On 09/25/2015 12:37 PM, Wen Congyang wrote:
On 09/25/2015 05:31 PM, Denis V. Lunev wrote:
On 09/25/2015 12:24 PM, Wen Congyang wrote:
On 09/25/2015 05:09 PM, Denis V. Lunev wrote:
we can omit calling of bitmap_set in migration_bitmap_extend and
ram_save_setup just after bitmap_new, which properly zeroes memory
inside.
This patch is wrong. bitmap_set() is set all bits of the memory to 1,
not 0.
OK, then I'll replace g_try_malloc0 with g_try_malloc in the next patch to avoid
double memset
No, bitmap_new() is called in many places. Some caller needs zero memory, and
some caller doesn't need...
you are correct with your proposal with double allocation for the next
patch.
With a specific function like I have did in v2 g_try_malloc0 can be
replaced.
Anyway, I am fine with your proposal, this is not a (very) big deal.
EINPROGRESS
Den
