Best regards,
Vladimir
On 09.01.2015 01:28, Paolo Bonzini wrote:
On 11/12/2014 15:17, Vladimir Sementsov-Ogievskiy wrote:
-qemu_mutex_lock_iothread();
+bdrv_reset_dirty_bitmap(bs, bmds->dirty_bitmap, cur_sector, nr_sectors);
+
blk->aiocb = bdrv_aio_readv(bs, cur_sector, &blk->qi
On 09.01.2015 00:24, John Snow wrote:
On 12/11/2014 09:17 AM, Vladimir Sementsov-Ogievskiy wrote:
Instead of locking iothread, we can just swap these calls. So, if some
write to our range occures before resetting the bitmap, then it will
get into subsequent aio read, becouse it occures, in any
On 11/12/2014 15:17, Vladimir Sementsov-Ogievskiy wrote:
> -qemu_mutex_lock_iothread();
> +bdrv_reset_dirty_bitmap(bs, bmds->dirty_bitmap, cur_sector, nr_sectors);
> +
> blk->aiocb = bdrv_aio_readv(bs, cur_sector, &blk->qiov,
> nr_sectors, blk_mig_rea
On 12/11/2014 09:17 AM, Vladimir Sementsov-Ogievskiy wrote:
Instead of locking iothread, we can just swap these calls. So, if some
write to our range occures before resetting the bitmap, then it will
get into subsequent aio read, becouse it occures, in any case, after
resetting the bitmap.
s
Instead of locking iothread, we can just swap these calls. So, if some
write to our range occures before resetting the bitmap, then it will
get into subsequent aio read, becouse it occures, in any case, after
resetting the bitmap.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
block-migration.c
