[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PATCH 16/16] block: Make cache=unsafe flush to the OS
From: |
Kevin Wolf |
Subject: |
[Qemu-devel] [PATCH 16/16] block: Make cache=unsafe flush to the OS |
Date: |
Fri, 11 Nov 2011 18:39:28 +0100 |
cache=unsafe completely ignored bdrv_flush, because flushing the host disk
costs a lot of performance. However, this means that qcow2 images (and
potentially any other format) can lose data even after the guest has issued a
flush if the qemu process crashes/is killed. In case of a host crash, data loss
is certainly expected with cache=unsafe, but if just the qemu process dies this
is a bit too unsafe.
Now that we have two separate flush functions, we can choose to flush
everythign to the OS, but don't enforce that it's physically written to the
disk.
Signed-off-by: Kevin Wolf <address@hidden>
---
block.c | 10 +++++++---
1 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/block.c b/block.c
index b1a4629..86910b0 100644
--- a/block.c
+++ b/block.c
@@ -2791,12 +2791,11 @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
{
int ret;
- if (bs->open_flags & BDRV_O_NO_FLUSH) {
- return 0;
- } else if (!bs->drv) {
+ if (!bs->drv) {
return 0;
}
+ /* Write back cached data to the OS even with cache=unsafe */
if (bs->drv->bdrv_co_flush_to_os) {
ret = bs->drv->bdrv_co_flush_to_os(bs);
if (ret < 0) {
@@ -2804,6 +2803,11 @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
}
}
+ /* But don't actually force it to the disk with cache=unsafe */
+ if (bs->open_flags & BDRV_O_NO_FLUSH) {
+ return 0;
+ }
+
if (bs->drv->bdrv_co_flush_to_disk) {
return bs->drv->bdrv_co_flush_to_disk(bs);
} else if (bs->drv->bdrv_aio_flush) {
--
1.7.6.4
- [Qemu-devel] [PATCH 02/16] block: add eject request callback, (continued)
- [Qemu-devel] [PATCH 02/16] block: add eject request callback, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 03/16] atapi: implement eject requests, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 04/16] scsi-disk: implement eject requests, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 05/16] nbd: treat EPIPE from NBD_DO_IT as success, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 07/16] qemu-nbd: rename socket variable, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 09/16] qemu-nbd: print error messages from the daemon through a pipe, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 08/16] qemu-nbd: move client to a thread, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 11/16] qemu-nbd: open the block device after starting the client thread, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 15/16] block: Introduce bdrv_co_flush_to_os, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 14/16] block: Rename bdrv_co_flush to bdrv_co_flush_to_disk, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 16/16] block: Make cache=unsafe flush to the OS,
Kevin Wolf <=
- [Qemu-devel] [PATCH 10/16] qemu-nbd: fix socket creation race, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 12/16] block: Fix vpc initialization of the Dynamic Disk Header, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 06/16] qemu-nbd: trap SIGTERM, Kevin Wolf, 2011/11/11
- [Qemu-devel] [PATCH 13/16] hw/pc.c: Fix use-while-uninitialized of fd_type[], Kevin Wolf, 2011/11/11
- Re: [Qemu-devel] [PULL 00/16] Block patches for 1.0, Anthony Liguori, 2011/11/13