Hello everyone,
This is yet another round of Contiguous Memory Allocator patches. I hope that I've managed to resolve all the items discussed during the Memory Management summit at Linaro Meeting in Budapest and pointed later on mailing lists. The goal is to integrate it as tight as possible with other kernel subsystems (like memory management and dma-mapping) and finally merge to mainline.
Previous version introduced integration with DMA-mapping subsystem for ARM architecture. In this version I've cleaned up it even more and prepared for easier integration on other than ARM architectures. I've also rebased all the code onto latest v3.0-rc6 kernel.
A few words for these who see CMA for the first time:
The Contiguous Memory Allocator (CMA) makes it possible for device drivers to allocate big contiguous chunks of memory after the system has booted.
The main difference from the similar frameworks is the fact that CMA allows to transparently reuse memory region reserved for the big chunk allocation as a system memory, so no memory is wasted when no big chunk is allocated. Once the alloc request is issued, the framework will migrate system pages to create a required big chunk of physically contiguous memory.
For more information you can refer to nice LWN article: http://lwn.net/Articles/447405/ and links to previous versions of CMA framework.
The CMA framework has been initially developed by Michal Nazarewicz at Samsung Poland R&D Center. Since version 9, I've taken over the development, because Michal has left the company.
The current version of CMA is a set of helper functions for DMA mapping framework that handles allocation of contiguous memory blocks. The difference between this patchset and Kamezawa's alloc_contig_pages() are:
1. alloc_contig_pages() requires MAX_ORDER alignment of allocations which may be unsuitable for embeded systems where a few MiBs are required.
Lack of the requirement on the alignment means that several threads might try to access the same pageblock/page. To prevent this from happening CMA uses a mutex so that only one allocating/releasing function may run at one point.
2. CMA may use its own migratetype (MIGRATE_CMA) which behaves similarly to ZONE_MOVABLE but can be put in arbitrary places.
This is required for us since we need to define two disjoint memory ranges inside system RAM. (ie. in two memory banks (do not confuse with nodes)).
3. alloc_contig_pages() scans memory in search for range that could be migrated. CMA on the other hand maintains its own allocator to decide where to allocate memory for device drivers and then tries to migrate pages from that part if needed. This is not strictly required but I somehow feel it might be faster.
The integration with ARM DMA-mapping subsystem is quite straightforward. Once cma context is available alloc_pages() can be replaced by dma_alloc_from_contiguous() call.
Current version have been tested on Samsung S5PC110 based Goni machine and s5p-fimc V4L2 driver. The driver itself uses videobuf2 dma-contig memory allocator, which in turn relies on dma_alloc_coherent() from DMA-mapping subsystem. By integrating CMA with DMA-mapping we managed to get this driver working with CMA without any single change required in the driver or videobuf2-dma-contig allocator.
TODO: - resolve double-mapping issues with ARMv6+ and coherent memory
Best regards
From: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com
Memory hotplug is a logic for making pages unused in the specified range of pfn. So, some of core logics can be used for other purpose as allocating a very large contigous memory block.
This patch moves some functions from mm/memory_hotplug.c to mm/page_isolation.c. This helps adding a function for large-alloc in page_isolation.c with memory-unplug technique.
Signed-off-by: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com [m.nazarewicz: reworded commit message] Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com [m.szyprowski: rebased and updated to Linux v3.0-rc1] Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com --- include/linux/page-isolation.h | 7 +++ mm/memory_hotplug.c | 111 -------------------------------------- mm/page_isolation.c | 115 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 122 insertions(+), 111 deletions(-)
diff --git a/include/linux/page-isolation.h b/include/linux/page-isolation.h index 051c1b1..58cdbac 100644 --- a/include/linux/page-isolation.h +++ b/include/linux/page-isolation.h @@ -33,5 +33,12 @@ test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn); extern int set_migratetype_isolate(struct page *page); extern void unset_migratetype_isolate(struct page *page);
+/* + * For migration. + */ + +int test_pages_in_a_zone(unsigned long start_pfn, unsigned long end_pfn); +unsigned long scan_lru_pages(unsigned long start, unsigned long end); +int do_migrate_range(unsigned long start_pfn, unsigned long end_pfn);
#endif diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index c46887b..c32ca23 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -645,117 +645,6 @@ int is_mem_section_removable(unsigned long start_pfn, unsigned long nr_pages) }
/* - * Confirm all pages in a range [start, end) is belongs to the same zone. - */ -static int test_pages_in_a_zone(unsigned long start_pfn, unsigned long end_pfn) -{ - unsigned long pfn; - struct zone *zone = NULL; - struct page *page; - int i; - for (pfn = start_pfn; - pfn < end_pfn; - pfn += MAX_ORDER_NR_PAGES) { - i = 0; - /* This is just a CONFIG_HOLES_IN_ZONE check.*/ - while ((i < MAX_ORDER_NR_PAGES) && !pfn_valid_within(pfn + i)) - i++; - if (i == MAX_ORDER_NR_PAGES) - continue; - page = pfn_to_page(pfn + i); - if (zone && page_zone(page) != zone) - return 0; - zone = page_zone(page); - } - return 1; -} - -/* - * Scanning pfn is much easier than scanning lru list. - * Scan pfn from start to end and Find LRU page. - */ -static unsigned long scan_lru_pages(unsigned long start, unsigned long end) -{ - unsigned long pfn; - struct page *page; - for (pfn = start; pfn < end; pfn++) { - if (pfn_valid(pfn)) { - page = pfn_to_page(pfn); - if (PageLRU(page)) - return pfn; - } - } - return 0; -} - -static struct page * -hotremove_migrate_alloc(struct page *page, unsigned long private, int **x) -{ - /* This should be improooooved!! */ - return alloc_page(GFP_HIGHUSER_MOVABLE); -} - -#define NR_OFFLINE_AT_ONCE_PAGES (256) -static int -do_migrate_range(unsigned long start_pfn, unsigned long end_pfn) -{ - unsigned long pfn; - struct page *page; - int move_pages = NR_OFFLINE_AT_ONCE_PAGES; - int not_managed = 0; - int ret = 0; - LIST_HEAD(source); - - for (pfn = start_pfn; pfn < end_pfn && move_pages > 0; pfn++) { - if (!pfn_valid(pfn)) - continue; - page = pfn_to_page(pfn); - if (!get_page_unless_zero(page)) - continue; - /* - * We can skip free pages. And we can only deal with pages on - * LRU. - */ - ret = isolate_lru_page(page); - if (!ret) { /* Success */ - put_page(page); - list_add_tail(&page->lru, &source); - move_pages--; - inc_zone_page_state(page, NR_ISOLATED_ANON + - page_is_file_cache(page)); - - } else { -#ifdef CONFIG_DEBUG_VM - printk(KERN_ALERT "removing pfn %lx from LRU failed\n", - pfn); - dump_page(page); -#endif - put_page(page); - /* Because we don't have big zone->lock. we should - check this again here. */ - if (page_count(page)) { - not_managed++; - ret = -EBUSY; - break; - } - } - } - if (!list_empty(&source)) { - if (not_managed) { - putback_lru_pages(&source); - goto out; - } - /* this function returns # of failed pages */ - ret = migrate_pages(&source, hotremove_migrate_alloc, 0, - true, true); - if (ret) - putback_lru_pages(&source); - } -out: - return ret; -} - -/* * remove from free_area[] and mark all as Reserved. */ static int diff --git a/mm/page_isolation.c b/mm/page_isolation.c index 4ae42bb..15b41ec 100644 --- a/mm/page_isolation.c +++ b/mm/page_isolation.c @@ -5,6 +5,9 @@ #include <linux/mm.h> #include <linux/page-isolation.h> #include <linux/pageblock-flags.h> +#include <linux/memcontrol.h> +#include <linux/migrate.h> +#include <linux/mm_inline.h> #include "internal.h"
static inline struct page * @@ -139,3 +142,115 @@ int test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn) spin_unlock_irqrestore(&zone->lock, flags); return ret ? 0 : -EBUSY; } + + +/* + * Confirm all pages in a range [start, end) is belongs to the same zone. + */ +int test_pages_in_a_zone(unsigned long start_pfn, unsigned long end_pfn) +{ + unsigned long pfn; + struct zone *zone = NULL; + struct page *page; + int i; + for (pfn = start_pfn; + pfn < end_pfn; + pfn += MAX_ORDER_NR_PAGES) { + i = 0; + /* This is just a CONFIG_HOLES_IN_ZONE check.*/ + while ((i < MAX_ORDER_NR_PAGES) && !pfn_valid_within(pfn + i)) + i++; + if (i == MAX_ORDER_NR_PAGES) + continue; + page = pfn_to_page(pfn + i); + if (zone && page_zone(page) != zone) + return 0; + zone = page_zone(page); + } + return 1; +} + +/* + * Scanning pfn is much easier than scanning lru list. + * Scan pfn from start to end and Find LRU page. + */ +unsigned long scan_lru_pages(unsigned long start, unsigned long end) +{ + unsigned long pfn; + struct page *page; + for (pfn = start; pfn < end; pfn++) { + if (pfn_valid(pfn)) { + page = pfn_to_page(pfn); + if (PageLRU(page)) + return pfn; + } + } + return 0; +} + +struct page * +hotremove_migrate_alloc(struct page *page, unsigned long private, int **x) +{ + /* This should be improooooved!! */ + return alloc_page(GFP_HIGHUSER_MOVABLE); +} + +#define NR_OFFLINE_AT_ONCE_PAGES (256) +int do_migrate_range(unsigned long start_pfn, unsigned long end_pfn) +{ + unsigned long pfn; + struct page *page; + int move_pages = NR_OFFLINE_AT_ONCE_PAGES; + int not_managed = 0; + int ret = 0; + LIST_HEAD(source); + + for (pfn = start_pfn; pfn < end_pfn && move_pages > 0; pfn++) { + if (!pfn_valid(pfn)) + continue; + page = pfn_to_page(pfn); + if (!get_page_unless_zero(page)) + continue; + /* + * We can skip free pages. And we can only deal with pages on + * LRU. + */ + ret = isolate_lru_page(page); + if (!ret) { /* Success */ + put_page(page); + list_add_tail(&page->lru, &source); + move_pages--; + inc_zone_page_state(page, NR_ISOLATED_ANON + + page_is_file_cache(page)); + + } else { +#ifdef CONFIG_DEBUG_VM + printk(KERN_ALERT "removing pfn %lx from LRU failed\n", + pfn); + dump_page(page); +#endif + put_page(page); + /* Because we don't have big zone->lock. we should + check this again here. */ + if (page_count(page)) { + not_managed++; + ret = -EBUSY; + break; + } + } + } + if (!list_empty(&source)) { + if (not_managed) { + putback_lru_pages(&source); + goto out; + } + /* this function returns # of failed pages */ + ret = migrate_pages(&source, hotremove_migrate_alloc, 0, + true, true); + if (ret) + putback_lru_pages(&source); + } +out: + return ret; +} +
On Tuesday 05 July 2011, Marek Szyprowski wrote:
From: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com
Memory hotplug is a logic for making pages unused in the specified range of pfn. So, some of core logics can be used for other purpose as allocating a very large contigous memory block.
This patch moves some functions from mm/memory_hotplug.c to mm/page_isolation.c. This helps adding a function for large-alloc in page_isolation.c with memory-unplug technique.
Signed-off-by: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com [m.nazarewicz: reworded commit message] Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com [m.szyprowski: rebased and updated to Linux v3.0-rc1] Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com
Acked-by: Arnd Bergmann arnd@arndb.de
From: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com
This commit introduces alloc_contig_freed_pages() function which allocates (ie. removes from buddy system) free pages in range. Caller has to guarantee that all pages in range are in buddy system.
Along with this function, a free_contig_pages() function is provided which frees all (or a subset of) pages allocated with alloc_contig_free_pages().
Michal Nazarewicz has modified the function to make it easier to allocate not MAX_ORDER_NR_PAGES aligned pages by making it return pfn of one-past-the-last allocated page.
Signed-off-by: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com --- include/linux/page-isolation.h | 3 ++ mm/page_alloc.c | 44 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 47 insertions(+), 0 deletions(-)
diff --git a/include/linux/page-isolation.h b/include/linux/page-isolation.h index 58cdbac..f1417ed 100644 --- a/include/linux/page-isolation.h +++ b/include/linux/page-isolation.h @@ -32,6 +32,9 @@ test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn); */ extern int set_migratetype_isolate(struct page *page); extern void unset_migratetype_isolate(struct page *page); +extern unsigned long alloc_contig_freed_pages(unsigned long start, + unsigned long end, gfp_t flag); +extern void free_contig_pages(struct page *page, int nr_pages);
/* * For migration. diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 4e8985a..00e9b24 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -5600,6 +5600,50 @@ out: spin_unlock_irqrestore(&zone->lock, flags); }
+unsigned long alloc_contig_freed_pages(unsigned long start, unsigned long end, + gfp_t flag) +{ + unsigned long pfn = start, count; + struct page *page; + struct zone *zone; + int order; + + VM_BUG_ON(!pfn_valid(start)); + zone = page_zone(pfn_to_page(start)); + + spin_lock_irq(&zone->lock); + + page = pfn_to_page(pfn); + for (;;) { + VM_BUG_ON(page_count(page) || !PageBuddy(page)); + list_del(&page->lru); + order = page_order(page); + zone->free_area[order].nr_free--; + rmv_page_order(page); + __mod_zone_page_state(zone, NR_FREE_PAGES, -(1UL << order)); + pfn += 1 << order; + if (pfn >= end) + break; + VM_BUG_ON(!pfn_valid(pfn)); + page += 1 << order; + } + + spin_unlock_irq(&zone->lock); + + /* After this, pages in the range can be freed one be one */ + page = pfn_to_page(start); + for (count = pfn - start; count; --count, ++page) + prep_new_page(page, 0, flag); + + return pfn; +} + +void free_contig_pages(struct page *page, int nr_pages) +{ + for (; nr_pages; --nr_pages, ++page) + __free_page(page); +} + #ifdef CONFIG_MEMORY_HOTREMOVE /* * All pages in the range must be isolated before calling this.
On Tuesday 05 July 2011, Marek Szyprowski wrote:
From: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com
This commit introduces alloc_contig_freed_pages() function which allocates (ie. removes from buddy system) free pages in range. Caller has to guarantee that all pages in range are in buddy system.
Along with this function, a free_contig_pages() function is provided which frees all (or a subset of) pages allocated with alloc_contig_free_pages().
Michal Nazarewicz has modified the function to make it easier to allocate not MAX_ORDER_NR_PAGES aligned pages by making it return pfn of one-past-the-last allocated page.
Signed-off-by: KAMEZAWA Hiroyuki kamezawa.hiroyu@jp.fujitsu.com Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com
Acked-by: Arnd Bergmann arnd@arndb.de
From: Michal Nazarewicz m.nazarewicz@samsung.com
This commit adds the alloc_contig_range() function which tries to allecate given range of pages. It tries to migrate all already allocated pages that fall in the range thus freeing them. Once all pages in the range are freed they are removed from the buddy system thus allocated for the caller to use.
Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com [m.szyprowski: renamed some variables for easier code reading] Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com --- include/linux/page-isolation.h | 2 + mm/page_alloc.c | 144 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 146 insertions(+), 0 deletions(-)
diff --git a/include/linux/page-isolation.h b/include/linux/page-isolation.h index f1417ed..c5d1a7c 100644 --- a/include/linux/page-isolation.h +++ b/include/linux/page-isolation.h @@ -34,6 +34,8 @@ extern int set_migratetype_isolate(struct page *page); extern void unset_migratetype_isolate(struct page *page); extern unsigned long alloc_contig_freed_pages(unsigned long start, unsigned long end, gfp_t flag); +extern int alloc_contig_range(unsigned long start, unsigned long end, + gfp_t flags); extern void free_contig_pages(struct page *page, int nr_pages);
/* diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 00e9b24..2cea044 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -5638,6 +5638,150 @@ unsigned long alloc_contig_freed_pages(unsigned long start, unsigned long end, return pfn; }
+static unsigned long pfn_to_maxpage(unsigned long pfn) +{ + return pfn & ~(MAX_ORDER_NR_PAGES - 1); +} + +static unsigned long pfn_to_maxpage_up(unsigned long pfn) +{ + return ALIGN(pfn, MAX_ORDER_NR_PAGES); +} + +#define MIGRATION_RETRY 5 +static int __alloc_contig_migrate_range(unsigned long start, unsigned long end) +{ + int migration_failed = 0, ret; + unsigned long pfn = start; + + /* + * Some code "borrowed" from KAMEZAWA Hiroyuki's + * __alloc_contig_pages(). + */ + + for (;;) { + pfn = scan_lru_pages(pfn, end); + if (!pfn || pfn >= end) + break; + + ret = do_migrate_range(pfn, end); + if (!ret) { + migration_failed = 0; + } else if (ret != -EBUSY + || ++migration_failed >= MIGRATION_RETRY) { + return ret; + } else { + /* There are unstable pages.on pagevec. */ + lru_add_drain_all(); + /* + * there may be pages on pcplist before + * we mark the range as ISOLATED. + */ + drain_all_pages(); + } + cond_resched(); + } + + if (!migration_failed) { + /* drop all pages in pagevec and pcp list */ + lru_add_drain_all(); + drain_all_pages(); + } + + /* Make sure all pages are isolated */ + if (WARN_ON(test_pages_isolated(start, end))) + return -EBUSY; + + return 0; +} + +/** + * alloc_contig_range() -- tries to allocate given range of pages + * @start: start PFN to allocate + * @end: one-past-the-last PFN to allocate + * @flags: flags passed to alloc_contig_freed_pages(). + * + * The PFN range does not have to be pageblock or MAX_ORDER_NR_PAGES + * aligned, hovewer it's callers responsibility to guarantee that we + * are the only thread that changes migrate type of pageblocks the + * pages fall in. + * + * Returns zero on success or negative error code. On success all + * pages which PFN is in (start, end) are allocated for the caller and + * need to be freed with free_contig_pages(). + */ +int alloc_contig_range(unsigned long start, unsigned long end, + gfp_t flags) +{ + unsigned long outer_start, outer_end; + int ret; + + /* + * What we do here is we mark all pageblocks in range as + * MIGRATE_ISOLATE. Because of the way page allocator work, we + * align the range to MAX_ORDER pages so that page allocator + * won't try to merge buddies from different pageblocks and + * change MIGRATE_ISOLATE to some other migration type. + * + * Once the pageblocks are marked as MIGRATE_ISOLATE, we + * migrate the pages from an unaligned range (ie. pages that + * we are interested in). This will put all the pages in + * range back to page allocator as MIGRATE_ISOLATE. + * + * When this is done, we take the pages in range from page + * allocator removing them from the buddy system. This way + * page allocator will never consider using them. + * + * This lets us mark the pageblocks back as + * MIGRATE_CMA/MIGRATE_MOVABLE so that free pages in the + * MAX_ORDER aligned range but not in the unaligned, original + * range are put back to page allocator so that buddy can use + * them. + */ + + ret = start_isolate_page_range(pfn_to_maxpage(start), + pfn_to_maxpage_up(end)); + if (ret) + goto done; + + ret = __alloc_contig_migrate_range(start, end); + if (ret) + goto done; + + /* + * Pages from [start, end) are within a MAX_ORDER_NR_PAGES + * aligned blocks that are marked as MIGRATE_ISOLATE. What's + * more, all pages in [start, end) are free in page allocator. + * What we are going to do is to allocate all pages from + * [start, end) (that is remove them from page allocater). + * + * The only problem is that pages at the beginning and at the + * end of interesting range may be not aligned with pages that + * page allocator holds, ie. they can be part of higher order + * pages. Because of this, we reserve the bigger range and + * once this is done free the pages we are not interested in. + */ + + ret = 0; + while (!PageBuddy(pfn_to_page(start & (~0UL << ret)))) + if (WARN_ON(++ret >= MAX_ORDER)) + return -EINVAL; + + outer_start = start & (~0UL << ret); + outer_end = alloc_contig_freed_pages(outer_start, end, flags); + + /* Free head and tail (if any) */ + if (start != outer_start) + free_contig_pages(pfn_to_page(outer_start), start - outer_start); + if (end != outer_end) + free_contig_pages(pfn_to_page(end), outer_end - end); + + ret = 0; +done: + undo_isolate_page_range(pfn_to_maxpage(start), pfn_to_maxpage_up(end)); + return ret; +} + void free_contig_pages(struct page *page, int nr_pages) { for (; nr_pages; --nr_pages, ++page)
On Tuesday 05 July 2011, Marek Szyprowski wrote:
From: Michal Nazarewicz m.nazarewicz@samsung.com
This commit adds the alloc_contig_range() function which tries to allecate given range of pages. It tries to migrate all already allocated pages that fall in the range thus freeing them. Once all pages in the range are freed they are removed from the buddy system thus allocated for the caller to use.
Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com [m.szyprowski: renamed some variables for easier code reading] Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com
Acked-by: Arnd Bergmann arnd@arndb.de
From: Michal Nazarewicz m.nazarewicz@samsung.com
The MIGRATE_CMA migration type has two main characteristics: (i) only movable pages can be allocated from MIGRATE_CMA pageblocks and (ii) page allocator will never change migration type of MIGRATE_CMA pageblocks.
This guarantees that page in a MIGRATE_CMA page block can always be migrated somewhere else (unless there's no memory left in the system).
It is designed to be used with Contiguous Memory Allocator (CMA) for allocating big chunks (eg. 10MiB) of physically contiguous memory. Once driver requests contiguous memory, CMA will migrate pages from MIGRATE_CMA pageblocks.
To minimise number of migrations, MIGRATE_CMA migration type is the last type tried when page allocator falls back to other migration types then requested.
Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com [m.szyprowski: cleaned up Kconfig, renamed some functions] Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com
cma migrate fixup --- include/linux/mmzone.h | 43 +++++++++++++++--- include/linux/page-isolation.h | 4 ++ mm/Kconfig | 8 +++- mm/compaction.c | 10 ++++ mm/page_alloc.c | 94 ++++++++++++++++++++++++++++++++-------- 5 files changed, 132 insertions(+), 27 deletions(-)
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 9f7c3eb..126014d 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -35,13 +35,37 @@ */ #define PAGE_ALLOC_COSTLY_ORDER 3
-#define MIGRATE_UNMOVABLE 0 -#define MIGRATE_RECLAIMABLE 1 -#define MIGRATE_MOVABLE 2 -#define MIGRATE_PCPTYPES 3 /* the number of types on the pcp lists */ -#define MIGRATE_RESERVE 3 -#define MIGRATE_ISOLATE 4 /* can't allocate from here */ -#define MIGRATE_TYPES 5 +enum { + MIGRATE_UNMOVABLE, + MIGRATE_RECLAIMABLE, + MIGRATE_MOVABLE, + MIGRATE_PCPTYPES, /* the number of types on the pcp lists */ + MIGRATE_RESERVE = MIGRATE_PCPTYPES, +#ifdef CONFIG_CMA_MIGRATE_TYPE + /* + * MIGRATE_CMA migration type is designed to mimic the way + * ZONE_MOVABLE works. Only movable pages can be allocated + * from MIGRATE_CMA pageblocks and page allocator never + * implicitly change migration type of MIGRATE_CMA pageblock. + * + * The way to use it is to change migratetype of a range of + * pageblocks to MIGRATE_CMA which can be done by + * __free_pageblock_cma() function. What is important though + * is that a range of pageblocks must be aligned to + * MAX_ORDER_NR_PAGES should biggest page be bigger then + * a single pageblock. + */ + MIGRATE_CMA, +#endif + MIGRATE_ISOLATE, /* can't allocate from here */ + MIGRATE_TYPES +}; + +#ifdef CONFIG_CMA_MIGRATE_TYPE +# define is_migrate_cma(migratetype) unlikely((migratetype) == MIGRATE_CMA) +#else +# define is_migrate_cma(migratetype) false +#endif
#define for_each_migratetype_order(order, type) \ for (order = 0; order < MAX_ORDER; order++) \ @@ -54,6 +78,11 @@ static inline int get_pageblock_migratetype(struct page *page) return get_pageblock_flags_group(page, PB_migrate, PB_migrate_end); }
+static inline bool is_pageblock_cma(struct page *page) +{ + return is_migrate_cma(get_pageblock_migratetype(page)); +} + struct free_area { struct list_head free_list[MIGRATE_TYPES]; unsigned long nr_free; diff --git a/include/linux/page-isolation.h b/include/linux/page-isolation.h index c5d1a7c..014ebb5 100644 --- a/include/linux/page-isolation.h +++ b/include/linux/page-isolation.h @@ -46,4 +46,8 @@ int test_pages_in_a_zone(unsigned long start_pfn, unsigned long end_pfn); unsigned long scan_lru_pages(unsigned long start, unsigned long end); int do_migrate_range(unsigned long start_pfn, unsigned long end_pfn);
+#ifdef CONFIG_CMA_MIGRATE_TYPE +extern void init_cma_reserved_pageblock(struct page *page); +#endif + #endif diff --git a/mm/Kconfig b/mm/Kconfig index 8ca47a5..6ffedd8 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -189,7 +189,7 @@ config COMPACTION config MIGRATION bool "Page migration" def_bool y - depends on NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION + depends on NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA_MIGRATE_TYPE help Allows the migration of the physical location of pages of processes while the virtual addresses are not changed. This is useful in @@ -198,6 +198,12 @@ config MIGRATION pages as migration can relocate pages to satisfy a huge page allocation instead of reclaiming.
+config CMA_MIGRATE_TYPE + bool + help + This enables the use the MIGRATE_CMA migrate type, which lets lets CMA + work on almost arbitrary memory range and not only inside ZONE_MOVABLE. + config PHYS_ADDR_T_64BIT def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
diff --git a/mm/compaction.c b/mm/compaction.c index 6cc604b..9e5cc59 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -119,6 +119,16 @@ static bool suitable_migration_target(struct page *page) if (migratetype == MIGRATE_ISOLATE || migratetype == MIGRATE_RESERVE) return false;
+ /* Keep MIGRATE_CMA alone as well. */ + /* + * XXX Revisit. We currently cannot let compaction touch CMA + * pages since compaction insists on changing their migration + * type to MIGRATE_MOVABLE (see split_free_page() called from + * isolate_freepages_block() above). + */ + if (is_migrate_cma(migratetype)) + return false; + /* If the page is a large free page, then allow migration */ if (PageBuddy(page) && page_order(page) >= pageblock_order) return true; diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 2cea044..1353a0c 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -719,6 +719,31 @@ void __meminit __free_pages_bootmem(struct page *page, unsigned int order) } }
+#ifdef CONFIG_CMA_MIGRATE_TYPE + +/* + * Free whole pageblock and set it's migration type to MIGRATE_CMA. + */ +void __init init_cma_reserved_pageblock(struct page *page) +{ + struct page *p = page; + unsigned i = pageblock_nr_pages; + + prefetchw(p); + do { + if (--i) + prefetchw(p + 1); + __ClearPageReserved(p); + set_page_count(p, 0); + } while (++p, i); + + set_page_refcounted(page); + set_pageblock_migratetype(page, MIGRATE_CMA); + __free_pages(page, pageblock_order); + ++totalram_pages; +} + +#endif
/* * The order of subdivision here is critical for the IO subsystem. @@ -827,11 +852,15 @@ struct page *__rmqueue_smallest(struct zone *zone, unsigned int order, * This array describes the order lists are fallen back to when * the free lists for the desirable migrate type are depleted */ -static int fallbacks[MIGRATE_TYPES][MIGRATE_TYPES-1] = { +static int fallbacks[MIGRATE_TYPES][4] = { [MIGRATE_UNMOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_MOVABLE, MIGRATE_RESERVE }, [MIGRATE_RECLAIMABLE] = { MIGRATE_UNMOVABLE, MIGRATE_MOVABLE, MIGRATE_RESERVE }, +#ifdef CONFIG_CMA_MIGRATE_TYPE + [MIGRATE_MOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_CMA , MIGRATE_RESERVE }, +#else [MIGRATE_MOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_RESERVE }, - [MIGRATE_RESERVE] = { MIGRATE_RESERVE, MIGRATE_RESERVE, MIGRATE_RESERVE }, /* Never used */ +#endif + [MIGRATE_RESERVE] = { MIGRATE_RESERVE }, /* Never used */ };
/* @@ -926,12 +955,12 @@ __rmqueue_fallback(struct zone *zone, int order, int start_migratetype) /* Find the largest possible block of pages in the other list */ for (current_order = MAX_ORDER-1; current_order >= order; --current_order) { - for (i = 0; i < MIGRATE_TYPES - 1; i++) { + for (i = 0; i < ARRAY_SIZE(fallbacks[0]); i++) { migratetype = fallbacks[start_migratetype][i];
/* MIGRATE_RESERVE handled later if necessary */ if (migratetype == MIGRATE_RESERVE) - continue; + break;
area = &(zone->free_area[current_order]); if (list_empty(&area->free_list[migratetype])) @@ -946,19 +975,29 @@ __rmqueue_fallback(struct zone *zone, int order, int start_migratetype) * pages to the preferred allocation list. If falling * back for a reclaimable kernel allocation, be more * aggressive about taking ownership of free pages + * + * On the other hand, never change migration + * type of MIGRATE_CMA pageblocks nor move CMA + * pages on different free lists. We don't + * want unmovable pages to be allocated from + * MIGRATE_CMA areas. */ - if (unlikely(current_order >= (pageblock_order >> 1)) || - start_migratetype == MIGRATE_RECLAIMABLE || - page_group_by_mobility_disabled) { - unsigned long pages; + if (!is_pageblock_cma(page) && + (unlikely(current_order >= pageblock_order / 2) || + start_migratetype == MIGRATE_RECLAIMABLE || + page_group_by_mobility_disabled)) { + int pages; pages = move_freepages_block(zone, page, - start_migratetype); + start_migratetype);
- /* Claim the whole block if over half of it is free */ + /* + * Claim the whole block if over half + * of it is free + */ if (pages >= (1 << (pageblock_order-1)) || - page_group_by_mobility_disabled) + page_group_by_mobility_disabled) set_pageblock_migratetype(page, - start_migratetype); + start_migratetype);
migratetype = start_migratetype; } @@ -968,11 +1007,14 @@ __rmqueue_fallback(struct zone *zone, int order, int start_migratetype) rmv_page_order(page);
/* Take ownership for orders >= pageblock_order */ - if (current_order >= pageblock_order) + if (current_order >= pageblock_order && + !is_pageblock_cma(page)) change_pageblock_range(page, current_order, start_migratetype);
- expand(zone, page, order, current_order, area, migratetype); + expand(zone, page, order, current_order, area, + is_migrate_cma(start_migratetype) + ? start_migratetype : migratetype);
trace_mm_page_alloc_extfrag(page, order, current_order, start_migratetype, migratetype); @@ -1044,7 +1086,12 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order, list_add(&page->lru, list); else list_add_tail(&page->lru, list); - set_page_private(page, migratetype); +#ifdef CONFIG_CMA_MIGRATE_TYPE + if (is_pageblock_cma(page)) + set_page_private(page, MIGRATE_CMA); + else +#endif + set_page_private(page, migratetype); list = &page->lru; } __mod_zone_page_state(zone, NR_FREE_PAGES, -(i << order)); @@ -1185,9 +1232,16 @@ void free_hot_cold_page(struct page *page, int cold) * offlined but treat RESERVE as movable pages so we can get those * areas back if necessary. Otherwise, we may have to free * excessively into the page allocator + * + * Still, do not change migration type of MIGRATE_CMA pages (if + * they'd be recorded as MIGRATE_MOVABLE an unmovable page could + * be allocated from MIGRATE_CMA block and we don't want to allow + * that). In this respect, treat MIGRATE_CMA like + * MIGRATE_ISOLATE. */ if (migratetype >= MIGRATE_PCPTYPES) { - if (unlikely(migratetype == MIGRATE_ISOLATE)) { + if (unlikely(migratetype == MIGRATE_ISOLATE + || is_migrate_cma(migratetype))) { free_one_page(zone, page, 0, migratetype); goto out; } @@ -1276,7 +1330,9 @@ int split_free_page(struct page *page) if (order >= pageblock_order - 1) { struct page *endpage = page + (1 << order) - 1; for (; page < endpage; page += pageblock_nr_pages) - set_pageblock_migratetype(page, MIGRATE_MOVABLE); + if (!is_pageblock_cma(page)) + set_pageblock_migratetype(page, + MIGRATE_MOVABLE); }
return 1 << order; @@ -5486,8 +5542,8 @@ __count_immobile_pages(struct zone *zone, struct page *page, int count) */ if (zone_idx(zone) == ZONE_MOVABLE) return true; - - if (get_pageblock_migratetype(page) == MIGRATE_MOVABLE) + if (get_pageblock_migratetype(page) == MIGRATE_MOVABLE || + is_pageblock_cma(page)) return true;
pfn = page_to_pfn(page);
On Tuesday 05 July 2011, Marek Szyprowski wrote:
From: Michal Nazarewicz m.nazarewicz@samsung.com
The MIGRATE_CMA migration type has two main characteristics: (i) only movable pages can be allocated from MIGRATE_CMA pageblocks and (ii) page allocator will never change migration type of MIGRATE_CMA pageblocks.
This guarantees that page in a MIGRATE_CMA page block can always be migrated somewhere else (unless there's no memory left in the system).
It is designed to be used with Contiguous Memory Allocator (CMA) for allocating big chunks (eg. 10MiB) of physically contiguous memory. Once driver requests contiguous memory, CMA will migrate pages from MIGRATE_CMA pageblocks.
To minimise number of migrations, MIGRATE_CMA migration type is the last type tried when page allocator falls back to other migration types then requested.
Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com [m.szyprowski: cleaned up Kconfig, renamed some functions] Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com
Acked-by: Arnd Bergmann arnd@arndb.de, but I noticed a few things:
cma migrate fixup
This text doesn't belong here.
+enum {
- MIGRATE_UNMOVABLE,
- MIGRATE_RECLAIMABLE,
- MIGRATE_MOVABLE,
- MIGRATE_PCPTYPES, /* the number of types on the pcp lists */
- MIGRATE_RESERVE = MIGRATE_PCPTYPES,
+#ifdef CONFIG_CMA_MIGRATE_TYPE
- /*
* MIGRATE_CMA migration type is designed to mimic the way* ZONE_MOVABLE works. Only movable pages can be allocated* from MIGRATE_CMA pageblocks and page allocator never* implicitly change migration type of MIGRATE_CMA pageblock.** The way to use it is to change migratetype of a range of* pageblocks to MIGRATE_CMA which can be done by* __free_pageblock_cma() function. What is important though* is that a range of pageblocks must be aligned to* MAX_ORDER_NR_PAGES should biggest page be bigger then* a single pageblock.*/- MIGRATE_CMA,
+#endif
- MIGRATE_ISOLATE, /* can't allocate from here */
- MIGRATE_TYPES
+};
It's not clear to me why you need this #ifdef. Does it hurt if the migration type is defined but not used?
@@ -198,6 +198,12 @@ config MIGRATION pages as migration can relocate pages to satisfy a huge page allocation instead of reclaiming. +config CMA_MIGRATE_TYPE
- bool
- help
This enables the use the MIGRATE_CMA migrate type, which lets lets CMAwork on almost arbitrary memory range and not only inside ZONE_MOVABLE.config PHYS_ADDR_T_64BIT def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
This is currently only selected on ARM with your patch set.
diff --git a/mm/compaction.c b/mm/compaction.c index 6cc604b..9e5cc59 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -119,6 +119,16 @@ static bool suitable_migration_target(struct page *page) if (migratetype == MIGRATE_ISOLATE || migratetype == MIGRATE_RESERVE) return false;
- /* Keep MIGRATE_CMA alone as well. */
- /*
* XXX Revisit. We currently cannot let compaction touch CMA* pages since compaction insists on changing their migration* type to MIGRATE_MOVABLE (see split_free_page() called from* isolate_freepages_block() above).*/- if (is_migrate_cma(migratetype))
return false;- /* If the page is a large free page, then allow migration */ if (PageBuddy(page) && page_order(page) >= pageblock_order) return true;
Do you plan to fix address this before merging the patch set, or is it harmless enough to get in this way?
/*
- The order of subdivision here is critical for the IO subsystem.
@@ -827,11 +852,15 @@ struct page *__rmqueue_smallest(struct zone *zone, unsigned int order,
- This array describes the order lists are fallen back to when
- the free lists for the desirable migrate type are depleted
*/ -static int fallbacks[MIGRATE_TYPES][MIGRATE_TYPES-1] = { +static int fallbacks[MIGRATE_TYPES][4] = { [MIGRATE_UNMOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_MOVABLE, MIGRATE_RESERVE }, [MIGRATE_RECLAIMABLE] = { MIGRATE_UNMOVABLE, MIGRATE_MOVABLE, MIGRATE_RESERVE }, +#ifdef CONFIG_CMA_MIGRATE_TYPE
- [MIGRATE_MOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_CMA , MIGRATE_RESERVE },
+#else [MIGRATE_MOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_RESERVE },
- [MIGRATE_RESERVE] = { MIGRATE_RESERVE, MIGRATE_RESERVE, MIGRATE_RESERVE }, /* Never used */
+#endif
- [MIGRATE_RESERVE] = { MIGRATE_RESERVE }, /* Never used */
}; /* @@ -1044,7 +1086,12 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order, list_add(&page->lru, list); else list_add_tail(&page->lru, list);
set_page_private(page, migratetype);+#ifdef CONFIG_CMA_MIGRATE_TYPE
if (is_pageblock_cma(page))set_page_private(page, MIGRATE_CMA);else+#endif
set_page_private(page, migratetype);
I guess if you can get rid of the first #ifdef I mentioned above, these two can be removed as well, without causing any run-time overhead.
Arnd
On Tue, Jul 05, 2011 at 01:44:31PM +0200, Arnd Bergmann wrote:
@@ -198,6 +198,12 @@ config MIGRATION pages as migration can relocate pages to satisfy a huge page allocation instead of reclaiming. +config CMA_MIGRATE_TYPE
- bool
- help
This enables the use the MIGRATE_CMA migrate type, which lets lets CMAwork on almost arbitrary memory range and not only inside ZONE_MOVABLE.config PHYS_ADDR_T_64BIT def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
This is currently only selected on ARM with your patch set.
That's because CMA is targeted at solving the "we need massive contiguous DMA areas" problem on ARM SoCs.
And it does this without addressing the technical architecture problems surrounding multiple aliasing mappings with differing attributes which actually make it unsuitable for use on ARM. This is not the first time I've pointed that out, and I'm now at the point of basically ignoring this CMA work because I'm tired of constantly pointing this out.
My silence on this subject must not be taken as placid acceptance of the approach, but revulsion at seemingly being constantly ignored and having these patches pushed time and time again with nothing really changing on that issue.
It will be a sad day if these patches make their way into mainline without that being addressed, and will show contempt for architecture maintainers if it does.
From: Michal Nazarewicz m.nazarewicz@samsung.com
This commit changes various functions that change pages and pageblocks migrate type between MIGRATE_ISOLATE and MIGRATE_MOVABLE in such a way as to allow to work with MIGRATE_CMA migrate type.
Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com --- include/linux/page-isolation.h | 40 +++++++++++++++++++++++++++------------- mm/page_alloc.c | 19 ++++++++++++------- mm/page_isolation.c | 15 ++++++++------- 3 files changed, 47 insertions(+), 27 deletions(-)
diff --git a/include/linux/page-isolation.h b/include/linux/page-isolation.h index 014ebb5..96e287d 100644 --- a/include/linux/page-isolation.h +++ b/include/linux/page-isolation.h @@ -3,39 +3,53 @@
/* * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE. - * If specified range includes migrate types other than MOVABLE, + * If specified range includes migrate types other than MOVABLE or CMA, * this will fail with -EBUSY. * * For isolating all pages in the range finally, the caller have to * free all pages in the range. test_page_isolated() can be used for * test it. */ -extern int -start_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn); +int __start_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn, + unsigned migratetype); + +static inline int +start_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn) +{ + return __start_isolate_page_range(start_pfn, end_pfn, MIGRATE_MOVABLE); +} + +int __undo_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn, + unsigned migratetype);
/* * Changes MIGRATE_ISOLATE to MIGRATE_MOVABLE. * target range is [start_pfn, end_pfn) */ -extern int -undo_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn); +static inline int +undo_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn) +{ + return __undo_isolate_page_range(start_pfn, end_pfn, MIGRATE_MOVABLE); +}
/* - * test all pages in [start_pfn, end_pfn)are isolated or not. + * Test all pages in [start_pfn, end_pfn) are isolated or not. */ -extern int -test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn); +int test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn);
/* - * Internal funcs.Changes pageblock's migrate type. - * Please use make_pagetype_isolated()/make_pagetype_movable(). + * Internal functions. Changes pageblock's migrate type. */ -extern int set_migratetype_isolate(struct page *page); -extern void unset_migratetype_isolate(struct page *page); +int set_migratetype_isolate(struct page *page); +void __unset_migratetype_isolate(struct page *page, unsigned migratetype); +static inline void unset_migratetype_isolate(struct page *page) +{ + __unset_migratetype_isolate(page, MIGRATE_MOVABLE); +} extern unsigned long alloc_contig_freed_pages(unsigned long start, unsigned long end, gfp_t flag); extern int alloc_contig_range(unsigned long start, unsigned long end, - gfp_t flags); + gfp_t flags, unsigned migratetype); extern void free_contig_pages(struct page *page, int nr_pages);
/* diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 1353a0c..a936a75 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -5642,7 +5642,7 @@ out: return ret; }
-void unset_migratetype_isolate(struct page *page) +void __unset_migratetype_isolate(struct page *page, unsigned migratetype) { struct zone *zone; unsigned long flags; @@ -5650,8 +5650,8 @@ void unset_migratetype_isolate(struct page *page) spin_lock_irqsave(&zone->lock, flags); if (get_pageblock_migratetype(page) != MIGRATE_ISOLATE) goto out; - set_pageblock_migratetype(page, MIGRATE_MOVABLE); - move_freepages_block(zone, page, MIGRATE_MOVABLE); + set_pageblock_migratetype(page, migratetype); + move_freepages_block(zone, page, migratetype); out: spin_unlock_irqrestore(&zone->lock, flags); } @@ -5756,6 +5756,10 @@ static int __alloc_contig_migrate_range(unsigned long start, unsigned long end) * @start: start PFN to allocate * @end: one-past-the-last PFN to allocate * @flags: flags passed to alloc_contig_freed_pages(). + * @migratetype: migratetype of the underlaying pageblocks (either + * #MIGRATE_MOVABLE or #MIGRATE_CMA). All pageblocks + * in range must have the same migratetype and it must + * be either of the two. * * The PFN range does not have to be pageblock or MAX_ORDER_NR_PAGES * aligned, hovewer it's callers responsibility to guarantee that we @@ -5767,7 +5771,7 @@ static int __alloc_contig_migrate_range(unsigned long start, unsigned long end) * need to be freed with free_contig_pages(). */ int alloc_contig_range(unsigned long start, unsigned long end, - gfp_t flags) + gfp_t flags, unsigned migratetype) { unsigned long outer_start, outer_end; int ret; @@ -5795,8 +5799,8 @@ int alloc_contig_range(unsigned long start, unsigned long end, * them. */
- ret = start_isolate_page_range(pfn_to_maxpage(start), - pfn_to_maxpage_up(end)); + ret = __start_isolate_page_range(pfn_to_maxpage(start), + pfn_to_maxpage_up(end), migratetype); if (ret) goto done;
@@ -5834,7 +5838,8 @@ int alloc_contig_range(unsigned long start, unsigned long end,
ret = 0; done: - undo_isolate_page_range(pfn_to_maxpage(start), pfn_to_maxpage_up(end)); + __undo_isolate_page_range(pfn_to_maxpage(start), pfn_to_maxpage_up(end), + migratetype); return ret; }
diff --git a/mm/page_isolation.c b/mm/page_isolation.c index 15b41ec..f8beab5 100644 --- a/mm/page_isolation.c +++ b/mm/page_isolation.c @@ -23,10 +23,11 @@ __first_valid_page(unsigned long pfn, unsigned long nr_pages) }
/* - * start_isolate_page_range() -- make page-allocation-type of range of pages + * __start_isolate_page_range() -- make page-allocation-type of range of pages * to be MIGRATE_ISOLATE. * @start_pfn: The lower PFN of the range to be isolated. * @end_pfn: The upper PFN of the range to be isolated. + * @migratetype: migrate type to set in error recovery. * * Making page-allocation-type to be MIGRATE_ISOLATE means free pages in * the range will never be allocated. Any free pages and pages freed in the @@ -35,8 +36,8 @@ __first_valid_page(unsigned long pfn, unsigned long nr_pages) * start_pfn/end_pfn must be aligned to pageblock_order. * Returns 0 on success and -EBUSY if any part of range cannot be isolated. */ -int -start_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn) +int __start_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn, + unsigned migratetype) { unsigned long pfn; unsigned long undo_pfn; @@ -59,7 +60,7 @@ undo: for (pfn = start_pfn; pfn < undo_pfn; pfn += pageblock_nr_pages) - unset_migratetype_isolate(pfn_to_page(pfn)); + __unset_migratetype_isolate(pfn_to_page(pfn), migratetype);
return -EBUSY; } @@ -67,8 +68,8 @@ undo: /* * Make isolated pages available again. */ -int -undo_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn) +int __undo_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn, + unsigned migratetype) { unsigned long pfn; struct page *page; @@ -80,7 +81,7 @@ undo_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn) page = __first_valid_page(pfn, pageblock_nr_pages); if (!page || get_pageblock_migratetype(page) != MIGRATE_ISOLATE) continue; - unset_migratetype_isolate(page); + __unset_migratetype_isolate(page, migratetype); } return 0; }
On Tuesday 05 July 2011, Marek Szyprowski wrote:
From: Michal Nazarewicz m.nazarewicz@samsung.com
This commit changes various functions that change pages and pageblocks migrate type between MIGRATE_ISOLATE and MIGRATE_MOVABLE in such a way as to allow to work with MIGRATE_CMA migrate type.
Signed-off-by: Michal Nazarewicz m.nazarewicz@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com CC: Michal Nazarewicz mina86@mina86.com
Acked-by: Arnd Bergmann arnd@arndb.de
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com CC: Michal Nazarewicz mina86@mina86.com --- drivers/base/Kconfig | 77 +++++++++ drivers/base/Makefile | 1 + drivers/base/dma-contiguous.c | 367 ++++++++++++++++++++++++++++++++++++++++ include/linux/dma-contiguous.h | 104 +++++++++++ 4 files changed, 549 insertions(+), 0 deletions(-) create mode 100644 drivers/base/dma-contiguous.c create mode 100644 include/linux/dma-contiguous.h
diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig index d57e8d0..95ae1a7 100644 --- a/drivers/base/Kconfig +++ b/drivers/base/Kconfig @@ -168,4 +168,81 @@ config SYS_HYPERVISOR bool default n
+config CMA + bool "Contiguous Memory Allocator" + depends HAVE_DMA_CONTIGUOUS && HAVE_MEMBLOCK + select MIGRATION + select CMA_MIGRATE_TYPE + help + This enables the Contiguous Memory Allocator which allows drivers + to allocate big physically-contiguous blocks of memory for use with + hardware components that do not support I/O map nor scatter-gather. + + For more information see <include/linux/dma-contiguous.h>. + If unsure, say "n". + +if CMA + +config CMA_DEBUG + bool "CMA debug messages (DEVELOPEMENT)" + help + Turns on debug messages in CMA. This produces KERN_DEBUG + messages for every CMA call as well as various messages while + processing calls such as dma_alloc_from_contiguous(). + This option does not affect warning and error messages. + +comment "Default contiguous memory area size:" + +config CMA_SIZE_ABSOLUTE + int "Absolute size (in MiB)" + default 16 + help + Defines the size (in MiB) of the default memory area for Contiguous + Memory Allocator. + +config CMA_SIZE_PERCENTAGE + int "Percentage of total memory" + default 10 + help + Defines the size of the default memory area for Contiguous Memory + Allocator as a percentage of the total memory in the system. + +choice + prompt "Selected region size" + default CMA_SIZE_SEL_ABSOLUTE + +config CMA_SIZE_SEL_ABSOLUTE + bool "Use absolute value only" + +config CMA_SIZE_SEL_PERCENTAGE + bool "Use percentage value only" + +config CMA_SIZE_SEL_MIN + bool "Use lower value (minimum)" + +config CMA_SIZE_SEL_MAX + bool "Use higher value (maximum)" + +endchoice + +config CMA_ALIGNMENT + int "Maximum PAGE_SIZE order of alignment for contiguous buffers" + range 4 9 + default 8 + help + DMA mapping framework by default aligns all buffers to the smallest + PAGE_SIZE order which is greater than or equal to the requested buffer + size. This works well for buffers up to a few hundreds kilobytes, but + for larger buffers it just a memory waste. With this parameter you can + specify the maximum PAGE_SIZE order for contiguous buffers. Larger + buffers will be aligned only to this specified order. The order is + expressed as a power of two multiplied by the PAGE_SIZE. + + For example, if your system defaults to 4KiB pages, the order value + of 8 means that the buffers will be aligned up to 1MiB only. + + If unsure, leave the default value "8". + +endif + endmenu diff --git a/drivers/base/Makefile b/drivers/base/Makefile index 4c5701c..be6aab4 100644 --- a/drivers/base/Makefile +++ b/drivers/base/Makefile @@ -5,6 +5,7 @@ obj-y := core.o sys.o bus.o dd.o syscore.o \ cpu.o firmware.o init.o map.o devres.o \ attribute_container.o transport_class.o obj-$(CONFIG_DEVTMPFS) += devtmpfs.o +obj-$(CONFIG_CMA) += dma-contiguous.o obj-y += power/ obj-$(CONFIG_HAS_DMA) += dma-mapping.o obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o diff --git a/drivers/base/dma-contiguous.c b/drivers/base/dma-contiguous.c new file mode 100644 index 0000000..707b901 --- /dev/null +++ b/drivers/base/dma-contiguous.c @@ -0,0 +1,367 @@ +/* + * Contiguous Memory Allocator for DMA mapping framework + * Copyright (c) 2010-2011 by Samsung Electronics. + * Written by: + * Marek Szyprowski m.szyprowski@samsung.com + * Michal Nazarewicz mina86@mina86.com + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of the + * License or (at your optional) any later version of the license. + */ + +#define pr_fmt(fmt) "cma: " fmt + +#ifdef CONFIG_CMA_DEBUG +#ifndef DEBUG +# define DEBUG +#endif +#endif + +#include <asm/page.h> +#include <asm/sizes.h> + +#include <linux/memblock.h> +#include <linux/err.h> +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/mutex.h> +#include <linux/page-isolation.h> +#include <linux/slab.h> +#include <linux/swap.h> +#include <linux/mm_types.h> +#include <linux/dma-mapping.h> +#include <linux/dma-contiguous.h> + +struct cma { + unsigned long base_pfn; + unsigned long count; + unsigned long *bitmap; +}; + +struct cma *dma_contiguous_default_area; + +static unsigned long size_abs = CONFIG_CMA_SIZE_ABSOLUTE * SZ_1M; +static unsigned long size_percent = CONFIG_CMA_SIZE_PERCENTAGE; +static long size_cmdline = -1; + +static int __init early_cma(char *p) +{ + pr_debug("%s(%s)\n", __func__, p); + size_cmdline = memparse(p, &p); + return 0; +} +early_param("cma", early_cma); + +/** + * dma_contiguous_reserve() - reserve area for contiguous memory handling + * + * This funtion reserves memory from memblock subsystem. It should be + * called by arch specific code once a memblock allocator has been activated + * and all other subsystems have already allocated/reserved memory. + */ +void __init dma_contiguous_reserve(void) +{ + struct memblock_region *reg; + unsigned long selected_size = 0; + unsigned long total_pages = 0; + + pr_debug("%s()\n", __func__); + + /* + * We cannot use memblock_phys_mem_size() here, because + * memblock_analyze() has not been called yet. + */ + for_each_memblock(memory, reg) + total_pages += memblock_region_memory_end_pfn(reg) - + memblock_region_memory_base_pfn(reg); + + size_percent *= (total_pages << PAGE_SHIFT) / 100; + + pr_debug("%s: available phys mem: %ld MiB\n", __func__, + (total_pages << PAGE_SHIFT) / SZ_1M); + +#ifdef CONFIG_CMA_SIZE_SEL_ABSOLUTE + selected_size = size_abs; +#endif +#ifdef CONFIG_CMA_SIZE_SEL_PERCENTAGE + selected_size = size_percent; +#endif +#ifdef CONFIG_CMA_SIZE_SEL_MIN + selected_size = min(size_abs, size_percent); +#endif +#ifdef CONFIG_CMA_SIZE_SEL_MAX + selected_size = max(size_abs, size_percent); +#endif + + if (size_cmdline != -1) + selected_size = size_cmdline; + + if (!selected_size) + return; + + pr_debug("%s: reserving %ld MiB for global area\n", __func__, + selected_size / SZ_1M); + + dma_declare_contiguous(NULL, selected_size, 0); +}; + +static DEFINE_MUTEX(cma_mutex); + +#ifdef CONFIG_DEBUG_VM + +static int __cma_activate_area(unsigned long base_pfn, unsigned long count) +{ + unsigned long pfn = base_pfn; + unsigned i = count; + struct zone *zone; + + pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count); + + VM_BUG_ON(!pfn_valid(pfn)); + zone = page_zone(pfn_to_page(pfn)); + + do { + VM_BUG_ON(!pfn_valid(pfn)); + VM_BUG_ON(page_zone(pfn_to_page(pfn)) != zone); + if (!(pfn & (pageblock_nr_pages - 1))) + init_cma_reserved_pageblock(pfn_to_page(pfn)); + ++pfn; + } while (--i); + + return 0; +} + +#else + +static int __cma_activate_area(unsigned long base_pfn, unsigned long count) +{ + unsigned i = count >> pageblock_order; + struct page *p = pfn_to_page(base_pfn); + + pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count); + + do { + init_cma_reserved_pageblock(p); + p += pageblock_nr_pages; + } while (--i); + + return 0; +} + +#endif + +static struct cma *__cma_create_area(unsigned long base_pfn, + unsigned long count) +{ + int bitmap_size = BITS_TO_LONGS(count) * sizeof(long); + struct cma *cma; + + pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count); + + cma = kmalloc(sizeof *cma, GFP_KERNEL); + if (!cma) + return ERR_PTR(-ENOMEM); + + cma->base_pfn = base_pfn; + cma->count = count; + cma->bitmap = kzalloc(bitmap_size, GFP_KERNEL); + + if (!cma->bitmap) + goto no_mem; + + __cma_activate_area(base_pfn, count); + + pr_debug("%s: returning <%p>\n", __func__, (void *)cma); + return cma; + +no_mem: + kfree(cma); + return ERR_PTR(-ENOMEM); +} + +static struct cma_reserved { + unsigned long start; + unsigned long size; + struct device *dev; +} cma_reserved[8] __initdata; +static unsigned cma_reserved_count __initdata; + +static int __init __cma_init_reserved_areas(void) +{ + struct cma_reserved *r = cma_reserved; + unsigned i = cma_reserved_count; + + pr_debug("%s()\n", __func__); + + for (; i; --i, ++r) { + struct cma *cma; + cma = __cma_create_area(page_to_pfn(phys_to_page(r->start)), + r->size >> PAGE_SHIFT); + if (!IS_ERR(cma)) { + pr_debug("%s: created area %p\n", __func__, cma); + if (r->dev) + set_dev_cma_area(r->dev, cma); + else + dma_contiguous_default_area = cma; + } + } + return 0; +} +core_initcall(__cma_init_reserved_areas); + +/** + * dma_declare_contiguous() - reserve area for contiguous memory handling + * for particular device + * @dev: Pointer to device structure. + * @size: Size of the reserved memory. + * @start: Start address of the reserved memory (optional, 0 for any). + * + * This funtion reserves memory for specified device. It should be + * called by board specific code once a memblock allocator has been activated + * and all other subsystems have already allocated/reserved memory. + */ +int __init dma_declare_contiguous(struct device *dev, unsigned long size, + phys_addr_t start) +{ + struct cma_reserved *r = &cma_reserved[cma_reserved_count]; + unsigned long alignment; + + pr_debug("%s(%p+%p)\n", __func__, (void *)start, (void *)size); + + /* Sanity checks */ + if (cma_reserved_count == ARRAY_SIZE(cma_reserved)) + return -ENOSPC; + + if (!size) + return -EINVAL; + + /* Sanitise input arguments */ + alignment = PAGE_SIZE << (MAX_ORDER + 1); + start = ALIGN(start, alignment); + size = ALIGN(size , alignment); + + /* Reserve memory */ + if (start) { + if (memblock_is_region_reserved(start, size) || + memblock_reserve(start, size) < 0) + return -EBUSY; + } else { + /* + * Use __memblock_alloc_base() since + * memblock_alloc_base() panic()s. + */ + u64 addr = __memblock_alloc_base(size, alignment, 0); + if (!addr) { + return -ENOMEM; + } else if (addr + size > ~(unsigned long)0) { + memblock_free(addr, size); + return -EOVERFLOW; + } else { + start = addr; + } + } + + /* + * Each reserved area must be initialised later, when more kernel + * subsystems (like slab allocator) are available. + */ + r->start = start; + r->size = size; + r->dev = dev; + cma_reserved_count++; + printk(KERN_INFO "%s: reserved %ld MiB area at 0x%p\n", __func__, + size / SZ_1M, (void *)start); + return 0; +} + +/** + * dma_alloc_from_contiguous() - allocate pages from contiguous area + * @dev: Pointer to device for which the allocation is performed. + * @count: Requested number of pages. + * @align: Requested alignment of pages (in PAGE_SIZE order). + * + * This funtion allocates memory buffer for specified device. It uses + * device specific contiguous memory area if available or the default + * global one. Requires architecture specific get_dev_cma_area() helper + * function. + */ +struct page *dma_alloc_from_contiguous(struct device *dev, int count, + unsigned int align) +{ + struct cma *cma = get_dev_cma_area(dev); + unsigned long pfn, pageno; + int ret; + + if (!cma) + return NULL; + + if (align > CONFIG_CMA_ALIGNMENT) + align = CONFIG_CMA_ALIGNMENT; + + pr_debug("%s(<%p>, %d/%d)\n", __func__, (void *)cma, count, align); + + if (!count) + return NULL; + + mutex_lock(&cma_mutex); + + pageno = bitmap_find_next_zero_area(cma->bitmap, cma->count, 0, count, + align); + if (pageno >= cma->count) { + ret = -ENOMEM; + goto error; + } + bitmap_set(cma->bitmap, pageno, count); + + pfn = cma->base_pfn + pageno; + ret = alloc_contig_range(pfn, pfn + count, 0, MIGRATE_CMA); + if (ret) + goto free; + + mutex_unlock(&cma_mutex); + + pr_debug("%s(): returning [%ld]\n", __func__, pfn); + return pfn_to_page(pfn); +free: + bitmap_clear(cma->bitmap, pageno, count); +error: + mutex_unlock(&cma_mutex); + return NULL; +} + +/** + * dma_release_from_contiguous() - release allocated pages + * @dev: Pointer to device for which the pages were allocated. + * @pages: Allocated pages. + * @count: Number of allocated pages. + * + * This funtion reserves memory for specified device. It should be + * called by board specific code once a memblock allocator has been activated + * and all other subsystems have already allocated/reserved memory. + */ +int dma_release_from_contiguous(struct device *dev, struct page *pages, + int count) +{ + struct cma *cma = get_dev_cma_area(dev); + unsigned long pfn; + + if (!cma || !pages) + return 0; + + pr_debug("%s([%p])\n", __func__, (void *)pages); + + pfn = page_to_pfn(pages); + + if (pfn < cma->base_pfn || pfn >= cma->base_pfn + count) + return 0; + + mutex_lock(&cma_mutex); + + bitmap_clear(cma->bitmap, pfn - cma->base_pfn, count); + free_contig_pages(pages, count); + + mutex_unlock(&cma_mutex); + return 1; +} diff --git a/include/linux/dma-contiguous.h b/include/linux/dma-contiguous.h new file mode 100644 index 0000000..98312c9 --- /dev/null +++ b/include/linux/dma-contiguous.h @@ -0,0 +1,104 @@ +#ifndef __LINUX_CMA_H +#define __LINUX_CMA_H + +/* + * Contiguous Memory Allocator for DMA mapping framework + * Copyright (c) 2010-2011 by Samsung Electronics. + * Written by: + * Marek Szyprowski m.szyprowski@samsung.com + * Michal Nazarewicz mina86@mina86.com + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of the + * License or (at your optional) any later version of the license. + */ + +/* + * Contiguous Memory Allocator + * + * The Contiguous Memory Allocator (CMA) makes it possible to + * allocate big contiguous chunks of memory after the system has + * booted. + * + * Why is it needed? + * + * Various devices on embedded systems have no scatter-getter and/or + * IO map support and require contiguous blocks of memory to + * operate. They include devices such as cameras, hardware video + * coders, etc. + * + * Such devices often require big memory buffers (a full HD frame + * is, for instance, more then 2 mega pixels large, i.e. more than 6 + * MB of memory), which makes mechanisms such as kmalloc() or + * alloc_page() ineffective. + * + * At the same time, a solution where a big memory region is + * reserved for a device is suboptimal since often more memory is + * reserved then strictly required and, moreover, the memory is + * inaccessible to page system even if device drivers don't use it. + * + * CMA tries to solve this issue by operating on memory regions + * where only movable pages can be allocated from. This way, kernel + * can use the memory for pagecache and when device driver requests + * it, allocated pages can be migrated. + * + * Driver usage + * + * CMA should not be used by the device drivers directly. It is + * only a helper framework for dma-mapping subsystem. + * + * For more information, see kernel-docs in drivers/base/dma-contiguous.c + */ + +#ifdef __KERNEL__ + +struct cma; +struct page; +struct device; + +#ifdef CONFIG_CMA + +extern struct cma *dma_contiguous_default_area; + +void dma_contiguous_reserve(void); +int dma_declare_contiguous(struct device *dev, unsigned long size, + phys_addr_t base); + +struct page *dma_alloc_from_contiguous(struct device *dev, int count, + unsigned int order); +int dma_release_from_contiguous(struct device *dev, struct page *pages, + int count); + +#else + +#define dna_contiguous_default_area NULL + +static inline void dma_contiguous_reserve(void) { } + +static inline +int dma_declare_contiguous(struct device *dev, unsigned long size, + unsigned long base) +{ + return -EINVAL; +} + +static inline +struct page *dma_alloc_from_contiguous(struct device *dev, int count, + unsigned int order) +{ + return NULL; +} + +static inline +int dma_release_from_contiguous(struct device *dev, struct page *pages, + int count) +{ + return 0; +} + +#endif + +#endif + +#endif
Hello,
On Tuesday, July 05, 2011 9:42 AM Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com CC: Michal Nazarewicz mina86@mina86.com
drivers/base/Kconfig | 77 +++++++++ drivers/base/Makefile | 1 + drivers/base/dma-contiguous.c | 367 ++++++++++++++++++++++++++++++++++++++++ include/linux/dma-contiguous.h | 104 +++++++++++ 4 files changed, 549 insertions(+), 0 deletions(-) create mode 100644 drivers/base/dma-contiguous.c create mode 100644 include/linux/dma-contiguous.h
diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig index d57e8d0..95ae1a7 100644 --- a/drivers/base/Kconfig +++ b/drivers/base/Kconfig @@ -168,4 +168,81 @@ config SYS_HYPERVISOR bool default n
+config CMA
- bool "Contiguous Memory Allocator"
- depends HAVE_DMA_CONTIGUOUS && HAVE_MEMBLOCK
The above line should be obviously "depends on HAVE_DMA_CONTIGUOUS && HAVE_MEMBLOCK". I'm sorry for posting broken version.
(snipped)
Best regards
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com CC: Michal Nazarewicz mina86@mina86.com --- arch/Kconfig | 3 + drivers/base/Kconfig | 77 +++++++++ drivers/base/Makefile | 1 + drivers/base/dma-contiguous.c | 367 ++++++++++++++++++++++++++++++++++++++++ include/linux/dma-contiguous.h | 104 +++++++++++ 5 files changed, 552 insertions(+), 0 deletions(-) create mode 100644 drivers/base/dma-contiguous.c create mode 100644 include/linux/dma-contiguous.h
diff --git a/arch/Kconfig b/arch/Kconfig index 26b0e23..228d761 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -124,6 +124,9 @@ config HAVE_ARCH_TRACEHOOK config HAVE_DMA_ATTRS bool
+config HAVE_DMA_CONTIGUOUS + bool + config USE_GENERIC_SMP_HELPERS bool
diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig index d57e8d0..c690d05 100644 --- a/drivers/base/Kconfig +++ b/drivers/base/Kconfig @@ -168,4 +168,81 @@ config SYS_HYPERVISOR bool default n
+config CMA + bool "Contiguous Memory Allocator" + depends on HAVE_DMA_CONTIGUOUS && HAVE_MEMBLOCK + select MIGRATION + select CMA_MIGRATE_TYPE + help + This enables the Contiguous Memory Allocator which allows drivers + to allocate big physically-contiguous blocks of memory for use with + hardware components that do not support I/O map nor scatter-gather. + + For more information see <include/linux/dma-contiguous.h>. + If unsure, say "n". + +if CMA + +config CMA_DEBUG + bool "CMA debug messages (DEVELOPEMENT)" + help + Turns on debug messages in CMA. This produces KERN_DEBUG + messages for every CMA call as well as various messages while + processing calls such as dma_alloc_from_contiguous(). + This option does not affect warning and error messages. + +comment "Default contiguous memory area size:" + +config CMA_SIZE_ABSOLUTE + int "Absolute size (in MiB)" + default 16 + help + Defines the size (in MiB) of the default memory area for Contiguous + Memory Allocator. + +config CMA_SIZE_PERCENTAGE + int "Percentage of total memory" + default 10 + help + Defines the size of the default memory area for Contiguous Memory + Allocator as a percentage of the total memory in the system. + +choice + prompt "Selected region size" + default CMA_SIZE_SEL_ABSOLUTE + +config CMA_SIZE_SEL_ABSOLUTE + bool "Use absolute value only" + +config CMA_SIZE_SEL_PERCENTAGE + bool "Use percentage value only" + +config CMA_SIZE_SEL_MIN + bool "Use lower value (minimum)" + +config CMA_SIZE_SEL_MAX + bool "Use higher value (maximum)" + +endchoice + +config CMA_ALIGNMENT + int "Maximum PAGE_SIZE order of alignment for contiguous buffers" + range 4 9 + default 8 + help + DMA mapping framework by default aligns all buffers to the smallest + PAGE_SIZE order which is greater than or equal to the requested buffer + size. This works well for buffers up to a few hundreds kilobytes, but + for larger buffers it just a memory waste. With this parameter you can + specify the maximum PAGE_SIZE order for contiguous buffers. Larger + buffers will be aligned only to this specified order. The order is + expressed as a power of two multiplied by the PAGE_SIZE. + + For example, if your system defaults to 4KiB pages, the order value + of 8 means that the buffers will be aligned up to 1MiB only. + + If unsure, leave the default value "8". + +endif + endmenu diff --git a/drivers/base/Makefile b/drivers/base/Makefile index 4c5701c..be6aab4 100644 --- a/drivers/base/Makefile +++ b/drivers/base/Makefile @@ -5,6 +5,7 @@ obj-y := core.o sys.o bus.o dd.o syscore.o \ cpu.o firmware.o init.o map.o devres.o \ attribute_container.o transport_class.o obj-$(CONFIG_DEVTMPFS) += devtmpfs.o +obj-$(CONFIG_CMA) += dma-contiguous.o obj-y += power/ obj-$(CONFIG_HAS_DMA) += dma-mapping.o obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o diff --git a/drivers/base/dma-contiguous.c b/drivers/base/dma-contiguous.c new file mode 100644 index 0000000..707b901 --- /dev/null +++ b/drivers/base/dma-contiguous.c @@ -0,0 +1,367 @@ +/* + * Contiguous Memory Allocator for DMA mapping framework + * Copyright (c) 2010-2011 by Samsung Electronics. + * Written by: + * Marek Szyprowski m.szyprowski@samsung.com + * Michal Nazarewicz mina86@mina86.com + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of the + * License or (at your optional) any later version of the license. + */ + +#define pr_fmt(fmt) "cma: " fmt + +#ifdef CONFIG_CMA_DEBUG +#ifndef DEBUG +# define DEBUG +#endif +#endif + +#include <asm/page.h> +#include <asm/sizes.h> + +#include <linux/memblock.h> +#include <linux/err.h> +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/mutex.h> +#include <linux/page-isolation.h> +#include <linux/slab.h> +#include <linux/swap.h> +#include <linux/mm_types.h> +#include <linux/dma-mapping.h> +#include <linux/dma-contiguous.h> + +struct cma { + unsigned long base_pfn; + unsigned long count; + unsigned long *bitmap; +}; + +struct cma *dma_contiguous_default_area; + +static unsigned long size_abs = CONFIG_CMA_SIZE_ABSOLUTE * SZ_1M; +static unsigned long size_percent = CONFIG_CMA_SIZE_PERCENTAGE; +static long size_cmdline = -1; + +static int __init early_cma(char *p) +{ + pr_debug("%s(%s)\n", __func__, p); + size_cmdline = memparse(p, &p); + return 0; +} +early_param("cma", early_cma); + +/** + * dma_contiguous_reserve() - reserve area for contiguous memory handling + * + * This funtion reserves memory from memblock subsystem. It should be + * called by arch specific code once a memblock allocator has been activated + * and all other subsystems have already allocated/reserved memory. + */ +void __init dma_contiguous_reserve(void) +{ + struct memblock_region *reg; + unsigned long selected_size = 0; + unsigned long total_pages = 0; + + pr_debug("%s()\n", __func__); + + /* + * We cannot use memblock_phys_mem_size() here, because + * memblock_analyze() has not been called yet. + */ + for_each_memblock(memory, reg) + total_pages += memblock_region_memory_end_pfn(reg) - + memblock_region_memory_base_pfn(reg); + + size_percent *= (total_pages << PAGE_SHIFT) / 100; + + pr_debug("%s: available phys mem: %ld MiB\n", __func__, + (total_pages << PAGE_SHIFT) / SZ_1M); + +#ifdef CONFIG_CMA_SIZE_SEL_ABSOLUTE + selected_size = size_abs; +#endif +#ifdef CONFIG_CMA_SIZE_SEL_PERCENTAGE + selected_size = size_percent; +#endif +#ifdef CONFIG_CMA_SIZE_SEL_MIN + selected_size = min(size_abs, size_percent); +#endif +#ifdef CONFIG_CMA_SIZE_SEL_MAX + selected_size = max(size_abs, size_percent); +#endif + + if (size_cmdline != -1) + selected_size = size_cmdline; + + if (!selected_size) + return; + + pr_debug("%s: reserving %ld MiB for global area\n", __func__, + selected_size / SZ_1M); + + dma_declare_contiguous(NULL, selected_size, 0); +}; + +static DEFINE_MUTEX(cma_mutex); + +#ifdef CONFIG_DEBUG_VM + +static int __cma_activate_area(unsigned long base_pfn, unsigned long count) +{ + unsigned long pfn = base_pfn; + unsigned i = count; + struct zone *zone; + + pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count); + + VM_BUG_ON(!pfn_valid(pfn)); + zone = page_zone(pfn_to_page(pfn)); + + do { + VM_BUG_ON(!pfn_valid(pfn)); + VM_BUG_ON(page_zone(pfn_to_page(pfn)) != zone); + if (!(pfn & (pageblock_nr_pages - 1))) + init_cma_reserved_pageblock(pfn_to_page(pfn)); + ++pfn; + } while (--i); + + return 0; +} + +#else + +static int __cma_activate_area(unsigned long base_pfn, unsigned long count) +{ + unsigned i = count >> pageblock_order; + struct page *p = pfn_to_page(base_pfn); + + pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count); + + do { + init_cma_reserved_pageblock(p); + p += pageblock_nr_pages; + } while (--i); + + return 0; +} + +#endif + +static struct cma *__cma_create_area(unsigned long base_pfn, + unsigned long count) +{ + int bitmap_size = BITS_TO_LONGS(count) * sizeof(long); + struct cma *cma; + + pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count); + + cma = kmalloc(sizeof *cma, GFP_KERNEL); + if (!cma) + return ERR_PTR(-ENOMEM); + + cma->base_pfn = base_pfn; + cma->count = count; + cma->bitmap = kzalloc(bitmap_size, GFP_KERNEL); + + if (!cma->bitmap) + goto no_mem; + + __cma_activate_area(base_pfn, count); + + pr_debug("%s: returning <%p>\n", __func__, (void *)cma); + return cma; + +no_mem: + kfree(cma); + return ERR_PTR(-ENOMEM); +} + +static struct cma_reserved { + unsigned long start; + unsigned long size; + struct device *dev; +} cma_reserved[8] __initdata; +static unsigned cma_reserved_count __initdata; + +static int __init __cma_init_reserved_areas(void) +{ + struct cma_reserved *r = cma_reserved; + unsigned i = cma_reserved_count; + + pr_debug("%s()\n", __func__); + + for (; i; --i, ++r) { + struct cma *cma; + cma = __cma_create_area(page_to_pfn(phys_to_page(r->start)), + r->size >> PAGE_SHIFT); + if (!IS_ERR(cma)) { + pr_debug("%s: created area %p\n", __func__, cma); + if (r->dev) + set_dev_cma_area(r->dev, cma); + else + dma_contiguous_default_area = cma; + } + } + return 0; +} +core_initcall(__cma_init_reserved_areas); + +/** + * dma_declare_contiguous() - reserve area for contiguous memory handling + * for particular device + * @dev: Pointer to device structure. + * @size: Size of the reserved memory. + * @start: Start address of the reserved memory (optional, 0 for any). + * + * This funtion reserves memory for specified device. It should be + * called by board specific code once a memblock allocator has been activated + * and all other subsystems have already allocated/reserved memory. + */ +int __init dma_declare_contiguous(struct device *dev, unsigned long size, + phys_addr_t start) +{ + struct cma_reserved *r = &cma_reserved[cma_reserved_count]; + unsigned long alignment; + + pr_debug("%s(%p+%p)\n", __func__, (void *)start, (void *)size); + + /* Sanity checks */ + if (cma_reserved_count == ARRAY_SIZE(cma_reserved)) + return -ENOSPC; + + if (!size) + return -EINVAL; + + /* Sanitise input arguments */ + alignment = PAGE_SIZE << (MAX_ORDER + 1); + start = ALIGN(start, alignment); + size = ALIGN(size , alignment); + + /* Reserve memory */ + if (start) { + if (memblock_is_region_reserved(start, size) || + memblock_reserve(start, size) < 0) + return -EBUSY; + } else { + /* + * Use __memblock_alloc_base() since + * memblock_alloc_base() panic()s. + */ + u64 addr = __memblock_alloc_base(size, alignment, 0); + if (!addr) { + return -ENOMEM; + } else if (addr + size > ~(unsigned long)0) { + memblock_free(addr, size); + return -EOVERFLOW; + } else { + start = addr; + } + } + + /* + * Each reserved area must be initialised later, when more kernel + * subsystems (like slab allocator) are available. + */ + r->start = start; + r->size = size; + r->dev = dev; + cma_reserved_count++; + printk(KERN_INFO "%s: reserved %ld MiB area at 0x%p\n", __func__, + size / SZ_1M, (void *)start); + return 0; +} + +/** + * dma_alloc_from_contiguous() - allocate pages from contiguous area + * @dev: Pointer to device for which the allocation is performed. + * @count: Requested number of pages. + * @align: Requested alignment of pages (in PAGE_SIZE order). + * + * This funtion allocates memory buffer for specified device. It uses + * device specific contiguous memory area if available or the default + * global one. Requires architecture specific get_dev_cma_area() helper + * function. + */ +struct page *dma_alloc_from_contiguous(struct device *dev, int count, + unsigned int align) +{ + struct cma *cma = get_dev_cma_area(dev); + unsigned long pfn, pageno; + int ret; + + if (!cma) + return NULL; + + if (align > CONFIG_CMA_ALIGNMENT) + align = CONFIG_CMA_ALIGNMENT; + + pr_debug("%s(<%p>, %d/%d)\n", __func__, (void *)cma, count, align); + + if (!count) + return NULL; + + mutex_lock(&cma_mutex); + + pageno = bitmap_find_next_zero_area(cma->bitmap, cma->count, 0, count, + align); + if (pageno >= cma->count) { + ret = -ENOMEM; + goto error; + } + bitmap_set(cma->bitmap, pageno, count); + + pfn = cma->base_pfn + pageno; + ret = alloc_contig_range(pfn, pfn + count, 0, MIGRATE_CMA); + if (ret) + goto free; + + mutex_unlock(&cma_mutex); + + pr_debug("%s(): returning [%ld]\n", __func__, pfn); + return pfn_to_page(pfn); +free: + bitmap_clear(cma->bitmap, pageno, count); +error: + mutex_unlock(&cma_mutex); + return NULL; +} + +/** + * dma_release_from_contiguous() - release allocated pages + * @dev: Pointer to device for which the pages were allocated. + * @pages: Allocated pages. + * @count: Number of allocated pages. + * + * This funtion reserves memory for specified device. It should be + * called by board specific code once a memblock allocator has been activated + * and all other subsystems have already allocated/reserved memory. + */ +int dma_release_from_contiguous(struct device *dev, struct page *pages, + int count) +{ + struct cma *cma = get_dev_cma_area(dev); + unsigned long pfn; + + if (!cma || !pages) + return 0; + + pr_debug("%s([%p])\n", __func__, (void *)pages); + + pfn = page_to_pfn(pages); + + if (pfn < cma->base_pfn || pfn >= cma->base_pfn + count) + return 0; + + mutex_lock(&cma_mutex); + + bitmap_clear(cma->bitmap, pfn - cma->base_pfn, count); + free_contig_pages(pages, count); + + mutex_unlock(&cma_mutex); + return 1; +} diff --git a/include/linux/dma-contiguous.h b/include/linux/dma-contiguous.h new file mode 100644 index 0000000..98312c9 --- /dev/null +++ b/include/linux/dma-contiguous.h @@ -0,0 +1,104 @@ +#ifndef __LINUX_CMA_H +#define __LINUX_CMA_H + +/* + * Contiguous Memory Allocator for DMA mapping framework + * Copyright (c) 2010-2011 by Samsung Electronics. + * Written by: + * Marek Szyprowski m.szyprowski@samsung.com + * Michal Nazarewicz mina86@mina86.com + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of the + * License or (at your optional) any later version of the license. + */ + +/* + * Contiguous Memory Allocator + * + * The Contiguous Memory Allocator (CMA) makes it possible to + * allocate big contiguous chunks of memory after the system has + * booted. + * + * Why is it needed? + * + * Various devices on embedded systems have no scatter-getter and/or + * IO map support and require contiguous blocks of memory to + * operate. They include devices such as cameras, hardware video + * coders, etc. + * + * Such devices often require big memory buffers (a full HD frame + * is, for instance, more then 2 mega pixels large, i.e. more than 6 + * MB of memory), which makes mechanisms such as kmalloc() or + * alloc_page() ineffective. + * + * At the same time, a solution where a big memory region is + * reserved for a device is suboptimal since often more memory is + * reserved then strictly required and, moreover, the memory is + * inaccessible to page system even if device drivers don't use it. + * + * CMA tries to solve this issue by operating on memory regions + * where only movable pages can be allocated from. This way, kernel + * can use the memory for pagecache and when device driver requests + * it, allocated pages can be migrated. + * + * Driver usage + * + * CMA should not be used by the device drivers directly. It is + * only a helper framework for dma-mapping subsystem. + * + * For more information, see kernel-docs in drivers/base/dma-contiguous.c + */ + +#ifdef __KERNEL__ + +struct cma; +struct page; +struct device; + +#ifdef CONFIG_CMA + +extern struct cma *dma_contiguous_default_area; + +void dma_contiguous_reserve(void); +int dma_declare_contiguous(struct device *dev, unsigned long size, + phys_addr_t base); + +struct page *dma_alloc_from_contiguous(struct device *dev, int count, + unsigned int order); +int dma_release_from_contiguous(struct device *dev, struct page *pages, + int count); + +#else + +#define dna_contiguous_default_area NULL + +static inline void dma_contiguous_reserve(void) { } + +static inline +int dma_declare_contiguous(struct device *dev, unsigned long size, + unsigned long base) +{ + return -EINVAL; +} + +static inline +struct page *dma_alloc_from_contiguous(struct device *dev, int count, + unsigned int order) +{ + return NULL; +} + +static inline +int dma_release_from_contiguous(struct device *dev, struct page *pages, + int count) +{ + return 0; +} + +#endif + +#endif + +#endif
On Tuesday 05 July 2011, Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com CC: Michal Nazarewicz mina86@mina86.com
Reviewed-by: Arnd Bergmann arnd@arndb.de, but I noticed two one-character mistakes:
+if CMA
+config CMA_DEBUG
- bool "CMA debug messages (DEVELOPEMENT)"
s/DEVELOPEMENT/DEVELOPMENT/
diff --git a/drivers/base/Makefile b/drivers/base/Makefile index 4c5701c..be6aab4 100644 --- a/drivers/base/Makefile +++ b/drivers/base/Makefile @@ -5,6 +5,7 @@ obj-y := core.o sys.o bus.o dd.o syscore.o \ cpu.o firmware.o init.o map.o devres.o \ attribute_container.o transport_class.o obj-$(CONFIG_DEVTMPFS) += devtmpfs.o +obj-$(CONFIG_CMA) += dma-contiguous.o obj-y += power/ obj-$(CONFIG_HAS_DMA) += dma-mapping.o
Please add another tab to indent the line in the same way as the others.
Arnd
On Tue, Jul 05, 2011 at 09:41:48AM +0200, Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
And how are you addressing the technical concerns about aliasing of cache attributes which I keep bringing up with this and you keep ignoring and telling me that I'm standing in your way.
On Tuesday 05 July 2011, Russell King - ARM Linux wrote:
On Tue, Jul 05, 2011 at 09:41:48AM +0200, Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
And how are you addressing the technical concerns about aliasing of cache attributes which I keep bringing up with this and you keep ignoring and telling me that I'm standing in your way.
This is of course an important issue, and it's the one item listed as TODO in the introductory mail that sent.
It's also a preexisting problem as far as I can tell, and it needs to be solved in __dma_alloc for both cases, dma_alloc_from_contiguous and __alloc_system_pages as introduced in patch 7.
We've discussed this back and forth, and it always comes down to one of two ugly solutions:
1. Put all of the MIGRATE_CMA and pages into highmem and change __alloc_system_pages so it also allocates only from highmem pages. The consequences of this are that we always need to build kernels with highmem enabled and that we have less lowmem on systems that are already small, both of which can be fairly expensive unless you have lots of highmem already.
2. Add logic to unmap pages from the linear mapping, which is very expensive because it forces the use of small pages in the linear mapping (or in parts of it), and possibly means walking all page tables to remove the PTEs on alloc and put them back in on free.
I believe that Chunsang Jeong from Linaro is planning to implement both variants and post them for review, so we can decide which one to merge, or even to merge both and make it a configuration option. See also https://blueprints.launchpad.net/linaro-mm-sig/+spec/engr-mm-dma-mapping-201...
I don't think we need to make merging the CMA patches depending on the other patches, it's clear that both need to be solved, and they are independent enough.
Arnd
On Tue, Jul 05, 2011 at 02:27:44PM +0200, Arnd Bergmann wrote:
It's also a preexisting problem as far as I can tell, and it needs to be solved in __dma_alloc for both cases, dma_alloc_from_contiguous and __alloc_system_pages as introduced in patch 7.
Which is now resolved in linux-next, and has been through this cycle as previously discussed.
It's taken some time because the guy who tested the patch for me said he'd review other platforms but never did, so I've just about given up waiting and stuffed it in ready for the 3.1 merge window irrespective of anything else.
On Tuesday 05 July 2011, Russell King - ARM Linux wrote:
On Tue, Jul 05, 2011 at 02:27:44PM +0200, Arnd Bergmann wrote:
It's also a preexisting problem as far as I can tell, and it needs to be solved in __dma_alloc for both cases, dma_alloc_from_contiguous and __alloc_system_pages as introduced in patch 7.
Which is now resolved in linux-next, and has been through this cycle as previously discussed.
It's taken some time because the guy who tested the patch for me said he'd review other platforms but never did, so I've just about given up waiting and stuffed it in ready for the 3.1 merge window irrespective of anything else.
Ah, sorry I missed that patch on the mailing list, found it now in your for-next branch.
If I'm reading your "ARM: DMA: steal memory for DMA coherent mappings" correctly, the idea is to have a per-platform compile-time amount of memory that is reserved purely for coherent allocations and taking out of the buddy allocator, right?
As you say, this solves the problem for the non-CMA case, and does not apply to CMA because the entire point of CMA is not to remove the pages from the buddy allocator in order to preserve memory.
So with your patch getting merged, patch 7/8 obviously has both a conflict and introduces a regression against the fix you did. Consequently that patch needs to be redone in a way that fits on top of your patch and avoids the double-mapping problem.
What about the rest? As I mentioned in private, adding invasive features to core code is obviously not nice if it can be avoided, but my feeling is that we can no longer claim that there is no need for this with so much hardware relying on large contiguous memory ranges for DMA. The patches have come a long way since the first version, especially regarding the device driver interface and I think they are about as good as it gets in that regard.
I do understand that without patch 7, there isn't a single architecture using the feature, which is somewhat silly, but I'm also convinced that other architectures will start using it, and that a solution for the double mapping in the ways I mentioned in my previous mail is going to happen. Probably not in 3.1 then, but we could put the patches into -mm anyway until we get there.
Arnd
On Tue, Jul 05, 2011 at 03:58:39PM +0200, Arnd Bergmann wrote:
Ah, sorry I missed that patch on the mailing list, found it now in your for-next branch.
I've been searching for this email to reply to for the last day or so...
If I'm reading your "ARM: DMA: steal memory for DMA coherent mappings" correctly, the idea is to have a per-platform compile-time amount of memory that is reserved purely for coherent allocations and taking out of the buddy allocator, right?
Yes, because every time I've looked at taking out memory mappings in the first level page tables, it's always been a major issue.
We have a method where we can remove first level mappings on uniprocessor systems in the ioremap code just fine - we use that so that systems can setup section and supersection mappings. They can tear them down as well - and we update other tasks L1 page tables when they get switched in.
This, however, doesn't work on SMP, because if you have a DMA allocation (which is permitted from IRQ context) you must have some way of removing the L1 page table entries from all CPUs TLBs and the page tables currently in use and any future page tables which those CPUs may switch to.
The easy bit is "future page tables" - that can be done in the same way as the ioremap() code does with a generation number, checked when a new page table is switched in. The problem is the current CPUs, and as we know trying to call smp_call_function() with IRQs disabled is not permitted due to deadlock.
So, in a SMP system, there is no safe way to remove L1 page table entries from IRQ context. That means if memory is mapped for the buddy allocators using L1 page table entries, then it is fixed for that application on a SMP system.
However, that's not really what I wanted to find this email for. That is I'm dropping the "ARM: DMA: steal memory for DMA coherent mappings" patch for this merge window because - as I found out yesterday - it prevents the Assabet platform booting, and so would be a regression.
Plus, I have a report of a regression with the streaming DMA API speculative prefetch fixes causing the IOP ADMA raid5 async offload stuff to explode - which may result in the streaming DMA API fixes being reverted (which will leave ARMv6+ vulnerable to data corruption.) As I have no time to work through the RAID5 code, async_tx code, and IOP ADMA code to get to the bottom of it (because of this flood of patches) I think a revert is looking likely - either that or I'll have to tell the bug reporter to go away, which really isn't on. It's on LKML if anyone's interested in trying to diagnose it, the "PROBLEM: ARM-dma-mapping-fix-for-speculative-prefetching cause OOPS" thread.
On Friday 08 July 2011, Russell King - ARM Linux wrote:
On Tue, Jul 05, 2011 at 03:58:39PM +0200, Arnd Bergmann wrote:
If I'm reading your "ARM: DMA: steal memory for DMA coherent mappings" correctly, the idea is to have a per-platform compile-time amount of memory that is reserved purely for coherent allocations and taking out of the buddy allocator, right?
Yes, because every time I've looked at taking out memory mappings in the first level page tables, it's always been a major issue.
We have a method where we can remove first level mappings on uniprocessor systems in the ioremap code just fine - we use that so that systems can setup section and supersection mappings. They can tear them down as well - and we update other tasks L1 page tables when they get switched in.
This, however, doesn't work on SMP, because if you have a DMA allocation (which is permitted from IRQ context) you must have some way of removing the L1 page table entries from all CPUs TLBs and the page tables currently in use and any future page tables which those CPUs may switch to.
Ah, interesting. So there is no tlb flush broadcast operation and it always goes through IPI?
So, in a SMP system, there is no safe way to remove L1 page table entries from IRQ context. That means if memory is mapped for the buddy allocators using L1 page table entries, then it is fixed for that application on a SMP system.
Ok. Can we limit GFP_ATOMIC to memory that doesn't need to be remapped then? I guess we can assume that there is no regression if we just skip the dma_alloc_contiguous step in dma_alloc_coherent for any atomic callers and immediately fall back to the regular allocator.
Unfortunately, this still means we have to keep both methods. I was hoping that with CMA doing dynamic remapping there would be no need for keeping a significant number of pages reserved for this.
Arnd
[cc to ks-discuss added, since this may be a relevant topic]
On Tue, 2011-07-05 at 14:27 +0200, Arnd Bergmann wrote:
On Tuesday 05 July 2011, Russell King - ARM Linux wrote:
On Tue, Jul 05, 2011 at 09:41:48AM +0200, Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
And how are you addressing the technical concerns about aliasing of cache attributes which I keep bringing up with this and you keep ignoring and telling me that I'm standing in your way.
Just to chime in here, parisc has an identical issue. If the CPU ever sees an alias with different attributes for the same page, it will HPMC the box (that's basically the bios will kill the system as being architecturally inconsistent), so an architecture neutral solution on this point is essential to us as well.
This is of course an important issue, and it's the one item listed as TODO in the introductory mail that sent.
It's also a preexisting problem as far as I can tell, and it needs to be solved in __dma_alloc for both cases, dma_alloc_from_contiguous and __alloc_system_pages as introduced in patch 7.
We've discussed this back and forth, and it always comes down to one of two ugly solutions:
- Put all of the MIGRATE_CMA and pages into highmem and change
__alloc_system_pages so it also allocates only from highmem pages. The consequences of this are that we always need to build kernels with highmem enabled and that we have less lowmem on systems that are already small, both of which can be fairly expensive unless you have lots of highmem already.
So this would require that systems using the API have a highmem? (parisc doesn't today).
- Add logic to unmap pages from the linear mapping, which is
very expensive because it forces the use of small pages in the linear mapping (or in parts of it), and possibly means walking all page tables to remove the PTEs on alloc and put them back in on free.
I believe that Chunsang Jeong from Linaro is planning to implement both variants and post them for review, so we can decide which one to merge, or even to merge both and make it a configuration option. See also https://blueprints.launchpad.net/linaro-mm-sig/+spec/engr-mm-dma-mapping-201...
I don't think we need to make merging the CMA patches depending on the other patches, it's clear that both need to be solved, and they are independent enough.
I assume from the above that ARM has a hardware page walker?
The way I'd fix this on parisc, because we have a software based TLB, is to rely on the fact that a page may only be used either for DMA or for Page Cache, so the aliases should never be interleaved. Since you know the point at which the page flips from DMA to Cache (and vice versa), I'd purge the TLB entry and flush the page at that point and rely on the usage guarantees to ensure that the alias TLB entry doesn't reappear. This isn't inexpensive but the majority of the cost is the cache flush which is a requirement to clean the aliases anyway (a TLB entry purge is pretty cheap).
Would this work for the ARM hardware walker as well? It would require you to have a TLB entry purge instruction as well as some architectural guarantees about not speculating the TLB.
James
Hello,
I'm sorry for the late reply. I must have missed this mail...
On Wednesday, August 03, 2011 7:44 PM James Bottomley wrote:
[cc to ks-discuss added, since this may be a relevant topic]
On Tue, 2011-07-05 at 14:27 +0200, Arnd Bergmann wrote:
On Tuesday 05 July 2011, Russell King - ARM Linux wrote:
On Tue, Jul 05, 2011 at 09:41:48AM +0200, Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
And how are you addressing the technical concerns about aliasing of cache attributes which I keep bringing up with this and you keep ignoring and telling me that I'm standing in your way.
Just to chime in here, parisc has an identical issue. If the CPU ever sees an alias with different attributes for the same page, it will HPMC the box (that's basically the bios will kill the system as being architecturally inconsistent), so an architecture neutral solution on this point is essential to us as well.
This is of course an important issue, and it's the one item listed as TODO in the introductory mail that sent.
It's also a preexisting problem as far as I can tell, and it needs to be solved in __dma_alloc for both cases, dma_alloc_from_contiguous and __alloc_system_pages as introduced in patch 7.
We've discussed this back and forth, and it always comes down to one of two ugly solutions:
- Put all of the MIGRATE_CMA and pages into highmem and change
__alloc_system_pages so it also allocates only from highmem pages. The consequences of this are that we always need to build kernels with highmem enabled and that we have less lowmem on systems that are already small, both of which can be fairly expensive unless you have lots of highmem already.
So this would require that systems using the API have a highmem? (parisc doesn't today).
Yes, such solution will require highmem. It will introduce the highmem issues to systems that typically don't use highmem, that's why I searched for other solutions.
- Add logic to unmap pages from the linear mapping, which is
very expensive because it forces the use of small pages in the linear mapping (or in parts of it), and possibly means walking all page tables to remove the PTEs on alloc and put them back in on free.
I believe that Chunsang Jeong from Linaro is planning to implement both variants and post them for review, so we can decide which one to merge, or even to merge both and make it a configuration option. See also https://blueprints.launchpad.net/linaro-mm-sig/+spec/engr-mm-dma-mapping-201...
I don't think we need to make merging the CMA patches depending on the other patches, it's clear that both need to be solved, and they are independent enough.
I assume from the above that ARM has a hardware page walker?
Right.
The way I'd fix this on parisc, because we have a software based TLB, is to rely on the fact that a page may only be used either for DMA or for Page Cache, so the aliases should never be interleaved. Since you know the point at which the page flips from DMA to Cache (and vice versa), I'd purge the TLB entry and flush the page at that point and rely on the usage guarantees to ensure that the alias TLB entry doesn't reappear. This isn't inexpensive but the majority of the cost is the cache flush which is a requirement to clean the aliases anyway (a TLB entry purge is pretty cheap).
Would this work for the ARM hardware walker as well? It would require you to have a TLB entry purge instruction as well as some architectural guarantees about not speculating the TLB.
The main problem with ARM linear mapping is the fact that it is created using 2MiB sections, so entries for kernel linear mapping fits entirely in first lever of process page table. This implies that direct changing this linear mapping is not easy task and must be performed for all tasks in the system. In my CMA v12+ patches I decided to use simpler way of solving this issue. I rely on the fact that DMA memory is allocated only from CMA regions, so during early boot I change the kernel linear mappings for these regions. Instead of 2MiB sections, I use regular 4KiB pages which create 2 level of page tables. Second level of page table for these regions can be easily shared for all processes in the system.
This way I can easily update cache attributes for any single 4KiB page that is used for DMA and avoid any aliasing at all. The only drawback of this method is larger TLB pressure what might result in some slowdown during heavy IO if pages with 4KiB linear mapping are used. However with my hardware has only slow io (with eMMC I get only about 30MiB/s) so I cannot notice any impact of the mapping method on the io speed.
Best regards
On Wed, Aug 03, 2011 at 12:43:50PM -0500, James Bottomley wrote:
I assume from the above that ARM has a hardware page walker?
Correct, and speculative prefetch (which isn't prevented by not having TLB entries), so you can't keep entries out of the TLB. If it's in the page tables it can end up in the TLB.
The problem is that we could end up with conflicting attributes available to the hardware for the same physical page, and it is _completely_ undefined how hardware behaves with that (except that it does not halt - and there's no exception path for the condition because there's no detection of the problem case.)
So, if you had one mapping which was fully cacheable and another mapping which wasn't, you can flush the TLB all you like - it could be possible that you still up with an access through the non-cacheable mapping being cached (either hitting speculatively prefetched cache lines via the cacheable mapping, or the cacheable attributes being applied to the non-cacheable mapping - or conversely uncacheable attributes applied to the cacheable mapping.)
Essentially, the condition is labelled 'unpredictable' in the TRMs, which basically means that not even observed behaviour can be relied upon, because there may be cases where the observed behaviour fails.
Hello,
On Tuesday, July 05, 2011 1:34 PM Russell King - ARM Linux wrote:
On Tue, Jul 05, 2011 at 09:41:48AM +0200, Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
And how are you addressing the technical concerns about aliasing of cache attributes which I keep bringing up with this and you keep ignoring and telling me that I'm standing in your way.
I'm perfectly aware of the issues with aliasing of cache attributes.
My idea is to change low memory linear mapping for all CMA areas on boot time to use 2 level page tables (4KiB mappings instead of super-section mappings). This way the page properties for a single page in CMA area can be changed/updated at any time to match required coherent/writecombine attributes. Linear mapping can be even removed completely if we want to create the it elsewhere in the address space.
The only problem that might need to be resolved is GFP_ATOMIC allocation (updating page properties probably requires some locking), but it can be served from a special area which is created on boot without low-memory mapping at all. None sane driver will call dma_alloc_coherent(GFP_ATOMIC) for large buffers anyway.
CMA limits the memory area from which coherent pages are being taken quite well, so the change in the linear mapping method should have no significant impact on the system performance.
I didn't implement such solution yet, because it is really hard to handle all issues at the same time and creating the allocator was just a first step.
Best regards
On Wednesday 06 July 2011, Marek Szyprowski wrote:
The only problem that might need to be resolved is GFP_ATOMIC allocation (updating page properties probably requires some locking), but it can be served from a special area which is created on boot without low-memory mapping at all. None sane driver will call dma_alloc_coherent(GFP_ATOMIC) for large buffers anyway.
Would it be easier to start with a version that only allocated from memory without a low-memory mapping at first?
This would be similar to the approach that Russell's fix for the regular dma_alloc_coherent has taken, except that you need to also allow the memory to be used as highmem user pages.
Maybe you can simply adapt the default location of the contiguous memory are like this: - make CONFIG_CMA depend on CONFIG_HIGHMEM on ARM, at compile time - if ZONE_HIGHMEM exist during boot, put the CMA area in there - otherwise, put the CMA area at the top end of lowmem, and change the zone sizes so ZONE_HIGHMEM stretches over all of the CMA memory.
Arnd
On Wed, Jul 06, 2011 at 04:09:29PM +0200, Arnd Bergmann wrote:
Maybe you can simply adapt the default location of the contiguous memory are like this:
- make CONFIG_CMA depend on CONFIG_HIGHMEM on ARM, at compile time
- if ZONE_HIGHMEM exist during boot, put the CMA area in there
- otherwise, put the CMA area at the top end of lowmem, and change the zone sizes so ZONE_HIGHMEM stretches over all of the CMA memory.
One of the requirements of the allocator is that the returned memory should be zero'd (because it can be exposed to userspace via ALSA and frame buffers.)
Zeroing the memory from all the contexts which dma_alloc_coherent is called from is a trivial matter if its in lowmem, but highmem is harder.
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
Do we encounter this in practice i.e. do those platforms requiring large contiguous allocations motivating this work have such DMA restrictions?
Nicolas
On Wednesday 06 July 2011, Nicolas Pitre wrote:
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
Do we encounter this in practice i.e. do those platforms requiring large contiguous allocations motivating this work have such DMA restrictions?
You can probably find one or two of those, but we don't have to optimize for that case. I would at least expect the maximum size of the allocation to be smaller than the DMA limit for these, and consequently mandate that they define a sufficiently large CONSISTENT_DMA_SIZE for the crazy devices, or possibly add a hack to unmap some low memory and call dma_declare_coherent_memory() for the device.
Arnd
On Wed, 6 Jul 2011 at 16:59:45 Arnd Bergmann wrote:
On Wednesday 06 July 2011, Nicolas Pitre wrote:
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
Do we encounter this in practice i.e. do those platforms requiring large contiguous allocations motivating this work have such DMA restrictions?
You can probably find one or two of those, but we don't have to optimize for that case. I would at least expect the maximum size of the allocation to be smaller than the DMA limit for these, and consequently mandate that they define a sufficiently large CONSISTENT_DMA_SIZE for the crazy devices, or possibly add a hack to unmap some low memory and call dma_declare_coherent_memory() for the device.
Once found that Russell has dropped his "ARM: DMA: steal memory for DMA coherent mappings" for now, let me get back to this idea of a hack that would allow for safely calling dma_declare_coherent_memory() in order to assign a device with a block of contiguous memory for exclusive use. Assuming there should be no problem with successfully allocating a large continuous block of coherent memory at boot time with dma_alloc_coherent(), this block could be reserved for the device. The only problem is with the dma_declare_coherent_memory() calling ioremap(), which was designed with a device's dedicated physical memory in mind, but shouldn't be called on a memory already mapped.
There were three approaches proposed, two of them in August 2010: http://www.spinics.net/lists/linux-media/msg22179.html, http://www.spinics.net/lists/arm-kernel/msg96318.html, and a third one in January 2011: http://www.spinics.net/lists/linux-arch/msg12637.html.
As far as I can understand the reason why both of the first two were NAKed, it was suggested that videobuf-dma-contig shouldn't use coherent if all it requires is a contiguous memory, and a new API should be invented, or dma_pool API extended, for providing contiguous memory. The CMA was pointed out as a new work in progress contiguous memory API. Now it turns out it's not, it's only a helper to ensure that dma_alloc_coherent() always succeeds, and videobuf2-dma-contig is still going to allocate buffers from coherent memory.
(CCing both authors, Marin Mitov and Guennadi Liakhovetski, and their main opponent, FUJITA Tomonori)
The third solution was not discussed much after it was pointed out as being not very different from those two in terms of the above mentioned rationale.
All three solutions was different from now suggested method of unmapping some low memory and then calling dma_declare_coherent_memory() which ioremaps it in that those tried to reserve some boot time allocated coherent memory, already mapped correctly, without (io)remapping it.
If there are still problems with the CMA on one hand, and a need for a hack to handle "crazy devices" is still seen, regardless of CMA available and working or not, on the other, maybe we should get back to the idea of adopting coherent API to new requirements, review those three proposals again and select one which seems most acceptable to everyone? Being a submitter of the third, I'll be happy to refresh it if selected.
Thanks, Janusz
Hello,
On Saturday, July 09, 2011 4:57 PM Janusz Krzysztofik wrote:
On Wed, 6 Jul 2011 at 16:59:45 Arnd Bergmann wrote:
On Wednesday 06 July 2011, Nicolas Pitre wrote:
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
Do we encounter this in practice i.e. do those platforms requiring large contiguous allocations motivating this work have such DMA restrictions?
You can probably find one or two of those, but we don't have to optimize for that case. I would at least expect the maximum size of the allocation to be smaller than the DMA limit for these, and consequently mandate that they define a sufficiently large CONSISTENT_DMA_SIZE for the crazy devices, or possibly add a hack to unmap some low memory and call dma_declare_coherent_memory() for the device.
Once found that Russell has dropped his "ARM: DMA: steal memory for DMA coherent mappings" for now, let me get back to this idea of a hack that would allow for safely calling dma_declare_coherent_memory() in order to assign a device with a block of contiguous memory for exclusive use.
We tested such approach and finally with 3.0-rc1 it works fine. You can find an example for dma_declare_coherent() together with required memblock_remove() calls in the following patch series: http://www.spinics.net/lists/linux-samsung-soc/msg05026.html "[PATCH 0/3 v2] ARM: S5P: Add support for MFC device on S5PV210 and EXYNOS4"
Assuming there should be no problem with successfully allocating a large continuous block of coherent memory at boot time with dma_alloc_coherent(), this block could be reserved for the device. The only problem is with the dma_declare_coherent_memory() calling ioremap(), which was designed with a device's dedicated physical memory in mind, but shouldn't be called on a memory already mapped.
All these issues with ioremap has been finally resolved in 3.0-rc1. Like Russell pointed me in http://www.spinics.net/lists/arm-kernel/msg127644.html, ioremap can be fixed to work on early reserved memory areas by selecting ARCH_HAS_HOLES_MEMORYMODEL Kconfig option.
There were three approaches proposed, two of them in August 2010: http://www.spinics.net/lists/linux-media/msg22179.html, http://www.spinics.net/lists/arm-kernel/msg96318.html, and a third one in January 2011: http://www.spinics.net/lists/linux-arch/msg12637.html.
As far as I can understand the reason why both of the first two were NAKed, it was suggested that videobuf-dma-contig shouldn't use coherent if all it requires is a contiguous memory, and a new API should be invented, or dma_pool API extended, for providing contiguous memory.
This is another story. DMA-mapping framework definitely needs some extensions to allow more detailed specification of the allocated memory (currently we have only coherent and nearly ARM-specific writecombine). During Linaro Memory Management summit we agreed that the dma_alloc_attrs() function might be needed to clean-up the API and provide a nice way of adding new memory parameters. Having a possibility to allocate contiguous cached buffers might be one of the new DMA attributes. Here are some details of my proposal: http://www.spinics.net/lists/linux-mm/msg21235.html
The CMA was pointed out as a new work in progress contiguous memory API.
That was probably the biggest mistake at the beginning. We definitely should have learned dma-mapping framework and its internals.
Now it turns out it's not, it's only a helper to ensure that dma_alloc_coherent() always succeeds, and videobuf2-dma-contig is still going to allocate buffers from coherent memory.
I hope that once the dma_alloc_attrs() API will be accepted, I will add support for memory attributes to videobuf2-dma-contig allocator.
(CCing both authors, Marin Mitov and Guennadi Liakhovetski, and their main opponent, FUJITA Tomonori)
The third solution was not discussed much after it was pointed out as being not very different from those two in terms of the above mentioned rationale.
All three solutions was different from now suggested method of unmapping some low memory and then calling dma_declare_coherent_memory() which ioremaps it in that those tried to reserve some boot time allocated coherent memory, already mapped correctly, without (io)remapping it.
If there are still problems with the CMA on one hand, and a need for a hack to handle "crazy devices" is still seen, regardless of CMA available and working or not, on the other, maybe we should get back to the idea of adopting coherent API to new requirements, review those three proposals again and select one which seems most acceptable to everyone? Being a submitter of the third, I'll be happy to refresh it if selected.
I'm open to discussion.
Best regards
Dnia poniedziałek, 11 lipca 2011 o 15:47:32 Marek Szyprowski napisał(a):
Hello,
On Saturday, July 09, 2011 4:57 PM Janusz Krzysztofik wrote:
On Wed, 6 Jul 2011 at 16:59:45 Arnd Bergmann wrote:
On Wednesday 06 July 2011, Nicolas Pitre wrote:
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
Do we encounter this in practice i.e. do those platforms requiring large contiguous allocations motivating this work have such DMA restrictions?
You can probably find one or two of those, but we don't have to optimize for that case. I would at least expect the maximum size of the allocation to be smaller than the DMA limit for these, and consequently mandate that they define a sufficiently large CONSISTENT_DMA_SIZE for the crazy devices, or possibly add a hack to unmap some low memory and call dma_declare_coherent_memory() for the device.
Once found that Russell has dropped his "ARM: DMA: steal memory for DMA coherent mappings" for now, let me get back to this idea of a hack that would allow for safely calling dma_declare_coherent_memory() in order to assign a device with a block of contiguous memory for exclusive use.
We tested such approach and finally with 3.0-rc1 it works fine. You can find an example for dma_declare_coherent() together with required memblock_remove() calls in the following patch series: http://www.spinics.net/lists/linux-samsung-soc/msg05026.html "[PATCH 0/3 v2] ARM: S5P: Add support for MFC device on S5PV210 and EXYNOS4"
Assuming there should be no problem with successfully allocating a large continuous block of coherent memory at boot time with dma_alloc_coherent(), this block could be reserved for the device. The only problem is with the dma_declare_coherent_memory() calling ioremap(), which was designed with a device's dedicated physical memory in mind, but shouldn't be called on a memory already mapped.
All these issues with ioremap has been finally resolved in 3.0-rc1. Like Russell pointed me in http://www.spinics.net/lists/arm-kernel/msg127644.html, ioremap can be fixed to work on early reserved memory areas by selecting ARCH_HAS_HOLES_MEMORYMODEL Kconfig option.
I'm not sure. Recently I tried to refresh my now 7 months old patch in which I used that 'memblock_remove() then dma_declare_coherent_memery()' method[1]. It was different from your S5P MFC example in that it didn't punch any holes in the system memory, only stole a block of SDRAM from its tail. But Russell reminded me again: "we should not be mapping SDRAM using device mappings."[2]. Would defining ARCH_HAS_HOLES_MEMORYMODEL (even if it was justified) make any diference in my case? I don't think so. Wnat I think, after Russell, is that we still need that obligatory ioremap() removed from dma_declare_coherent_memory(), or made it optional, or a separate dma_declare_coherent_memory()-like function without (obligatory) ioremap() provided by the DMA API, in order to get the dma_declare_coherent_memery() method being accepted without any reservations when used inside arch/arm, I'm afraid.
Thanks, Janusz
[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2010-December/034644.h... [2] http://lists.infradead.org/pipermail/linux-arm-kernel/2011-June/052488.html
Hello,
On Monday, July 11, 2011 9:01 PM Janusz Krzysztofik wrote:
Dnia poniedziałek, 11 lipca 2011 o 15:47:32 Marek Szyprowski napisał(a):
Hello,
On Saturday, July 09, 2011 4:57 PM Janusz Krzysztofik wrote:
On Wed, 6 Jul 2011 at 16:59:45 Arnd Bergmann wrote:
On Wednesday 06 July 2011, Nicolas Pitre wrote:
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
Do we encounter this in practice i.e. do those platforms requiring large contiguous allocations motivating this work have such DMA restrictions?
You can probably find one or two of those, but we don't have to optimize for that case. I would at least expect the maximum size of the allocation to be smaller than the DMA limit for these, and consequently mandate that they define a sufficiently large CONSISTENT_DMA_SIZE for the crazy devices, or possibly add a hack to unmap some low memory and call dma_declare_coherent_memory() for the device.
Once found that Russell has dropped his "ARM: DMA: steal memory for DMA coherent mappings" for now, let me get back to this idea of a hack that would allow for safely calling dma_declare_coherent_memory() in order to assign a device with a block of contiguous memory for exclusive use.
We tested such approach and finally with 3.0-rc1 it works fine. You can find an example for dma_declare_coherent() together with required memblock_remove() calls in the following patch series: http://www.spinics.net/lists/linux-samsung-soc/msg05026.html "[PATCH 0/3 v2] ARM: S5P: Add support for MFC device on S5PV210 and EXYNOS4"
Assuming there should be no problem with successfully allocating a large continuous block of coherent memory at boot time with dma_alloc_coherent(), this block could be reserved for the device. The only problem is with the dma_declare_coherent_memory() calling ioremap(), which was designed with a device's dedicated physical memory in mind, but shouldn't be called on a memory already mapped.
All these issues with ioremap has been finally resolved in 3.0-rc1. Like Russell pointed me in http://www.spinics.net/lists/arm-kernel/msg127644.html, ioremap can be fixed to work on early reserved memory areas by selecting ARCH_HAS_HOLES_MEMORYMODEL Kconfig option.
I'm not sure. Recently I tried to refresh my now 7 months old patch in which I used that 'memblock_remove() then dma_declare_coherent_memery()' method[1]. It was different from your S5P MFC example in that it didn't punch any holes in the system memory, only stole a block of SDRAM from its tail. But Russell reminded me again: "we should not be mapping SDRAM using device mappings."[2]. Would defining ARCH_HAS_HOLES_MEMORYMODEL (even if it was justified) make any diference in my case? I don't think so.
Defining ARCH_HAS_HOLES_MEMORYMODEL changes the behavior of valid_pfn() macro/function, which is used in the ioremap(). When defined, valid_pfn() checks if the selected pfn is inside system memory or not (using memblock information). If the area is removed with memblock_remove(), then a check with valid_pfn() fails and ioremap() doesn't complain about mapping system memory.
Wnat I think, after Russell, is that we still need that obligatory ioremap() removed from dma_declare_coherent_memory(), or made it optional, or a separate dma_declare_coherent_memory()-like function without (obligatory) ioremap() provided by the DMA API, in order to get the dma_declare_coherent_memery() method being accepted without any reservations when used inside arch/arm, I'm afraid.
Please check again with 3.0-rc1. ARCH_HAS_HOLES_MEMORYMODEL solution was suggested by Russell. It looks like this is the correct solution for this problem, because I don't believe that ioremap() will be removed from dma_declare_coherent() anytime soon.
Best regards
On Wednesday 06 July 2011, Russell King - ARM Linux wrote:
On Wed, Jul 06, 2011 at 04:09:29PM +0200, Arnd Bergmann wrote:
Maybe you can simply adapt the default location of the contiguous memory are like this:
- make CONFIG_CMA depend on CONFIG_HIGHMEM on ARM, at compile time
- if ZONE_HIGHMEM exist during boot, put the CMA area in there
- otherwise, put the CMA area at the top end of lowmem, and change the zone sizes so ZONE_HIGHMEM stretches over all of the CMA memory.
One of the requirements of the allocator is that the returned memory should be zero'd (because it can be exposed to userspace via ALSA and frame buffers.)
Zeroing the memory from all the contexts which dma_alloc_coherent is called from is a trivial matter if its in lowmem, but highmem is harder.
I don't see how. The pages get allocated from an unmapped area or memory, mapped into the kernel address space as uncached or wc and then cleared. This should be the same for lowmem or highmem pages.
What am I missing?
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
True. The dmabounce code would consequently have to allocate the memory through an internal function that avoids the contiguous allocation area and goes straight to ZONE_DMA memory as it does today.
Arnd
On Wed, Jul 06, 2011 at 04:51:49PM +0200, Arnd Bergmann wrote:
On Wednesday 06 July 2011, Russell King - ARM Linux wrote:
On Wed, Jul 06, 2011 at 04:09:29PM +0200, Arnd Bergmann wrote:
Maybe you can simply adapt the default location of the contiguous memory are like this:
- make CONFIG_CMA depend on CONFIG_HIGHMEM on ARM, at compile time
- if ZONE_HIGHMEM exist during boot, put the CMA area in there
- otherwise, put the CMA area at the top end of lowmem, and change the zone sizes so ZONE_HIGHMEM stretches over all of the CMA memory.
One of the requirements of the allocator is that the returned memory should be zero'd (because it can be exposed to userspace via ALSA and frame buffers.)
Zeroing the memory from all the contexts which dma_alloc_coherent is called from is a trivial matter if its in lowmem, but highmem is harder.
I don't see how. The pages get allocated from an unmapped area or memory, mapped into the kernel address space as uncached or wc and then cleared. This should be the same for lowmem or highmem pages.
You don't want to clear them via their uncached or WC mapping, but via their cached mapping _before_ they get their alternative mapping, and flush any cached out of that mapping - both L1 and L2 caches.
For lowmem pages, that's easy. For highmem pages, they need to be individually kmap'd to zero them etc. (alloc_pages() warns on GFP_HIGHMEM + GFP_ZERO from atomic contexts - and dma_alloc_coherent must be callable from such contexts.)
That may be easier now that we don't have the explicit indicies for kmap_atomics, but at that time it wasn't easily possible.
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
True. The dmabounce code would consequently have to allocate the memory through an internal function that avoids the contiguous allocation area and goes straight to ZONE_DMA memory as it does today.
CMA's whole purpose for existing is to provide _dma-able_ contiguous memory for things like cameras and such like found on crippled non- scatter-gather hardware. If that memory is not DMA-able what's the point?
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
True. The dmabounce code would consequently have to allocate the memory through an internal function that avoids the contiguous allocation area and goes straight to ZONE_DMA memory as it does today.
CMA's whole purpose for existing is to provide _dma-able_ contiguous memory for things like cameras and such like found on crippled non- scatter-gather hardware. If that memory is not DMA-able what's the point?
ZONE_DMA is a zone for memory of legacy (crippled) devices that cannot DMA into all of memory (and so is ZONE_DMA32). Memory from ZONE_NORMAL can be used for DMA as well and a fully capable device would be expected to handle any memory in the system for DMA transfers.
"guaranteed" dmaable memory? DMA abilities are device specific. Well maybe you can call ZONE_DMA memory to be guaranteed if you guarantee that any device must at mininum be able to perform DMA into ZONE_DMA memory. But there may not be much of that memory around so you would want to limit the use of that scarce resource.
On Wed, 06 Jul 2011 18:05:00 +0200, Christoph Lameter cl@linux.com wrote:
ZONE_DMA is a zone for memory of legacy (crippled) devices that cannot DMA into all of memory (and so is ZONE_DMA32). Memory from ZONE_NORMAL can be used for DMA as well and a fully capable device would be expected to handle any memory in the system for DMA transfers.
"guaranteed" dmaable memory? DMA abilities are device specific. Well maybe you can call ZONE_DMA memory to be guaranteed if you guarantee that any device must at mininum be able to perform DMA into ZONE_DMA memory. But there may not be much of that memory around so you would want to limit the use of that scarce resource.
As pointed in Marek's other mail, this reasoning is not helping in any way. In case of video codec on various Samsung devices (and from some other threads this is not limited to Samsung), the codec needs separate buffers in separate memory banks.
On Wed, 6 Jul 2011, Michal Nazarewicz wrote:
On Wed, 06 Jul 2011 18:05:00 +0200, Christoph Lameter cl@linux.com wrote:
ZONE_DMA is a zone for memory of legacy (crippled) devices that cannot DMA into all of memory (and so is ZONE_DMA32). Memory from ZONE_NORMAL can be used for DMA as well and a fully capable device would be expected to handle any memory in the system for DMA transfers.
"guaranteed" dmaable memory? DMA abilities are device specific. Well maybe you can call ZONE_DMA memory to be guaranteed if you guarantee that any device must at mininum be able to perform DMA into ZONE_DMA memory. But there may not be much of that memory around so you would want to limit the use of that scarce resource.
As pointed in Marek's other mail, this reasoning is not helping in any way. In case of video codec on various Samsung devices (and from some other threads this is not limited to Samsung), the codec needs separate buffers in separate memory banks.
What I described is the basic memory architecture of Linux. I am not that familiar with ARM and the issue discussed here. Only got involved because ZONE_DMA was mentioned. The nature of ZONE_DMA is often misunderstood.
The allocation of the memory banks for the Samsung devices has to fit somehow into one of these zones. Its probably best to put the memory banks into ZONE_NORMAL and not have any dependency on ZONE_DMA at all.
On Wed, Jul 06, 2011 at 11:19:00AM -0500, Christoph Lameter wrote:
What I described is the basic memory architecture of Linux. I am not that familiar with ARM and the issue discussed here. Only got involved because ZONE_DMA was mentioned. The nature of ZONE_DMA is often misunderstood.
The allocation of the memory banks for the Samsung devices has to fit somehow into one of these zones. Its probably best to put the memory banks into ZONE_NORMAL and not have any dependency on ZONE_DMA at all.
Let me teach you about the ARM memory management on Linux.
Firstly, lets go over the structure of zones in Linux. There are three zones - ZONE_DMA, ZONE_NORMAL and ZONE_HIGHMEM. These zones are filled in that order. So, ZONE_DMA starts at zero. Following on from ZONE_DMA is ZONE_NORMAL memory, and lastly ZONE_HIGHMEM.
At boot, we pass all memory over to the kernel as follows:
1. If there is no DMA zone, then we pass all low memory over as ZONE_NORMAL.
2. If there is a DMA zone, by default we pass all low memory as ZONE_DMA. This is required so drivers which use GFP_DMA can work.
Platforms with restricted DMA requirements can modify that layout to move memory from ZONE_DMA into ZONE_NORMAL, thereby restricting the upper address which the kernel allocators will give for GFP_DMA allocations.
3. In either case, any high memory as ZONE_HIGHMEM if configured (or memory is truncated if not.)
So, when we have (eg) a platform where only the _even_ MBs of memory are DMA-able, we have a 1MB DMA zone at the beginning of system memory, and everything else in ZONE_NORMAL. This means GFP_DMA will return either memory from the first 1MB or fail if it can't. This is the behaviour we desire.
Normal allocations will come from ZONE_NORMAL _first_ and then try ZONE_DMA if there's no other alternative. This is the same desired behaviour as x86.
So, ARM is no different from x86, with the exception that the 16MB DMA zone due to ISA ends up being different sizes on ARM depending on our restrictions.
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
So, ARM is no different from x86, with the exception that the 16MB DMA zone due to ISA ends up being different sizes on ARM depending on our restrictions.
Sounds good. Thank you.
On Wed, Jul 06, 2011 at 11:05:00AM -0500, Christoph Lameter wrote:
On Wed, 6 Jul 2011, Russell King - ARM Linux wrote:
they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
True. The dmabounce code would consequently have to allocate the memory through an internal function that avoids the contiguous allocation area and goes straight to ZONE_DMA memory as it does today.
CMA's whole purpose for existing is to provide _dma-able_ contiguous memory for things like cameras and such like found on crippled non- scatter-gather hardware. If that memory is not DMA-able what's the point?
ZONE_DMA is a zone for memory of legacy (crippled) devices that cannot DMA into all of memory (and so is ZONE_DMA32). Memory from ZONE_NORMAL can be used for DMA as well and a fully capable device would be expected to handle any memory in the system for DMA transfers.
"guaranteed" dmaable memory? DMA abilities are device specific. Well maybe you can call ZONE_DMA memory to be guaranteed if you guarantee that any device must at mininum be able to perform DMA into ZONE_DMA memory. But there may not be much of that memory around so you would want to limit the use of that scarce resource.
Precisely, which is what ZONE_DMA is all about. I *have* been a Linux kernel hacker for the last 18 years and do know these things, especially as ARM has had various issues with DMA memory limitations over those years - and have successfully had platforms working reliably given that and ZONE_DMA.
On Wednesday 06 July 2011, Russell King - ARM Linux wrote:
On Wed, Jul 06, 2011 at 04:51:49PM +0200, Arnd Bergmann wrote:
On Wednesday 06 July 2011, Russell King - ARM Linux wrote:
I don't see how. The pages get allocated from an unmapped area or memory, mapped into the kernel address space as uncached or wc and then cleared. This should be the same for lowmem or highmem pages.
You don't want to clear them via their uncached or WC mapping, but via their cached mapping _before_ they get their alternative mapping, and flush any cached out of that mapping - both L1 and L2 caches.
But there can't be any other mapping, which is the whole point of the exercise to use highmem. Quoting from the new dma_alloc_area() function:
c = arm_vmregion_alloc(&area->vm, align, size, gfp & ~(__GFP_DMA | __GFP_HIGHMEM)); if (!c) return NULL; memset((void *)c->vm_start, 0, size);
area->vm here points to an uncached location, which means that we already zero the data through the uncached mapping. I don't see how it's getting worse than it is already.
Another issue is that when a platform has restricted DMA regions, they typically don't fall into the highmem zone. As the dmabounce code allocates from the DMA coherent allocator to provide it with guaranteed DMA-able memory, that would be rather inconvenient.
True. The dmabounce code would consequently have to allocate the memory through an internal function that avoids the contiguous allocation area and goes straight to ZONE_DMA memory as it does today.
CMA's whole purpose for existing is to provide _dma-able_ contiguous memory for things like cameras and such like found on crippled non- scatter-gather hardware. If that memory is not DMA-able what's the point?
I mean not any ZONE_DMA memory, but the memory backing coherent_areas[], which is by definition DMA-able from any device and is what is currently being used for the purpose.
Arnd
On Wed, 6 Jul 2011, Arnd Bergmann wrote:
On Wednesday 06 July 2011, Russell King - ARM Linux wrote:
On Wed, Jul 06, 2011 at 04:51:49PM +0200, Arnd Bergmann wrote:
On Wednesday 06 July 2011, Russell King - ARM Linux wrote:
I don't see how. The pages get allocated from an unmapped area or memory, mapped into the kernel address space as uncached or wc and then cleared. This should be the same for lowmem or highmem pages.
You don't want to clear them via their uncached or WC mapping, but via their cached mapping _before_ they get their alternative mapping, and flush any cached out of that mapping - both L1 and L2 caches.
But there can't be any other mapping, which is the whole point of the exercise to use highmem. Quoting from the new dma_alloc_area() function:
c = arm_vmregion_alloc(&area->vm, align, size, gfp & ~(__GFP_DMA | __GFP_HIGHMEM)); if (!c) return NULL; memset((void *)c->vm_start, 0, size);area->vm here points to an uncached location, which means that we already zero the data through the uncached mapping. I don't see how it's getting worse than it is already.
If you get a highmem page, because the cache is VIPT, that page might still be cached even if it wasn't mapped. With a VIVT cache we must flush the cache whenever a highmem page is unmapped. There is no such restriction with VIPT i.e. ARMv6 and above. Therefore to make sure the highmem page you get doesn't have cache lines associated to it, you must first map it cacheable, then perform cache invalidation on it, and eventually remap it as non-cacheable. This is necessary because there is no way to perform cache maintenance on L1 cache using physical addresses unfortunately. See commit 7e5a69e83b for an example of what this entails (fortunately commit 3e4d3af501 made things much easier and therefore commit 39af22a79 greatly simplified things).
Nicolas
On Wednesday 06 July 2011 21:10:07 Nicolas Pitre wrote:
If you get a highmem page, because the cache is VIPT, that page might still be cached even if it wasn't mapped. With a VIVT cache we must flush the cache whenever a highmem page is unmapped. There is no such restriction with VIPT i.e. ARMv6 and above. Therefore to make sure the highmem page you get doesn't have cache lines associated to it, you must first map it cacheable, then perform cache invalidation on it, and eventually remap it as non-cacheable. This is necessary because there is no way to perform cache maintenance on L1 cache using physical addresses unfortunately. See commit 7e5a69e83b for an example of what this entails (fortunately commit 3e4d3af501 made things much easier and therefore commit 39af22a79 greatly simplified things).
Ok, thanks for the explanation. This definitely makes the highmem approach much harder to get right, and slower. Let's hope then that Marek's approach of using small pages for the contiguous memory region and changing their attributes on the fly works out better than this.
Arnd
On Wed, 6 Jul 2011, Arnd Bergmann wrote:
On Wednesday 06 July 2011 21:10:07 Nicolas Pitre wrote:
If you get a highmem page, because the cache is VIPT, that page might still be cached even if it wasn't mapped. With a VIVT cache we must flush the cache whenever a highmem page is unmapped. There is no such restriction with VIPT i.e. ARMv6 and above. Therefore to make sure the highmem page you get doesn't have cache lines associated to it, you must first map it cacheable, then perform cache invalidation on it, and eventually remap it as non-cacheable. This is necessary because there is no way to perform cache maintenance on L1 cache using physical addresses unfortunately. See commit 7e5a69e83b for an example of what this entails (fortunately commit 3e4d3af501 made things much easier and therefore commit 39af22a79 greatly simplified things).
Ok, thanks for the explanation. This definitely makes the highmem approach much harder to get right, and slower. Let's hope then that Marek's approach of using small pages for the contiguous memory region and changing their attributes on the fly works out better than this.
I would say that both approaches have fairly equivalent complexity.
Nicolas
Hello,
On Wednesday, July 06, 2011 4:09 PM Arnd Bergmann wrote:
On Wednesday 06 July 2011, Marek Szyprowski wrote:
The only problem that might need to be resolved is GFP_ATOMIC allocation (updating page properties probably requires some locking), but it can be served from a special area which is created on boot without low-memory mapping at all. None sane driver will call dma_alloc_coherent(GFP_ATOMIC) for large buffers anyway.
Would it be easier to start with a version that only allocated from memory without a low-memory mapping at first?
This would be similar to the approach that Russell's fix for the regular dma_alloc_coherent has taken, except that you need to also allow the memory to be used as highmem user pages.
Maybe you can simply adapt the default location of the contiguous memory are like this:
- make CONFIG_CMA depend on CONFIG_HIGHMEM on ARM, at compile time
- if ZONE_HIGHMEM exist during boot, put the CMA area in there
- otherwise, put the CMA area at the top end of lowmem, and change the zone sizes so ZONE_HIGHMEM stretches over all of the CMA memory.
This will not solve our problems. We need CMA also to create at least one device private area that for sure will be in low memory (video codec).
I will rewrite ARM dma-mapping & CMA integration patch basing on the latest ARM for-next patches and add proof-of-concept of the solution presented in my previous mail (2-level page tables and unmapping pages from low-mem).
Best regards
On Wed, Jul 06, 2011 at 04:56:23PM +0200, Marek Szyprowski wrote:
This will not solve our problems. We need CMA also to create at least one device private area that for sure will be in low memory (video codec).
You make these statements but you don't say why. Can you please explain why the video codec needs low memory - does it have a restricted number of memory address bits which it can manipulate?
Hello,
On Wednesday, July 06, 2011 5:37 PM Russell King - ARM Linux wrote:
On Wed, Jul 06, 2011 at 04:56:23PM +0200, Marek Szyprowski wrote:
This will not solve our problems. We need CMA also to create at least one device private area that for sure will be in low memory (video codec).
You make these statements but you don't say why. Can you please explain why the video codec needs low memory - does it have a restricted number of memory address bits which it can manipulate?
Nope, it only needs to put some type of memory buffers in first bank (effectively in 30000000-34ffffff area) and the others in the second bank (40000000-57ffffff area). The values are given for Samsung GONI board.
Best regards
Hello,
I've just found two nasty bugs in this version of CMA. Sadly, both are the results of posting the patches in a big hurry. I'm really sorry.
Alignment argument was not passed correctly to the bitmap_find_next_zero_area() function and there was an ugly bug in the dma_release_from_contiguous() function.
On Tuesday, July 05, 2011 9:42 AM Marek Szyprowski wrote:
The Contiguous Memory Allocator is a set of helper functions for DMA mapping framework that improves allocations of contiguous memory chunks.
CMA grabs memory on system boot, marks it with CMA_MIGRATE_TYPE and gives back to the system. Kernel is allowed to allocate movable pages within CMA's managed memory so that it can be used for example for page cache when DMA mapping do not use it. On dma_alloc_from_contiguous() request such pages are migrated out of CMA area to free required contiguous block and fulfill the request. This allows to allocate large contiguous chunks of memory at any time assuming that there is enough free memory available in the system.
This code is heavily based on earlier works by Michal Nazarewicz.
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com CC: Michal Nazarewicz mina86@mina86.com
drivers/base/Kconfig | 77 +++++++++ drivers/base/Makefile | 1 + drivers/base/dma-contiguous.c | 367 ++++++++++++++++++++++++++++++++++++++++ include/linux/dma-contiguous.h | 104 +++++++++++ 4 files changed, 549 insertions(+), 0 deletions(-) create mode 100644 drivers/base/dma-contiguous.c create mode 100644 include/linux/dma-contiguous.h
diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig index d57e8d0..95ae1a7 100644 --- a/drivers/base/Kconfig +++ b/drivers/base/Kconfig @@ -168,4 +168,81 @@ config SYS_HYPERVISOR bool default n
+config CMA
- bool "Contiguous Memory Allocator"
- depends HAVE_DMA_CONTIGUOUS && HAVE_MEMBLOCK
- select MIGRATION
- select CMA_MIGRATE_TYPE
- help
This enables the Contiguous Memory Allocator which allows driversto allocate big physically-contiguous blocks of memory for use withhardware components that do not support I/O map nor scatter-gather.For more information see <include/linux/dma-contiguous.h>.If unsure, say "n".+if CMA
+config CMA_DEBUG
- bool "CMA debug messages (DEVELOPEMENT)"
- help
Turns on debug messages in CMA. This produces KERN_DEBUGmessages for every CMA call as well as various messages whileprocessing calls such as dma_alloc_from_contiguous().This option does not affect warning and error messages.+comment "Default contiguous memory area size:"
+config CMA_SIZE_ABSOLUTE
- int "Absolute size (in MiB)"
- default 16
- help
Defines the size (in MiB) of the default memory area for ContiguousMemory Allocator.+config CMA_SIZE_PERCENTAGE
- int "Percentage of total memory"
- default 10
- help
Defines the size of the default memory area for Contiguous MemoryAllocator as a percentage of the total memory in the system.+choice
- prompt "Selected region size"
- default CMA_SIZE_SEL_ABSOLUTE
+config CMA_SIZE_SEL_ABSOLUTE
- bool "Use absolute value only"
+config CMA_SIZE_SEL_PERCENTAGE
- bool "Use percentage value only"
+config CMA_SIZE_SEL_MIN
- bool "Use lower value (minimum)"
+config CMA_SIZE_SEL_MAX
- bool "Use higher value (maximum)"
+endchoice
+config CMA_ALIGNMENT
- int "Maximum PAGE_SIZE order of alignment for contiguous buffers"
- range 4 9
- default 8
- help
DMA mapping framework by default aligns all buffers to the smallestPAGE_SIZE order which is greater than or equal to the requestedbuffer
size. This works well for buffers up to a few hundreds kilobytes,but
for larger buffers it just a memory waste. With this parameter youcan
specify the maximum PAGE_SIZE order for contiguous buffers. Largerbuffers will be aligned only to this specified order. The order isexpressed as a power of two multiplied by the PAGE_SIZE.For example, if your system defaults to 4KiB pages, the order valueof 8 means that the buffers will be aligned up to 1MiB only.If unsure, leave the default value "8".+endif
endmenu diff --git a/drivers/base/Makefile b/drivers/base/Makefile index 4c5701c..be6aab4 100644 --- a/drivers/base/Makefile +++ b/drivers/base/Makefile @@ -5,6 +5,7 @@ obj-y := core.o sys.o bus.o dd.o syscore.o \ cpu.o firmware.o init.o map.o devres.o \ attribute_container.o transport_class.o obj-$(CONFIG_DEVTMPFS) += devtmpfs.o +obj-$(CONFIG_CMA) += dma-contiguous.o obj-y += power/ obj-$(CONFIG_HAS_DMA) += dma-mapping.o obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o diff --git a/drivers/base/dma-contiguous.c b/drivers/base/dma-contiguous.c new file mode 100644 index 0000000..707b901 --- /dev/null +++ b/drivers/base/dma-contiguous.c @@ -0,0 +1,367 @@ +/*
- Contiguous Memory Allocator for DMA mapping framework
- Copyright (c) 2010-2011 by Samsung Electronics.
- Written by:
- Marek Szyprowski m.szyprowski@samsung.com
- Michal Nazarewicz mina86@mina86.com
- This program is free software; you can redistribute it and/or
- modify it under the terms of the GNU General Public License as
- published by the Free Software Foundation; either version 2 of the
- License or (at your optional) any later version of the license.
- */
+#define pr_fmt(fmt) "cma: " fmt
+#ifdef CONFIG_CMA_DEBUG +#ifndef DEBUG +# define DEBUG +#endif +#endif
+#include <asm/page.h> +#include <asm/sizes.h>
+#include <linux/memblock.h> +#include <linux/err.h> +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/mutex.h> +#include <linux/page-isolation.h> +#include <linux/slab.h> +#include <linux/swap.h> +#include <linux/mm_types.h> +#include <linux/dma-mapping.h> +#include <linux/dma-contiguous.h>
+struct cma {
- unsigned long base_pfn;
- unsigned long count;
- unsigned long *bitmap;
+};
+struct cma *dma_contiguous_default_area;
+static unsigned long size_abs = CONFIG_CMA_SIZE_ABSOLUTE * SZ_1M; +static unsigned long size_percent = CONFIG_CMA_SIZE_PERCENTAGE; +static long size_cmdline = -1;
+static int __init early_cma(char *p) +{
- pr_debug("%s(%s)\n", __func__, p);
- size_cmdline = memparse(p, &p);
- return 0;
+} +early_param("cma", early_cma);
+/**
- dma_contiguous_reserve() - reserve area for contiguous memory handling
- This funtion reserves memory from memblock subsystem. It should be
- called by arch specific code once a memblock allocator has been
activated
- and all other subsystems have already allocated/reserved memory.
- */
+void __init dma_contiguous_reserve(void) +{
- struct memblock_region *reg;
- unsigned long selected_size = 0;
- unsigned long total_pages = 0;
- pr_debug("%s()\n", __func__);
- /*
* We cannot use memblock_phys_mem_size() here, because* memblock_analyze() has not been called yet.*/- for_each_memblock(memory, reg)
total_pages += memblock_region_memory_end_pfn(reg) -memblock_region_memory_base_pfn(reg);- size_percent *= (total_pages << PAGE_SHIFT) / 100;
- pr_debug("%s: available phys mem: %ld MiB\n", __func__,
(total_pages << PAGE_SHIFT) / SZ_1M);+#ifdef CONFIG_CMA_SIZE_SEL_ABSOLUTE
- selected_size = size_abs;
+#endif +#ifdef CONFIG_CMA_SIZE_SEL_PERCENTAGE
- selected_size = size_percent;
+#endif +#ifdef CONFIG_CMA_SIZE_SEL_MIN
- selected_size = min(size_abs, size_percent);
+#endif +#ifdef CONFIG_CMA_SIZE_SEL_MAX
- selected_size = max(size_abs, size_percent);
+#endif
- if (size_cmdline != -1)
selected_size = size_cmdline;- if (!selected_size)
return;- pr_debug("%s: reserving %ld MiB for global area\n", __func__,
selected_size / SZ_1M);- dma_declare_contiguous(NULL, selected_size, 0);
+};
+static DEFINE_MUTEX(cma_mutex);
+#ifdef CONFIG_DEBUG_VM
+static int __cma_activate_area(unsigned long base_pfn, unsigned long count) +{
- unsigned long pfn = base_pfn;
- unsigned i = count;
- struct zone *zone;
- pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count);
- VM_BUG_ON(!pfn_valid(pfn));
- zone = page_zone(pfn_to_page(pfn));
- do {
VM_BUG_ON(!pfn_valid(pfn));VM_BUG_ON(page_zone(pfn_to_page(pfn)) != zone);if (!(pfn & (pageblock_nr_pages - 1)))init_cma_reserved_pageblock(pfn_to_page(pfn));++pfn;- } while (--i);
- return 0;
+}
+#else
+static int __cma_activate_area(unsigned long base_pfn, unsigned long count) +{
- unsigned i = count >> pageblock_order;
- struct page *p = pfn_to_page(base_pfn);
- pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count);
- do {
init_cma_reserved_pageblock(p);p += pageblock_nr_pages;- } while (--i);
- return 0;
+}
+#endif
+static struct cma *__cma_create_area(unsigned long base_pfn,
unsigned long count)+{
- int bitmap_size = BITS_TO_LONGS(count) * sizeof(long);
- struct cma *cma;
- pr_debug("%s(0x%08lx+0x%lx)\n", __func__, base_pfn, count);
- cma = kmalloc(sizeof *cma, GFP_KERNEL);
- if (!cma)
return ERR_PTR(-ENOMEM);- cma->base_pfn = base_pfn;
- cma->count = count;
- cma->bitmap = kzalloc(bitmap_size, GFP_KERNEL);
- if (!cma->bitmap)
goto no_mem;- __cma_activate_area(base_pfn, count);
- pr_debug("%s: returning <%p>\n", __func__, (void *)cma);
- return cma;
+no_mem:
- kfree(cma);
- return ERR_PTR(-ENOMEM);
+}
+static struct cma_reserved {
- unsigned long start;
- unsigned long size;
- struct device *dev;
+} cma_reserved[8] __initdata; +static unsigned cma_reserved_count __initdata;
+static int __init __cma_init_reserved_areas(void) +{
- struct cma_reserved *r = cma_reserved;
- unsigned i = cma_reserved_count;
- pr_debug("%s()\n", __func__);
- for (; i; --i, ++r) {
struct cma *cma;cma = __cma_create_area(page_to_pfn(phys_to_page(r->start)),r->size >> PAGE_SHIFT);if (!IS_ERR(cma)) {pr_debug("%s: created area %p\n", __func__, cma);if (r->dev)set_dev_cma_area(r->dev, cma);elsedma_contiguous_default_area = cma;}- }
- return 0;
+} +core_initcall(__cma_init_reserved_areas);
+/**
- dma_declare_contiguous() - reserve area for contiguous memory handling
for particular device
- @dev: Pointer to device structure.
- @size: Size of the reserved memory.
- @start: Start address of the reserved memory (optional, 0 for any).
- This funtion reserves memory for specified device. It should be
- called by board specific code once a memblock allocator has been
activated
- and all other subsystems have already allocated/reserved memory.
- */
+int __init dma_declare_contiguous(struct device *dev, unsigned long size,
phys_addr_t start)+{
- struct cma_reserved *r = &cma_reserved[cma_reserved_count];
- unsigned long alignment;
- pr_debug("%s(%p+%p)\n", __func__, (void *)start, (void *)size);
- /* Sanity checks */
- if (cma_reserved_count == ARRAY_SIZE(cma_reserved))
return -ENOSPC;- if (!size)
return -EINVAL;- /* Sanitise input arguments */
- alignment = PAGE_SIZE << (MAX_ORDER + 1);
- start = ALIGN(start, alignment);
- size = ALIGN(size , alignment);
- /* Reserve memory */
- if (start) {
if (memblock_is_region_reserved(start, size) ||memblock_reserve(start, size) < 0)return -EBUSY;- } else {
/** Use __memblock_alloc_base() since* memblock_alloc_base() panic()s.*/u64 addr = __memblock_alloc_base(size, alignment, 0);if (!addr) {return -ENOMEM;} else if (addr + size > ~(unsigned long)0) {memblock_free(addr, size);return -EOVERFLOW;} else {start = addr;}- }
- /*
* Each reserved area must be initialised later, when more kernel* subsystems (like slab allocator) are available.*/- r->start = start;
- r->size = size;
- r->dev = dev;
- cma_reserved_count++;
- printk(KERN_INFO "%s: reserved %ld MiB area at 0x%p\n", __func__,
size / SZ_1M, (void *)start);- return 0;
+}
+/**
- dma_alloc_from_contiguous() - allocate pages from contiguous area
- @dev: Pointer to device for which the allocation is performed.
- @count: Requested number of pages.
- @align: Requested alignment of pages (in PAGE_SIZE order).
- This funtion allocates memory buffer for specified device. It uses
- device specific contiguous memory area if available or the default
- global one. Requires architecture specific get_dev_cma_area() helper
- function.
- */
+struct page *dma_alloc_from_contiguous(struct device *dev, int count,
unsigned int align)+{
- struct cma *cma = get_dev_cma_area(dev);
- unsigned long pfn, pageno;
- int ret;
- if (!cma)
return NULL;- if (align > CONFIG_CMA_ALIGNMENT)
align = CONFIG_CMA_ALIGNMENT;- pr_debug("%s(<%p>, %d/%d)\n", __func__, (void *)cma, count, align);
- if (!count)
return NULL;- mutex_lock(&cma_mutex);
- pageno = bitmap_find_next_zero_area(cma->bitmap, cma->count, 0, count,
align);
Fixed version: pageno = bitmap_find_next_zero_area(cma->bitmap, cma->count, 0, count, (1 << align) - 1);
- if (pageno >= cma->count) {
ret = -ENOMEM;goto error;- }
- bitmap_set(cma->bitmap, pageno, count);
- pfn = cma->base_pfn + pageno;
- ret = alloc_contig_range(pfn, pfn + count, 0, MIGRATE_CMA);
- if (ret)
goto free;- mutex_unlock(&cma_mutex);
- pr_debug("%s(): returning [%ld]\n", __func__, pfn);
- return pfn_to_page(pfn);
+free:
- bitmap_clear(cma->bitmap, pageno, count);
+error:
- mutex_unlock(&cma_mutex);
- return NULL;
+}
+/**
- dma_release_from_contiguous() - release allocated pages
- @dev: Pointer to device for which the pages were allocated.
- @pages: Allocated pages.
- @count: Number of allocated pages.
- This funtion reserves memory for specified device. It should be
- called by board specific code once a memblock allocator has been
activated
- and all other subsystems have already allocated/reserved memory.
- */
+int dma_release_from_contiguous(struct device *dev, struct page *pages,
int count)+{
- struct cma *cma = get_dev_cma_area(dev);
- unsigned long pfn;
- if (!cma || !pages)
return 0;- pr_debug("%s([%p])\n", __func__, (void *)pages);
- pfn = page_to_pfn(pages);
- if (pfn < cma->base_pfn || pfn >= cma->base_pfn + count)
Fixed version: if (pfn < cma->base_pfn || pfn >= cma->base_pfn + cma->count)
return 0;- mutex_lock(&cma_mutex);
- bitmap_clear(cma->bitmap, pfn - cma->base_pfn, count);
- free_contig_pages(pages, count);
- mutex_unlock(&cma_mutex);
- return 1;
+} diff --git a/include/linux/dma-contiguous.h b/include/linux/dma- contiguous.h new file mode 100644 index 0000000..98312c9 --- /dev/null +++ b/include/linux/dma-contiguous.h @@ -0,0 +1,104 @@ +#ifndef __LINUX_CMA_H +#define __LINUX_CMA_H
+/*
- Contiguous Memory Allocator for DMA mapping framework
- Copyright (c) 2010-2011 by Samsung Electronics.
- Written by:
- Marek Szyprowski m.szyprowski@samsung.com
- Michal Nazarewicz mina86@mina86.com
- This program is free software; you can redistribute it and/or
- modify it under the terms of the GNU General Public License as
- published by the Free Software Foundation; either version 2 of the
- License or (at your optional) any later version of the license.
- */
+/*
- Contiguous Memory Allocator
- The Contiguous Memory Allocator (CMA) makes it possible to
- allocate big contiguous chunks of memory after the system has
- booted.
- Why is it needed?
- Various devices on embedded systems have no scatter-getter and/or
- IO map support and require contiguous blocks of memory to
- operate. They include devices such as cameras, hardware video
- coders, etc.
- Such devices often require big memory buffers (a full HD frame
- is, for instance, more then 2 mega pixels large, i.e. more than 6
- MB of memory), which makes mechanisms such as kmalloc() or
- alloc_page() ineffective.
- At the same time, a solution where a big memory region is
- reserved for a device is suboptimal since often more memory is
- reserved then strictly required and, moreover, the memory is
- inaccessible to page system even if device drivers don't use it.
- CMA tries to solve this issue by operating on memory regions
- where only movable pages can be allocated from. This way, kernel
- can use the memory for pagecache and when device driver requests
- it, allocated pages can be migrated.
- Driver usage
- CMA should not be used by the device drivers directly. It is
- only a helper framework for dma-mapping subsystem.
- For more information, see kernel-docs in drivers/base/dma-
contiguous.c
- */
+#ifdef __KERNEL__
+struct cma; +struct page; +struct device;
+#ifdef CONFIG_CMA
+extern struct cma *dma_contiguous_default_area;
+void dma_contiguous_reserve(void); +int dma_declare_contiguous(struct device *dev, unsigned long size,
phys_addr_t base);+struct page *dma_alloc_from_contiguous(struct device *dev, int count,
unsigned int order);+int dma_release_from_contiguous(struct device *dev, struct page *pages,
int count);+#else
+#define dna_contiguous_default_area NULL
+static inline void dma_contiguous_reserve(void) { }
+static inline +int dma_declare_contiguous(struct device *dev, unsigned long size,
unsigned long base)+{
- return -EINVAL;
+}
+static inline +struct page *dma_alloc_from_contiguous(struct device *dev, int count,
unsigned int order)+{
- return NULL;
+}
+static inline +int dma_release_from_contiguous(struct device *dev, struct page *pages,
int count)+{
- return 0;
+}
+#endif
+#endif
+#endif
1.7.1.569.g6f426
Best regards
This patch adds support for CMA to dma-mapping subsystem for ARM architecture. By default a global CMA area is used, but specific devices are allowed to have their private memory areas if required (they can be created with dma_declare_contiguous() function during board initialization).
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com --- arch/arm/Kconfig | 1 + arch/arm/include/asm/device.h | 3 ++ arch/arm/include/asm/dma-mapping.h | 20 ++++++++++++++ arch/arm/mm/dma-mapping.c | 51 +++++++++++++++++++++++++++-------- arch/arm/mm/init.c | 3 ++ 5 files changed, 66 insertions(+), 12 deletions(-)
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 9adc278..3cca8cc 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -3,6 +3,7 @@ config ARM default y select HAVE_AOUT select HAVE_DMA_API_DEBUG + select HAVE_DMA_CONTIGUOUS select HAVE_IDE select HAVE_MEMBLOCK select RTC_LIB diff --git a/arch/arm/include/asm/device.h b/arch/arm/include/asm/device.h index 9f390ce..942913e 100644 --- a/arch/arm/include/asm/device.h +++ b/arch/arm/include/asm/device.h @@ -10,6 +10,9 @@ struct dev_archdata { #ifdef CONFIG_DMABOUNCE struct dmabounce_device_info *dmabounce; #endif +#ifdef CONFIG_CMA + struct cma *cma_area; +#endif };
struct pdev_archdata { diff --git a/arch/arm/include/asm/dma-mapping.h b/arch/arm/include/asm/dma-mapping.h index 4fff837..a3e1e48c 100644 --- a/arch/arm/include/asm/dma-mapping.h +++ b/arch/arm/include/asm/dma-mapping.h @@ -6,6 +6,7 @@ #include <linux/mm_types.h> #include <linux/scatterlist.h> #include <linux/dma-debug.h> +#include <linux/dma-contiguous.h>
#include <asm-generic/dma-coherent.h> #include <asm/memory.h> @@ -14,6 +15,25 @@ #error Please update to __arch_pfn_to_dma #endif
+#ifdef CONFIG_CMA +static inline struct cma *get_dev_cma_area(struct device *dev) +{ + if (dev->archdata.cma_area) + return dev->archdata.cma_area; + return dma_contiguous_default_area; +} + +static inline void set_dev_cma_area(struct device *dev, struct cma *cma) +{ + dev->archdata.cma_area = cma; +} +#else +static inline struct cma *get_dev_cma_area(struct device *dev) +{ + return NULL; +} +#endif + /* * dma_to_pfn/pfn_to_dma/dma_to_virt/virt_to_dma are architecture private * functions used internally by the DMA-mapping API to provide DMA diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index 82a093c..1d4e916 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -17,6 +17,7 @@ #include <linux/init.h> #include <linux/device.h> #include <linux/dma-mapping.h> +#include <linux/dma-contiguous.h> #include <linux/highmem.h>
#include <asm/memory.h> @@ -52,16 +53,35 @@ static u64 get_coherent_dma_mask(struct device *dev) return mask; }
+ +static struct page *__alloc_system_pages(size_t count, unsigned int order, gfp_t gfp) +{ + struct page *page, *p, *e; + + page = alloc_pages(gfp, order); + if (!page) + return NULL; + + /* + * Now split the huge page and free the excess pages + */ + split_page(page, order); + for (p = page + count, e = page + (1 << order); p < e; p++) + __free_page(p); + return page; +} + /* * Allocate a DMA buffer for 'dev' of size 'size' using the * specified gfp mask. Note that 'size' must be page aligned. */ static struct page *__dma_alloc_buffer(struct device *dev, size_t size, gfp_t gfp) { - unsigned long order = get_order(size); - struct page *page, *p, *e; + struct page *page; + size_t count = size >> PAGE_SHIFT; void *ptr; u64 mask = get_coherent_dma_mask(dev); + unsigned long order = get_order(count << PAGE_SHIFT);
#ifdef CONFIG_DMA_API_DEBUG u64 limit = (mask + 1) & ~mask; @@ -78,16 +98,19 @@ static struct page *__dma_alloc_buffer(struct device *dev, size_t size, gfp_t gf if (mask < 0xffffffffULL) gfp |= GFP_DMA;
- page = alloc_pages(gfp, order); - if (!page) - return NULL; + /* + * First, try to allocate memory from contiguous area + */ + page = dma_alloc_from_contiguous(dev, count, order);
/* - * Now split the huge page and free the excess pages + * Fallback if contiguous alloc fails or is not available */ - split_page(page, order); - for (p = page + (size >> PAGE_SHIFT), e = page + (1 << order); p < e; p++) - __free_page(p); + if (!page) + page = __alloc_system_pages(count, order, gfp); + + if (!page) + return NULL;
/* * Ensure that the allocated pages are zeroed, and that any data @@ -104,9 +127,13 @@ static struct page *__dma_alloc_buffer(struct device *dev, size_t size, gfp_t gf /* * Free a DMA buffer. 'size' must be page aligned. */ -static void __dma_free_buffer(struct page *page, size_t size) +static void __dma_free_buffer(struct device *dev, struct page *page, size_t size) { - struct page *e = page + (size >> PAGE_SHIFT); + size_t count = size >> PAGE_SHIFT; + struct page *e = page + count; + + if (dma_release_from_contiguous(dev, page, count)) + return;
while (page < e) { __free_page(page); @@ -416,7 +443,7 @@ void dma_free_coherent(struct device *dev, size_t size, void *cpu_addr, dma_addr if (!arch_is_coherent()) __dma_free_remap(cpu_addr, size);
- __dma_free_buffer(pfn_to_page(dma_to_pfn(dev, handle)), size); + __dma_free_buffer(dev, pfn_to_page(dma_to_pfn(dev, handle)), size); } EXPORT_SYMBOL(dma_free_coherent);
diff --git a/arch/arm/mm/init.c b/arch/arm/mm/init.c index c19571c..b2dfdeb 100644 --- a/arch/arm/mm/init.c +++ b/arch/arm/mm/init.c @@ -20,6 +20,7 @@ #include <linux/gfp.h> #include <linux/memblock.h> #include <linux/sort.h> +#include <linux/dma-contiguous.h>
#include <asm/mach-types.h> #include <asm/prom.h> @@ -358,6 +359,8 @@ void __init arm_memblock_init(struct meminfo *mi, struct machine_desc *mdesc) if (mdesc->reserve) mdesc->reserve();
+ dma_contiguous_reserve(); + memblock_analyze(); memblock_dump_all(); }
On Tuesday 05 July 2011, Marek Szyprowski wrote:
This patch adds support for CMA to dma-mapping subsystem for ARM architecture. By default a global CMA area is used, but specific devices are allowed to have their private memory areas if required (they can be created with dma_declare_contiguous() function during board initialization).
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com
Reviewed-by: Arnd Bergmann arnd@arndb.de
This patch is an example how device private CMA area can be activated. It creates one CMA region and assigns it to the first s5p-fimc device on Samsung Goni S5PC110 board.
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com --- arch/arm/mach-s5pv210/Kconfig | 1 + arch/arm/mach-s5pv210/mach-goni.c | 8 ++++++++ 2 files changed, 9 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mach-s5pv210/Kconfig b/arch/arm/mach-s5pv210/Kconfig index 37b5a97..c09a92c 100644 --- a/arch/arm/mach-s5pv210/Kconfig +++ b/arch/arm/mach-s5pv210/Kconfig @@ -64,6 +64,7 @@ menu "S5PC110 Machines" config MACH_AQUILA bool "Aquila" select CPU_S5PV210 + select CMA select S3C_DEV_FB select S5P_DEV_FIMC0 select S5P_DEV_FIMC1 diff --git a/arch/arm/mach-s5pv210/mach-goni.c b/arch/arm/mach-s5pv210/mach-goni.c index 31d5aa7..d9e565d 100644 --- a/arch/arm/mach-s5pv210/mach-goni.c +++ b/arch/arm/mach-s5pv210/mach-goni.c @@ -26,6 +26,7 @@ #include <linux/input.h> #include <linux/gpio.h> #include <linux/interrupt.h> +#include <linux/dma-contiguous.h>
#include <asm/mach/arch.h> #include <asm/mach/map.h> @@ -886,6 +887,12 @@ static void __init goni_machine_init(void) platform_add_devices(goni_devices, ARRAY_SIZE(goni_devices)); }
+static void __init goni_reserve(void) +{ + /* Create private 16MiB contiguous memory area for s5p-fimc.0 device */ + dma_declare_contiguous(&s5p_device_fimc0.dev, 16*SZ_1M, 0); +} + MACHINE_START(GONI, "GONI") /* Maintainers: Kyungmin Park kyungmin.park@samsung.com */ .boot_params = S5P_PA_SDRAM + 0x100, @@ -893,4 +900,5 @@ MACHINE_START(GONI, "GONI") .map_io = goni_map_io, .init_machine = goni_machine_init, .timer = &s5p_timer, + .reserve = goni_reserve, MACHINE_END
On Tuesday 05 July 2011, Marek Szyprowski wrote:
This patch is an example how device private CMA area can be activated. It creates one CMA region and assigns it to the first s5p-fimc device on Samsung Goni S5PC110 board.
Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com Signed-off-by: Kyungmin Park kyungmin.park@samsung.com
Reviewed-by: Arnd Bergmann arnd@arndb.de
On Tuesday 05 July 2011, Marek Szyprowski wrote:
This is yet another round of Contiguous Memory Allocator patches. I hope that I've managed to resolve all the items discussed during the Memory Management summit at Linaro Meeting in Budapest and pointed later on mailing lists. The goal is to integrate it as tight as possible with other kernel subsystems (like memory management and dma-mapping) and finally merge to mainline.
You have certainly addressed all of my concerns, this looks really good now!
Andrew, can you add this to your -mm tree? What's your opinion on the current state, do you think this is ready for merging in 3.1 or would you want to have more reviews from core memory management people?
My reviews were mostly on the driver and platform API side, and I think we're fine there now, but I don't really understand the impacts this has in mm.
Arnd
On Tue, Jul 05, 2011 at 02:07:17PM +0200, Arnd Bergmann wrote:
On Tuesday 05 July 2011, Marek Szyprowski wrote:
This is yet another round of Contiguous Memory Allocator patches. I hope that I've managed to resolve all the items discussed during the Memory Management summit at Linaro Meeting in Budapest and pointed later on mailing lists. The goal is to integrate it as tight as possible with other kernel subsystems (like memory management and dma-mapping) and finally merge to mainline.
You have certainly addressed all of my concerns, this looks really good now!
Andrew, can you add this to your -mm tree? What's your opinion on the current state, do you think this is ready for merging in 3.1 or would you want to have more reviews from core memory management people?
See my other mails. It is not ready for mainline.
On Tue, 5 Jul 2011 14:07:17 +0200 Arnd Bergmann arnd@arndb.de wrote:
On Tuesday 05 July 2011, Marek Szyprowski wrote:
This is yet another round of Contiguous Memory Allocator patches. I hope that I've managed to resolve all the items discussed during the Memory Management summit at Linaro Meeting in Budapest and pointed later on mailing lists. The goal is to integrate it as tight as possible with other kernel subsystems (like memory management and dma-mapping) and finally merge to mainline.
You have certainly addressed all of my concerns, this looks really good now!
Andrew, can you add this to your -mm tree? What's your opinion on the current state, do you think this is ready for merging in 3.1 or would you want to have more reviews from core memory management people?
My reviews were mostly on the driver and platform API side, and I think we're fine there now, but I don't really understand the impacts this has in mm.
I could review it and put it in there on a preliminary basis for some runtime testing. But the question in my mind is how different will the code be after the problems which rmk has identified have been fixed?
If "not very different" then that effort and testing will have been worthwhile.
If "very different" or "unworkable" then it was all for naught.
So. Do we have a feeling for the magnitude of the changes which will be needed to fix these things up?
On Thursday 07 July 2011 00:11:12 Andrew Morton wrote:
I could review it and put it in there on a preliminary basis for some runtime testing. But the question in my mind is how different will the code be after the problems which rmk has identified have been fixed?
If "not very different" then that effort and testing will have been worthwhile.
If "very different" or "unworkable" then it was all for naught.
So. Do we have a feeling for the magnitude of the changes which will be needed to fix these things up?
As far as I can tell, the changes that we still need are mostly in the ARM specific portion of the series. All architectures that have cache coherent DMA by default (most of the other interesting ones) can just call dma_alloc_from_contiguous() from their dma_alloc_coherent() function without having to do extra work.
It's possible that there will be small changes to simplify to the first six patches in order to simplify the ARM port, but I expect them to stay basically as they are, unless someone complains about them.
Arnd
Hello,
On Thursday, July 07, 2011 12:11 AM Andrew Morton wrote:
On Tue, 5 Jul 2011 14:07:17 +0200 Arnd Bergmann arnd@arndb.de wrote:
On Tuesday 05 July 2011, Marek Szyprowski wrote:
This is yet another round of Contiguous Memory Allocator patches. I
hope
that I've managed to resolve all the items discussed during the Memory Management summit at Linaro Meeting in Budapest and pointed later on mailing lists. The goal is to integrate it as tight as possible with other kernel subsystems (like memory management and dma-mapping) and finally merge to mainline.
You have certainly addressed all of my concerns, this looks really good
now!
Andrew, can you add this to your -mm tree? What's your opinion on the current state, do you think this is ready for merging in 3.1 or would you want to have more reviews from core memory management people?
My reviews were mostly on the driver and platform API side, and I think we're fine there now, but I don't really understand the impacts this has in mm.
I could review it and put it in there on a preliminary basis for some runtime testing. But the question in my mind is how different will the code be after the problems which rmk has identified have been fixed?
If "not very different" then that effort and testing will have been worthwhile.
The issue reported by Russell is very ARM specific and can be solved mostly in arch/arm/mm/dma-mapping.c, maybe with some minor changes/helpers in drivers/base/dma-contiguous.c The core part in linux/mm probably won't be affected by these changes at all.
Best regards
linaro-mm-sig@lists.linaro.org
-
Andrew Morton -
Arnd Bergmann -
Christoph Lameter -
James Bottomley -
Janusz Krzysztofik -
Marek Szyprowski -
Michal Nazarewicz -
Nicolas Pitre -
Russell King - ARM Linux