Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 22 Dec 2014 10:42:49 -0800
From:      Adrian Chadd <adrian@freebsd.org>
To:        Konstantin Belousov <kostikbel@gmail.com>
Cc:        "svn-src-head@freebsd.org" <svn-src-head@freebsd.org>, "svn-src-all@freebsd.org" <svn-src-all@freebsd.org>, "src-committers@freebsd.org" <src-committers@freebsd.org>, Andriy Gapon <avg@freebsd.org>, John Baldwin <jhb@freebsd.org>
Subject:   Re: svn commit: r276008 - in head/sys: kern sys
Message-ID:  <CAJ-Vmo=gNeCTKr9Em05O0Bs=8Oj5KeRAbsDRwCUcTgb8csJVEA@mail.gmail.com>
In-Reply-To: <20141222172810.GY2148@kib.kiev.ua>
References:  <201412211332.sBLDW8VC030548@svn.freebsd.org> <20141221151425.GX2148@kib.kiev.ua> <5496F37D.5030903@FreeBSD.org> <201412220940.02509.jhb@freebsd.org> <20141222172810.GY2148@kib.kiev.ua>

next in thread | previous in thread | raw e-mail | index | archive | help
On 22 December 2014 at 09:28, Konstantin Belousov <kostikbel@gmail.com> wrote:
> On Mon, Dec 22, 2014 at 09:40:02AM -0500, John Baldwin wrote:
>> On Sunday, December 21, 2014 11:21:17 am Andriy Gapon wrote:
>> > On 21/12/2014 17:14, Konstantin Belousov wrote:
>> > > On Sun, Dec 21, 2014 at 04:45:28PM +0200, Andriy Gapon wrote:
>> > >> On 21/12/2014 16:41, Konstantin Belousov wrote:
>> > >>> Or, are you asking why caching of the name could be needed for
>> > >>> core dump files at all ?
>> > >>
>> > >> Sort of. Why VN_OPEN_NAMECACHE is useful in the case of core dumps?
>> > >> What does it make better?
>> > > The vn_fullpath() mostly works for the core files after the change,
>> > > comparing with the non-working state at all before.
>> > >
>> >
>> > Ah, vn_fullpath(). Thank you.
>>
>> Is there something specific to core dumps that makes vn_fullpath() more
>> useful to have working before a process tries to open the core?  (As
>> compared to other newly-created files)
>
> See other Rui' reply in the thread.  It was done by his request.
>
> Basically, we cannot enable caching for CREATE, since operations like
> extracting large archive, would flush the cache.  Doing it rarely
> and when needed is acceptable.  The explained use case seems to
> be warranted.

.. 2c, sounds like having an ARC style replacement policy for
namecache entries would be useful.

(I mean the policy as described in the ARC paper, not the ZFS
implementation as a block cache.)

-adrian



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmo=gNeCTKr9Em05O0Bs=8Oj5KeRAbsDRwCUcTgb8csJVEA>