Re: [Qemu-devel] bitmap

[Juan Quintela <quintela@redhat.com>]

Stefan Hajnoczi <stefanha@gmail.com> wrote:
> On Mon, May 22, 2017 at 06:00:12PM +0430, ali saeedi wrote:
>> does this code 'atomic_rcu_read(&migration_bitmap_rcu)->bmap' return bitmap
>> of dirty blocks or bitmap of pages?
>
> This code changed recently.  Please refer to qemu.git/master when asking
> questions on qemu-devel.
>
> I think you're asking about what is now RAMBlock->bmap.  This bitmap
> describes pages (each page is TARGET_PAGE_SIZE bytes).
>
> This comment looks outdated, I don't think it returns a byte offset:
>
> /**
>  * migration_bitmap_find_dirty: find the next dirty page from start
>  *
>  * Called with rcu_read_lock() to protect migration_bitmap
>  *
>  * Returns the byte offset within memory region of the start of a dirty page
>  *
>  * @rs: current RAM state
>  * @rb: RAMBlock where to search for dirty pages
>  * @start: page where we start the search
>  */
> static inline
> unsigned long migration_bitmap_find_dirty(RAMState *rs, RAMBlock *rb,
>                                           unsigned long start)
>
> Evidence:
>
> static bool find_dirty_block(RAMState *rs, PageSearchStatus *pss, bool *again)
> {
>     pss->page = migration_bitmap_find_dirty(rs, pss->block, pss->page);
>     ...
>     if ((pss->page << TARGET_PAGE_BITS) >= pss->block->used_length) {
>
> Stefan

Oops, thanks you are right.

We used to have a global dirty bitmap.  As far as I know, it has always
worked on pages.  TARGET_PAGE_SIZE to be exact.  Since 2.10 open out, we
moved to a bitmap by RAMBlock.  It used to return an absolute page
number, now it returns a page number relative to the start of the block.

Later, Juan.