[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Pine.LNX.4.64.0803111323130.18261@blonde.site>
Date: Tue, 11 Mar 2008 13:41:54 +0000 (GMT)
From: Hugh Dickins <hugh@...itas.com>
To: Akinobu Mita <akinobu.mita@...il.com>
cc: linux-kernel@...r.kernel.org, akpm@...ux-foundation.org,
linux-fsdevel@...r.kernel.org, William Irwin <wli@...omorphy.com>,
Matt Mackall <mpm@...enic.com>
Subject: Re: [PATCH 4/5] shmem: use call_once()
On Tue, 11 Mar 2008, Akinobu Mita wrote:
> 2008/3/11, Hugh Dickins <hugh@...itas.com>:
> > On Tue, 11 Mar 2008, Akinobu Mita wrote:
> > > This patch defers mounting tmpfs till shmem_file_setup() is
> > > called first time by using call_once().
> >
> > Please explain why we might need this patch: is something changing
> > elsewhere? Or are you misled by that "module_init(init_tmpfs)"
> > into thinking that mm/shmem.c is sometimes built modular?
>
> If no processes call shmem_file_setup() (via shm_get(2)), it is unnecessary
or shmem_zero_setup, not very common
> to do vfs_kern_mount(&tmpfs_fs_type, ...) unconditionary in boot-time.
> So I thought it is suitable example to demonstrate how to use "call_once()"
> in this patch set.
Oh, I see, thanks. Well, I don't feel all that strongly about it; but
on the whole I'd prefer we leave it as part of the __init, than change
it around to provide this example (and risk introducing some weird issue
e.g. related to its "dev"?). I guess the same should go for the huge
and the tiny, whereas you have better justification in the idr case.
Call me over-cautious.
Hugh
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists