[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PATCH 35/38] memory: syncronize kvm bitmap using bitmaps o
From: |
Juan Quintela |
Subject: |
[Qemu-devel] [PATCH 35/38] memory: syncronize kvm bitmap using bitmaps operations |
Date: |
Tue, 17 Dec 2013 16:26:11 +0100 |
If bitmaps are aligned properly, use bitmap operations. If they are
not, just use old bit at a time code.
Signed-off-by: Juan Quintela <address@hidden>
---
include/exec/ram_addr.h | 54 ++++++++++++++++++++++++++++++++-----------------
1 file changed, 36 insertions(+), 18 deletions(-)
diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h
index c6736ed..33c8acc 100644
--- a/include/exec/ram_addr.h
+++ b/include/exec/ram_addr.h
@@ -83,29 +83,47 @@ static inline void
cpu_physical_memory_set_dirty_lebitmap(unsigned long *bitmap,
ram_addr_t start,
ram_addr_t pages)
{
- unsigned int i, j;
+ unsigned long i, j;
unsigned long page_number, c;
hwaddr addr;
ram_addr_t ram_addr;
- unsigned int len = (pages + HOST_LONG_BITS - 1) / HOST_LONG_BITS;
+ unsigned long len = (pages + HOST_LONG_BITS - 1) / HOST_LONG_BITS;
unsigned long hpratio = getpagesize() / TARGET_PAGE_SIZE;
+ unsigned long page = BIT_WORD(start >> TARGET_PAGE_BITS);
- /*
- * bitmap-traveling is faster than memory-traveling (for addr...)
- * especially when most of the memory is not dirty.
- */
- for (i = 0; i < len; i++) {
- if (bitmap[i] != 0) {
- c = leul_to_cpu(bitmap[i]);
- do {
- j = ffsl(c) - 1;
- c &= ~(1ul << j);
- page_number = (i * HOST_LONG_BITS + j) * hpratio;
- addr = page_number * TARGET_PAGE_SIZE;
- ram_addr = start + addr;
- cpu_physical_memory_set_dirty_range(ram_addr,
- TARGET_PAGE_SIZE *
hpratio);
- } while (c != 0);
+ /* start address is aligned at the start of a word? */
+ if (((page * BITS_PER_LONG) << TARGET_PAGE_BITS) == start) {
+ long k;
+ long nr = BITS_TO_LONGS(pages);
+
+ for (k = 0; k < nr; k++) {
+ if (bitmap[k]) {
+ unsigned long temp = leul_to_cpu(bitmap[k]);
+
+ ram_list.dirty_memory[DIRTY_MEMORY_MIGRATION][page + k] |=
temp;
+ ram_list.dirty_memory[DIRTY_MEMORY_VGA][page + k] |= temp;
+ ram_list.dirty_memory[DIRTY_MEMORY_CODE][page + k] |= temp;
+ }
+ }
+ xen_modified_memory(start, pages);
+ } else {
+ /*
+ * bitmap-traveling is faster than memory-traveling (for addr...)
+ * especially when most of the memory is not dirty.
+ */
+ for (i = 0; i < len; i++) {
+ if (bitmap[i] != 0) {
+ c = leul_to_cpu(bitmap[i]);
+ do {
+ j = ffsl(c) - 1;
+ c &= ~(1ul << j);
+ page_number = (i * HOST_LONG_BITS + j) * hpratio;
+ addr = page_number * TARGET_PAGE_SIZE;
+ ram_addr = start + addr;
+ cpu_physical_memory_set_dirty_range(ram_addr,
+ TARGET_PAGE_SIZE * hpratio);
+ } while (c != 0);
+ }
}
}
}
--
1.8.3.1
- Re: [Qemu-devel] [PATCH 30/38] memory: split cpu_physical_memory_* functions to its own include, (continued)
- [Qemu-devel] [PATCH 31/38] memory: unfold memory_region_test_and_clear(), Juan Quintela, 2013/12/17
- [Qemu-devel] [PATCH 32/38] kvm: use directly cpu_physical_memory_* api for tracking dirty pages, Juan Quintela, 2013/12/17
- [Qemu-devel] [PATCH 33/38] kvm: refactor start address calculation, Juan Quintela, 2013/12/17
- [Qemu-devel] [PATCH 34/38] memory: move bitmap synchronization to its own function, Juan Quintela, 2013/12/17
- [Qemu-devel] [PATCH 35/38] memory: syncronize kvm bitmap using bitmaps operations,
Juan Quintela <=
- [Qemu-devel] [PATCH 36/38] ram: split function that synchronizes a range, Juan Quintela, 2013/12/17
- [Qemu-devel] [PATCH 37/38] migration: synchronize memory bitmap 64bits at a time, Juan Quintela, 2013/12/17
- [Qemu-devel] [PATCH 38/38] ram: align ram_addr_t's regions in multiples of 64, Juan Quintela, 2013/12/17