Re: [PATCH] mm/hugetlb: Warn the user when issues arise on boot due to hugepages

From: Michal Hocko <mhocko@suse.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: "Liam R. Howlett" <Liam.Howlett@Oracle.com>,
	linux-mm@kvack.org, mike.kravetz@Oracle.com,
	n-horiguchi@ah.jp.nec.com, aneesh.kumar@linux.vnet.ibm.com,
	gerald.schaefer@de.ibm.com, zhongjiang@huawei.com,
	aarcange@redhat.com, kirill.shutemov@linux.intel.com
Subject: Re: [PATCH] mm/hugetlb: Warn the user when issues arise on boot due to hugepages
Date: Sat, 17 Jun 2017 08:51:43 +0200	[thread overview]
Message-ID: <20170617065141.GB26698@dhcp22.suse.cz> (raw)
In-Reply-To: <20170616120755.c56d205f49d93a6e3dffb14f@linux-foundation.org>

On Fri 16-06-17 12:07:55, Andrew Morton wrote:
> On Mon, 12 Jun 2017 21:35:17 -0400 "Liam R. Howlett" <Liam.Howlett@Oracle.com> wrote:
> 
> > > 
> > > > If there's no message stating any
> > > > configuration issue, then many admins would probably think something is
> > > > seriously broken and it's not just a simple typo of K vs M.
> > > > 
> > > > Even though this doesn't catch all errors, I think it's a worth while
> > > > change since this is currently a silent failure which results in a
> > > > system crash.
> > > 
> > > Seriously, this warning just doesn't help in _most_ miscofigurations. It
> > > just focuses on one particular which really requires to misconfigure
> > > really badly. And there are way too many other ways to screw your system
> > > that way, yet we do not warn about many of those. So just try to step
> > > back and think whether this is something we actually do care about and
> > > if yes then try to come up with a more reasonable warning which would
> > > cover a wider range of misconfigurations.
> > 
> > Understood.  Again, I appreciate all the time you have taken on my
> > patch and explaining your points.  I will look at this again as you
> > have suggested.
> 
> So do we want to drop
> mm-hugetlb-warn-the-user-when-issues-arise-on-boot-due-to-hugepages.patch?
> 
> I'd be inclined to keep it if Liam found it a bit useful - it does have
> some overhead, but half the patch is in __init code...

I would rather see a more generic warning that would catch more
misconfiguration than those ultimately broken ones. If we find out that
such a warning is not feasible then I would not oppose to go with the
current approach but let's try a bit harder before we go that way.

Liam, are you willing to go that way?
-- 
Michal Hocko
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>