From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <vdavydov.dev@gmail.com>
Date: Mon, 6 May 2019 13:55:52 +0300
From: Vladimir Davydov <vdavydov.dev@gmail.com>
Subject: Re: [PATCH 3/3] test: slab_arena -- Verify madvise
Message-ID: <20190506105552.ckbg5baemwroablb@esperanza>
References: <20190501155006.14546-1-gorcunov@gmail.com>
 <20190501155006.14546-4-gorcunov@gmail.com>
 <20190506104520.it7uem5woxykg5xt@esperanza>
 <20190506104804.GO2488@uranus.lan>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20190506104804.GO2488@uranus.lan>
To: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: tml <tarantool-patches@freelists.org>, Alexander Turenko <alexander.turenko@tarantool.org>
List-ID: <tarantool-patches.dev.tarantool.org>

On Mon, May 06, 2019 at 01:48:04PM +0300, Cyrill Gorcunov wrote:
> On Mon, May 06, 2019 at 01:45:20PM +0300, Vladimir Davydov wrote:
> > [Cc += Alexander re tests]
> > 
> > On Wed, May 01, 2019 at 06:50:06PM +0300, Cyrill Gorcunov wrote:
> > > Since madvise support depends on sys libraries
> > > and kernel version we print error if only small
> > > supports it and /proc/self/smaps provide more
> > > less decent VmFlags.
> > 
> > TBO I wouldn't bother testing this feature at all, because it's way too
> > platform dependent.
> 
> Test does check for /proc/self/smaps file present and for VmFlags entry
> present, if neither of it found we just exit silently, without errors.

Okay, I'm fine with it.

To be prudent though, we'd have to check that the flag is set not only
for the preallocated memory region, but also for slabs allocated
afterwards (see slab_map -> mmap_checked). Not sure if it's worth it
though.