> Am 21.02.2017 um 13:42 schrieb Yann Ylavic <ylavic....@gmail.com>:
>
> On Mon, Feb 20, 2017 at 6:29 PM, Yann Ylavic <ylavic....@gmail.com> wrote:
>> On Mon, Feb 20, 2017 at 6:21 PM, Stefan Eissing
>> <stefan.eiss...@greenbytes.de> wrote:
>>>
>>> 1+2, so the only allocator is from master conn_rec.
>>
>> The allocator with a mutex (after r1783755)?
>
> Sorry to insist Stefan, I see a reason why it would fail without
> r1783755 (this commit) but not with, so maybe there is wolf in there
> (french expression, dunno if it translates literally to english :)
You have me confused now. If you did that all only for h2, then, I think, we
can live without them all nowadays, because:
1. master conn_rec->pool (own allocator, all ops in one thread)
* h2_session->pool
- h2_stream->pool
- h2_mplx->pool
2. h2_mplx->pool (own allocator, all ops guarded h2_mplx->lock mutex):
* h2_slave->pool
3. h2_slave->pool (own allocator, all ops in one thread)
* h2_task->pool
* r->pool
This is how it is intended to be used. And 2+3 run without allocator mutex at
Stefan Priebe's sites.
Stefan Eissing
<green/>bytes GmbH
Hafenstrasse 16
48155 Münster
www.greenbytes.de
