Re: [Qemu-devel] [PATCH 2/3] exec: adjuest rcu_read_lock requiement

[Fam Zheng]

In the subject line, s/adjuest/adjust/ and s/requiement/requirement/.

On Wed, 04/20 11:34, Gonglei wrote:
> qemu_ram_unset_idstr() don't need rcu lock anymore,

s/don't/doesn't/

> meanwhile adjuest the range of rcu lock in

s/adjuest/adjust/ or just s/adjuest/make/

> qemu_ram_set_idstr() as small as possible.
> 
> Signed-off-by: Gonglei <address@hidden>
> ---
>  exec.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/exec.c b/exec.c
> index 101f0f4..b0ecbcd 100644
> --- a/exec.c
> +++ b/exec.c
> @@ -1418,8 +1418,6 @@ void qemu_ram_set_idstr(RAMBlock *block, const char 
> *name, DeviceState *dev)
>      RAMBlock *new_block = block;
>      RAMBlock *old_block;
>  
> -    rcu_read_lock();
> -
>      assert(new_block);
>      assert(!new_block->idstr[0]);
>  
> @@ -1432,6 +1430,7 @@ void qemu_ram_set_idstr(RAMBlock *block, const char 
> *name, DeviceState *dev)
>      }
>      pstrcat(new_block->idstr, sizeof(new_block->idstr), name);
>  
> +    rcu_read_lock();
>      QLIST_FOREACH_RCU(old_block, &ram_list.blocks, next) {
>          if (old_block != new_block &&
>              !strcmp(old_block->idstr, new_block->idstr)) {
> @@ -1443,7 +1442,6 @@ void qemu_ram_set_idstr(RAMBlock *block, const char 
> *name, DeviceState *dev)
>      rcu_read_unlock();
>  }
>  
> -/* Called with iothread lock held.  */

This still seems true to me, why removing?

>  void qemu_ram_unset_idstr(RAMBlock *block)
>  {
>      /* FIXME: arch_init.c assumes that this is not called throughout
> @@ -1451,11 +1449,9 @@ void qemu_ram_unset_idstr(RAMBlock *block)
>       * does not work anyway.
>       */
>  
> -    rcu_read_lock();
>      if (block) {
>          memset(block->idstr, 0, sizeof(block->idstr));
>      }
> -    rcu_read_unlock();
>  }
>  
>  static int memory_try_enable_merging(void *addr, size_t len)
> -- 
> 1.7.12.4
> 
>