_again: continue; if (!zone_watermark_ok(zone, order, - zone->pages_high, 0, 0, 0)) { + zone->pages_high, 0, 0)) { end_zone = i; goto scan; } @@ -1111,7 +1111,7 @@ scan: if (nr_pages == 0) { /* Not software suspend */ if (!zone_watermark_ok(zone, order, - zone->pages_high, end_zone, 0, 0)) + zone->pages_high, end_zone, 0)) all_zones_ok = 0; } zone->temp_priority = priority; @@ -1259,7 +1259,7 @@ void wakeup_kswapd(struct zone *zone, int order) return; pgdat = zone->zone_pgdat; - if (zone_watermark_ok(zone, order, zone->pages_low, 0, 0, 0)) + if (zone_watermark_ok(zone, order, zone->pages_low, 0, 0)) return; if (pgdat->kswapd_max_order < order) pgdat->kswapd_max_order = order; -- cgit v1.2.3 From 669ed17521b9b78cdbeac8a53c30599aca9527ce Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Sun, 13 Nov 2005 16:06:45 -0800 Subject: [PATCH] mm: highmem watermarks The pages_high - pages_low and pages_low - pages_min deltas are the asynch reclaim watermarks. As such, the should be in the same ratios as any other zone for highmem zones. It is the pages_min - 0 delta which is the PF_MEMALLOC reserve, and this is the region that isn't very useful for highmem. This patch ensures highmem systems have similar characteristics as non highmem ones with the same amount of memory, and also that highmem zones get similar reclaim pressures to other zones. Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 845b91749a42..3c5cf664abd2 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -2397,13 +2397,18 @@ void setup_per_zone_pages_min(void) } for_each_zone(zone) { + unsigned long tmp; spin_lock_irqsave(&zone->lru_lock, flags); + tmp = (pages_min * zone->present_pages) / lowmem_pages; if (is_highmem(zone)) { /* - * Often, highmem doesn't need to reserve any pages. - * But the pages_min/low/high values are also used for - * batching up page reclaim activity so we need a - * decent value here. + * __GFP_HIGH and PF_MEMALLOC allocations usually don't + * need highmem pages, so cap pages_min to a small + * value here. + * + * The (pages_high-pages_low) and (pages_low-pages_min) + * deltas controls asynch page reclaim, and so should + * not be capped for highmem. */ int min_pages; @@ -2414,19 +2419,15 @@ void setup_per_zone_pages_min(void) min_pages = 128; zone->pages_min = min_pages; } else { - /* if it's a lowmem zone, reserve a number of pages + /* + * If it's a lowmem zone, reserve a number of pages * proportionate to the zone's size. */ - zone->pages_min = (pages_min * zone->present_pages) / - lowmem_pages; + zone->pages_min = tmp; } - /* - * When interpreting these watermarks, just keep in mind that: - * zone->pages_min == (zone->pages_min * 4) / 4; - */ - zone->pages_low = (zone->pages_min * 5) / 4; - zone->pages_high = (zone->pages_min * 6) / 4; + zone->pages_low = zone->pages_min + tmp / 4; + zone->pages_high = zone->pages_min + tmp / 2; spin_unlock_irqrestore(&zone->lru_lock, flags); } } -- cgit v1.2.3 From a2f1b424900715ed9d1699c3bb88a434a2b42bc0 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 5 Nov 2005 17:25:53 +0100 Subject: [PATCH] x86_64: Add 4GB DMA32 zone Add a new 4GB GFP_DMA32 zone between the GFP_DMA and GFP_NORMAL zones. As a bit of historical background: when the x86-64 port was originally designed we had some discussion if we should use a 16MB DMA zone like i386 or a 4GB DMA zone like IA64 or both. Both was ruled out at this point because it was in early 2.4 when VM is still quite shakey and had bad troubles even dealing with one DMA zone. We settled on the 16MB DMA zone mainly because we worried about older soundcards and the floppy. But this has always caused problems since then because device drivers had trouble getting enough DMA able memory. These days the VM works much better and the wide use of NUMA has proven it can deal with many zones successfully. So this patch adds both zones. This helps drivers who need a lot of memory below 4GB because their hardware is not accessing more (graphic drivers - proprietary and free ones, video frame buffer drivers, sound drivers etc.). Previously they could only use IOMMU+16MB GFP_DMA, which was not enough memory. Another common problem is that hardware who has full memory addressing for >4GB misses it for some control structures in memory (like transmit rings or other metadata). They tended to allocate memory in the 16MB GFP_DMA or the IOMMU/swiotlb then using pci_alloc_consistent, but that can tie up a lot of precious 16MB GFPDMA/IOMMU/swiotlb memory (even on AMD systems the IOMMU tends to be quite small) especially if you have many devices. With the new zone pci_alloc_consistent can just put this stuff into memory below 4GB which works better. One argument was still if the zone should be 4GB or 2GB. The main motivation for 2GB would be an unnamed not so unpopular hardware raid controller (mostly found in older machines from a particular four letter company) who has a strange 2GB restriction in firmware. But that one works ok with swiotlb/IOMMU anyways, so it doesn't really need GFP_DMA32. I chose 4GB to be compatible with IA64 and because it seems to be the most common restriction. The new zone is so far added only for x86-64. For other architectures who don't set up this new zone nothing changes. Architectures can set a compatibility define in Kconfig CONFIG_DMA_IS_DMA32 that will define GFP_DMA32 as GFP_DMA. Otherwise it's a nop because on 32bit architectures it's normally not needed because GFP_NORMAL (=0) is DMA able enough. One problem is still that GFP_DMA means different things on different architectures. e.g. some drivers used to have #ifdef ia64 use GFP_DMA (trusting it to be 4GB) #elif __x86_64__ (use other hacks like the swiotlb because 16MB is not enough) ... . This was quite ugly and is now obsolete. These should be now converted to use GFP_DMA32 unconditionally. I haven't done this yet. Or best only use pci_alloc_consistent/dma_alloc_coherent which will use GFP_DMA32 transparently. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/mm/init.c | 65 ++++++++++++++++++++++++++++++---------------- arch/x86_64/mm/numa.c | 25 ++++-------------- include/asm-x86_64/dma.h | 11 ++++++-- include/asm-x86_64/proto.h | 2 ++ include/linux/gfp.h | 11 ++++++++ include/linux/mmzone.h | 16 +++++++----- mm/page_alloc.c | 15 ++++++++--- 7 files changed, 90 insertions(+), 55 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/arch/x86_64/mm/init.c b/arch/x86_64/mm/init.c index e60a1a848de8..a1ad4cc423a7 100644 --- a/arch/x86_64/mm/init.c +++ b/arch/x86_64/mm/init.c @@ -318,32 +318,51 @@ void zap_low_mappings(void) flush_tlb_all(); } +/* Compute zone sizes for the DMA and DMA32 zones in a node. */ +__init void +size_zones(unsigned long *z, unsigned long *h, + unsigned long start_pfn, unsigned long end_pfn) +{ + int i; + unsigned long w; + + for (i = 0; i < MAX_NR_ZONES; i++) + z[i] = 0; + + if (start_pfn < MAX_DMA_PFN) + z[ZONE_DMA] = MAX_DMA_PFN - start_pfn; + if (start_pfn < MAX_DMA32_PFN) { + unsigned long dma32_pfn = MAX_DMA32_PFN; + if (dma32_pfn > end_pfn) + dma32_pfn = end_pfn; + z[ZONE_DMA32] = dma32_pfn - start_pfn; + } + z[ZONE_NORMAL] = end_pfn - start_pfn; + + /* Remove lower zones from higher ones. */ + w = 0; + for (i = 0; i < MAX_NR_ZONES; i++) { + if (z[i]) + z[i] -= w; + w += z[i]; + } + + /* Compute holes */ + w = 0; + for (i = 0; i < MAX_NR_ZONES; i++) { + unsigned long s = w; + w += z[i]; + h[i] = e820_hole_size(s, w); + } +} + #ifndef CONFIG_NUMA void __init paging_init(void) { - { - unsigned long zones_size[MAX_NR_ZONES]; - unsigned long holes[MAX_NR_ZONES]; - unsigned int max_dma; - - memset(zones_size, 0, sizeof(zones_size)); - memset(holes, 0, sizeof(holes)); - - max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT; - - if (end_pfn < max_dma) { - zones_size[ZONE_DMA] = end_pfn; - holes[ZONE_DMA] = e820_hole_size(0, end_pfn); - } else { - zones_size[ZONE_DMA] = max_dma; - holes[ZONE_DMA] = e820_hole_size(0, max_dma); - zones_size[ZONE_NORMAL] = end_pfn - max_dma; - holes[ZONE_NORMAL] = e820_hole_size(max_dma, end_pfn); - } - free_area_init_node(0, NODE_DATA(0), zones_size, - __pa(PAGE_OFFSET) >> PAGE_SHIFT, holes); - } - return; + unsigned long zones[MAX_NR_ZONES], holes[MAX_NR_ZONES]; + size_zones(zones, holes, 0, end_pfn); + free_area_init_node(0, NODE_DATA(0), zones, + __pa(PAGE_OFFSET) >> PAGE_SHIFT, holes); } #endif diff --git a/arch/x86_64/mm/numa.c b/arch/x86_64/mm/numa.c index 214803821001..18e86e2eac2d 100644 --- a/arch/x86_64/mm/numa.c +++ b/arch/x86_64/mm/numa.c @@ -132,29 +132,14 @@ void __init setup_node_zones(int nodeid) unsigned long start_pfn, end_pfn; unsigned long zones[MAX_NR_ZONES]; unsigned long holes[MAX_NR_ZONES]; - unsigned long dma_end_pfn; - memset(zones, 0, sizeof(unsigned long) * MAX_NR_ZONES); - memset(holes, 0, sizeof(unsigned long) * MAX_NR_ZONES); + start_pfn = node_start_pfn(nodeid); + end_pfn = node_end_pfn(nodeid); - start_pfn = node_start_pfn(nodeid); - end_pfn = node_end_pfn(nodeid); + Dprintk(KERN_INFO "setting up node %d %lx-%lx\n", + nodeid, start_pfn, end_pfn); - Dprintk(KERN_INFO "setting up node %d %lx-%lx\n", nodeid, start_pfn, end_pfn); - - /* All nodes > 0 have a zero length zone DMA */ - dma_end_pfn = __pa(MAX_DMA_ADDRESS) >> PAGE_SHIFT; - if (start_pfn < dma_end_pfn) { - zones[ZONE_DMA] = dma_end_pfn - start_pfn; - holes[ZONE_DMA] = e820_hole_size(start_pfn, dma_end_pfn); - zones[ZONE_NORMAL] = end_pfn - dma_end_pfn; - holes[ZONE_NORMAL] = e820_hole_size(dma_end_pfn, end_pfn); - - } else { - zones[ZONE_NORMAL] = end_pfn - start_pfn; - holes[ZONE_NORMAL] = e820_hole_size(start_pfn, end_pfn); - } - + size_zones(zones, holes, start_pfn, end_pfn); free_area_init_node(nodeid, NODE_DATA(nodeid), zones, start_pfn, holes); } diff --git a/include/asm-x86_64/dma.h b/include/asm-x86_64/dma.h index 16fa3a064d0c..6f2a817b6a7c 100644 --- a/include/asm-x86_64/dma.h +++ b/include/asm-x86_64/dma.h @@ -72,8 +72,15 @@ #define MAX_DMA_CHANNELS 8 -/* The maximum address that we can perform a DMA transfer to on this platform */ -#define MAX_DMA_ADDRESS (PAGE_OFFSET+0x1000000) + +/* 16MB ISA DMA zone */ +#define MAX_DMA_PFN ((16*1024*1024) >> PAGE_SHIFT) + +/* 4GB broken PCI/AGP hardware bus master zone */ +#define MAX_DMA32_PFN ((4UL*1024*1024*1024) >> PAGE_SHIFT) + +/* Compat define for old dma zone */ +#define MAX_DMA_ADDRESS ((unsigned long)__va(MAX_DMA_PFN << PAGE_SHIFT)) /* 8237 DMA controllers */ #define IO_DMA1_BASE 0x00 /* 8 bit slave DMA, channels 0..3 */ diff --git a/include/asm-x86_64/proto.h b/include/asm-x86_64/proto.h index dbb37b0adb43..c251152a0658 100644 --- a/include/asm-x86_64/proto.h +++ b/include/asm-x86_64/proto.h @@ -22,6 +22,8 @@ extern void mtrr_bp_init(void); #define mtrr_bp_init() do {} while (0) #endif extern void init_memory_mapping(unsigned long start, unsigned long end); +extern void size_zones(unsigned long *z, unsigned long *h, + unsigned long start_pfn, unsigned long end_pfn); extern void system_call(void); extern int kernel_syscall(void); diff --git a/include/linux/gfp.h b/include/linux/gfp.h index c3779432a723..4351e6bb5a79 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -14,6 +14,13 @@ struct vm_area_struct; /* Zone modifiers in GFP_ZONEMASK (see linux/mmzone.h - low two bits) */ #define __GFP_DMA ((__force gfp_t)0x01u) #define __GFP_HIGHMEM ((__force gfp_t)0x02u) +#ifdef CONFIG_DMA_IS_DMA32 +#define __GFP_DMA32 ((__force gfp_t)0x01) /* ZONE_DMA is ZONE_DMA32 */ +#elif BITS_PER_LONG < 64 +#define __GFP_DMA32 ((__force gfp_t)0x00) /* ZONE_NORMAL is ZONE_DMA32 */ +#else +#define __GFP_DMA32 ((__force gfp_t)0x04) /* Has own ZONE_DMA32 */ +#endif /* * Action modifiers - doesn't change the zoning @@ -64,6 +71,10 @@ struct vm_area_struct; #define GFP_DMA __GFP_DMA +/* 4GB DMA on some platforms */ +#define GFP_DMA32 __GFP_DMA32 + + #define gfp_zone(mask) ((__force int)((mask) & (__force gfp_t)GFP_ZONEMASK)) /* diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index f5fa3082fd6a..da7a829f8561 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -71,10 +71,11 @@ struct per_cpu_pageset { #endif #define ZONE_DMA 0 -#define ZONE_NORMAL 1 -#define ZONE_HIGHMEM 2 +#define ZONE_DMA32 1 +#define ZONE_NORMAL 2 +#define ZONE_HIGHMEM 3 -#define MAX_NR_ZONES 3 /* Sync this with ZONES_SHIFT */ +#define MAX_NR_ZONES 4 /* Sync this with ZONES_SHIFT */ #define ZONES_SHIFT 2 /* ceil(log2(MAX_NR_ZONES)) */ @@ -108,9 +109,10 @@ struct per_cpu_pageset { /* * On machines where it is needed (eg PCs) we divide physical memory - * into multiple physical zones. On a PC we have 3 zones: + * into multiple physical zones. On a PC we have 4 zones: * * ZONE_DMA < 16 MB ISA DMA capable memory + * ZONE_DMA32 0 MB Empty * ZONE_NORMAL 16-896 MB direct mapped by the kernel * ZONE_HIGHMEM > 896 MB only page cache and user processes */ @@ -455,10 +457,10 @@ extern struct pglist_data contig_page_data; #if BITS_PER_LONG == 32 || defined(ARCH_HAS_ATOMIC_UNSIGNED) /* - * with 32 bit page->flags field, we reserve 8 bits for node/zone info. - * there are 3 zones (2 bits) and this leaves 8-2=6 bits for nodes. + * with 32 bit page->flags field, we reserve 9 bits for node/zone info. + * there are 4 zones (3 bits) and this leaves 9-3=6 bits for nodes. */ -#define FLAGS_RESERVED 8 +#define FLAGS_RESERVED 9 #elif BITS_PER_LONG == 64 /* diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 2dbdd98426fd..9b43511dbefd 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -60,8 +60,11 @@ long nr_swap_pages; * NORMAL allocation will leave 784M/256 of ram reserved in the ZONE_DMA * HIGHMEM allocation will leave 224M/32 of ram reserved in ZONE_NORMAL * HIGHMEM allocation will (224M+784M)/256 of ram reserved in ZONE_DMA + * + * TBD: should special case ZONE_DMA32 machines here - in those we normally + * don't need any ZONE_NORMAL reservation */ -int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { 256, 32 }; +int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { 256, 256, 32 }; EXPORT_SYMBOL(totalram_pages); EXPORT_SYMBOL(nr_swap_pages); @@ -73,7 +76,7 @@ EXPORT_SYMBOL(nr_swap_pages); struct zone *zone_table[1 << ZONETABLE_SHIFT] __read_mostly; EXPORT_SYMBOL(zone_table); -static char *zone_names[MAX_NR_ZONES] = { "DMA", "Normal", "HighMem" }; +static char *zone_names[MAX_NR_ZONES] = { "DMA", "DMA32", "Normal", "HighMem" }; int min_free_kbytes = 1024; unsigned long __initdata nr_kernel_pages; @@ -1442,6 +1445,10 @@ static int __init build_zonelists_node(pg_data_t *pgdat, struct zonelist *zoneli zone = pgdat->node_zones + ZONE_NORMAL; if (zone->present_pages) zonelist->zones[j++] = zone; + case ZONE_DMA32: + zone = pgdat->node_zones + ZONE_DMA32; + if (zone->present_pages) + zonelist->zones[j++] = zone; case ZONE_DMA: zone = pgdat->node_zones + ZONE_DMA; if (zone->present_pages) @@ -1456,6 +1463,8 @@ static inline int highest_zone(int zone_bits) int res = ZONE_NORMAL; if (zone_bits & (__force int)__GFP_HIGHMEM) res = ZONE_HIGHMEM; + if (zone_bits & (__force int)__GFP_DMA32) + res = ZONE_DMA32; if (zone_bits & (__force int)__GFP_DMA) res = ZONE_DMA; return res; @@ -1976,7 +1985,7 @@ static void __init free_area_init_core(struct pglist_data *pgdat, if (zholes_size) realsize -= zholes_size[j]; - if (j == ZONE_DMA || j == ZONE_NORMAL) + if (j < ZONE_HIGHMEM) nr_kernel_pages += realsize; nr_all_pages += realsize; -- cgit v1.2.3 From b0d41693217b3bb5b837940dc7465e82a9d49476 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 5 Nov 2005 17:25:53 +0100 Subject: [PATCH] x86_64: When cpu_up fails clean up page allocator properly Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 9b43511dbefd..e8810a7e0264 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1876,11 +1876,10 @@ static int __devinit pageset_cpuup_callback(struct notifier_block *nfb, if (process_zones(cpu)) ret = NOTIFY_BAD; break; -#ifdef CONFIG_HOTPLUG_CPU + case CPU_UP_CANCELED: case CPU_DEAD: free_zone_pagesets(cpu); break; -#endif default: break; } -- cgit v1.2.3 From 07808b74e7dab1aa385e698795875337d72daf7d Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 5 Nov 2005 17:25:53 +0100 Subject: [PATCH] x86_64: Remove obsolete ARCH_HAS_ATOMIC_UNSIGNED and page_flags_t Has been introduced for x86-64 at some point to save memory in struct page, but has been obsolete for some time. Just remove it. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- include/linux/mm.h | 10 ++-------- include/linux/mmzone.h | 2 +- mm/filemap.c | 2 +- mm/page_alloc.c | 2 +- 4 files changed, 5 insertions(+), 11 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/linux/mm.h b/include/linux/mm.h index 5c1fb0a2e806..23fad4dae23c 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -206,12 +206,6 @@ struct vm_operations_struct { struct mmu_gather; struct inode; -#ifdef ARCH_HAS_ATOMIC_UNSIGNED -typedef unsigned page_flags_t; -#else -typedef unsigned long page_flags_t; -#endif - /* * Each physical page in the system has a struct page associated with * it to keep track of whatever it is we are using the page for at the @@ -219,7 +213,7 @@ typedef unsigned long page_flags_t; * a page. */ struct page { - page_flags_t flags; /* Atomic flags, some possibly + unsigned long flags; /* Atomic flags, some possibly * updated asynchronously */ atomic_t _count; /* Usage count, see below. */ atomic_t _mapcount; /* Count of ptes mapped in mms, @@ -435,7 +429,7 @@ static inline void put_page(struct page *page) #endif /* Page flags: | [SECTION] | [NODE] | ZONE | ... | FLAGS | */ -#define SECTIONS_PGOFF ((sizeof(page_flags_t)*8) - SECTIONS_WIDTH) +#define SECTIONS_PGOFF ((sizeof(unsigned long)*8) - SECTIONS_WIDTH) #define NODES_PGOFF (SECTIONS_PGOFF - NODES_WIDTH) #define ZONES_PGOFF (NODES_PGOFF - ZONES_WIDTH) diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index da7a829f8561..57fc99c67c31 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -455,7 +455,7 @@ extern struct pglist_data contig_page_data; #include #endif -#if BITS_PER_LONG == 32 || defined(ARCH_HAS_ATOMIC_UNSIGNED) +#if BITS_PER_LONG == 32 /* * with 32 bit page->flags field, we reserve 9 bits for node/zone info. * there are 4 zones (3 bits) and this leaves 9-3=6 bits for nodes. diff --git a/mm/filemap.c b/mm/filemap.c index 5d6e4c2000dc..33a28bfde158 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -134,7 +134,7 @@ static int sync_page(void *word) struct address_space *mapping; struct page *page; - page = container_of((page_flags_t *)word, struct page, flags); + page = container_of((unsigned long *)word, struct page, flags); /* * page_mapping() is being called without PG_locked held. diff --git a/mm/page_alloc.c b/mm/page_alloc.c index e8810a7e0264..259a71bacca4 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -128,7 +128,7 @@ static void bad_page(const char *function, struct page *page) printk(KERN_EMERG "Bad page state at %s (in process '%s', page %p)\n", function, current->comm, page); printk(KERN_EMERG "flags:0x%0*lx mapping:%p mapcount:%d count:%d\n", - (int)(2*sizeof(page_flags_t)), (unsigned long)page->flags, + (int)(2*sizeof(unsigned long)), (unsigned long)page->flags, page->mapping, page_mapcount(page), page_count(page)); printk(KERN_EMERG "Backtrace:\n"); dump_stack(); -- cgit v1.2.3 From 6b1de9161e973bac8c4675db608fe4f38d2689bd Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Thu, 17 Nov 2005 21:35:02 +0100 Subject: [PATCH] VM: fix zone list restart in page allocatate We must reassign z before looping through the zones kicking kswapd, since it will be NULL if we hit an OOM condition and jump back to the beginning again. 'z' is initially assigned before the restart: label. So move the restart label up a little. Signed-off-by: Jens Axboe --- mm/page_alloc.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 104e69ca55e0..bd4de592dc23 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -845,21 +845,22 @@ __alloc_pages(gfp_t gfp_mask, unsigned int order, might_sleep_if(wait); +restart: z = zonelist->zones; /* the list of zones suitable for gfp_mask */ if (unlikely(*z == NULL)) { /* Should this ever happen?? */ return NULL; } -restart: + page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, zonelist, ALLOC_CPUSET); if (page) goto got_pg; - do + do { wakeup_kswapd(*z, order); - while (*(++z)); + } while (*(++z)); /* * OK, we're below the kswapd watermark and have kicked background -- cgit v1.2.3