|
From: | zhanghailiang |
Subject: | Re: [Qemu-devel] [PATCH COLO-Frame v10 15/38] ram/COLO: Record pages received from PVM by re-using migration dirty bitmap |
Date: | Mon, 16 Nov 2015 17:07:48 +0800 |
User-agent: | Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0) Gecko/20100101 Thunderbird/38.1.0 |
On 2015/11/14 0:19, Dr. David Alan Gilbert wrote:
* zhanghailiang (address@hidden) wrote:We need to record the address of the dirty pages that received from PVM, It will help flushing pages that cached into SVM. Signed-off-by: zhanghailiang <address@hidden> --- v10: - New patch split from v9's patch 13 - Rebase to master to use 'migration_bitmap_rcu' --- migration/ram.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) diff --git a/migration/ram.c b/migration/ram.c index b094dc3..70879bd 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -1448,6 +1448,18 @@ static inline void *host_from_stream_offset(QEMUFile *f, } if (ram_cache_enable) { + unsigned long *bitmap; + long k = (block->mr->ram_addr + offset) >> TARGET_PAGE_BITS; + + bitmap = atomic_rcu_read(&migration_bitmap_rcu)->bmap; + /* + * During colo checkpoint, we need bitmap of these migrated pages. + * It help us to decide which pages in ram cache should be flushed + * into VM's RAM later. + */ + if (!test_and_set_bit(k, bitmap)) { + migration_dirty_pages++; + }I don't like having this in host_from_stream_offset; if you look at the current ram_load there is only a single call to host_from_stream_offset, so it's now much easier for you to move it into a separate function.
Hmm, that's really a good idea, i will split it in next version.
return block->host_cache + offset; } else { return block->host + offset; @@ -1462,6 +1474,13 @@ static inline void *host_from_stream_offset(QEMUFile *f, if (!strncmp(id, block->idstr, sizeof(id)) && block->max_length > offset) { if (ram_cache_enable) { + unsigned long *bitmap; + long k = (block->mr->ram_addr + offset) >> TARGET_PAGE_BITS; + + bitmap = atomic_rcu_read(&migration_bitmap_rcu)->bmap; + if (!test_and_set_bit(k, bitmap)) { + migration_dirty_pages++; + } return block->host_cache + offset; } else { return block->host + offset; @@ -1723,6 +1742,7 @@ static int ram_load(QEMUFile *f, void *opaque, int version_id) int colo_init_ram_cache(void) { RAMBlock *block; + int64_t ram_cache_pages = last_ram_offset() >> TARGET_PAGE_BITS; rcu_read_lock(); QLIST_FOREACH_RCU(block, &ram_list.blocks, next) { @@ -1734,6 +1754,15 @@ int colo_init_ram_cache(void) } rcu_read_unlock(); ram_cache_enable = true; + /* + * Record the dirty pages that sent by PVM, we use this dirty bitmap together + * with to decide which page in cache should be flushed into SVM's RAM. Here + * we use the same name 'migration_bitmap_rcu' as for migration. + */ + migration_bitmap_rcu = g_new(struct BitmapRcu, 1);Please update that to g_new0 (I changed the other use when I added postcopy).
OK, thanks.
Dave+ migration_bitmap_rcu->bmap = bitmap_new(ram_cache_pages); + migration_dirty_pages = 0; + return 0; out_locked: @@ -1751,9 +1780,15 @@ out_locked: void colo_release_ram_cache(void) { RAMBlock *block; + struct BitmapRcu *bitmap = migration_bitmap_rcu; ram_cache_enable = false; + atomic_rcu_set(&migration_bitmap_rcu, NULL); + if (bitmap) { + call_rcu(bitmap, migration_bitmap_free, rcu); + } + rcu_read_lock(); QLIST_FOREACH_RCU(block, &ram_list.blocks, next) { if (block->host_cache) { -- 1.8.3.1-- Dr. David Alan Gilbert / address@hidden / Manchester, UK .
[Prev in Thread] | Current Thread | [Next in Thread] |