Re: [Qemu-devel] [RFC v2 05/11] tcg: protect TBContext with tb_lock.

[Sergey Fedorov]

On 11/05/16 15:52, Paolo Bonzini wrote:
> Just a couple answers/remarks.
>
> On 11/05/2016 14:45, Sergey Fedorov wrote:
(snip)
>>> @@ -1350,6 +1352,7 @@ void tb_invalidate_phys_page_range(tb_page_addr_t 
>>> start, tb_page_addr_t end,
>>>      /* we remove all the TBs in the range [start, end[ */
>>>      /* XXX: see if in some cases it could be faster to invalidate all
>>>         the code */
>>> +    tb_lock();
>> Don't we need also protect a call to page_find() above? page_find()
>> calls page_find_alloc() which is noted to be called with 'tb_lock' held.
> Only if alloc=1; page_find calls it with alloc=0.

Year, right :)

>> However, it might depend on the way we treat 'mmap_lock' in system mode
>> emulation.
> It's just not there; generally speaking it's replaced with tb_lock.

So why do we need yet another lock, 'mmap_lock', for user-mode emulation
and don't need it for system mode?

Kind regards,
Sergey