[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZHBqGosY0tWkNdIR@infradead.org>
Date: Fri, 26 May 2023 01:13:14 -0700
From: Christoph Hellwig <hch@...radead.org>
To: Luis Chamberlain <mcgrof@...nel.org>
Cc: axboe@...nel.dk, agk@...hat.com, snitzer@...nel.org,
philipp.reisner@...bit.com, lars.ellenberg@...bit.com,
christoph.boehmwalder@...bit.com, hch@...radead.org,
djwong@...nel.org, minchan@...nel.org, senozhatsky@...omium.org,
patches@...ts.linux.dev, linux-block@...r.kernel.org,
linux-mm@...ck.org, linux-xfs@...r.kernel.org,
linux-fsdevel@...r.kernel.org, dm-devel@...hat.com,
drbd-dev@...ts.linbit.com, linux-kernel@...r.kernel.org,
willy@...radead.org, hare@...e.de, p.raghav@...sung.com,
da.gomez@...sung.com, rohan.puri@...sung.com,
rpuri.linux@...il.com, kbusch@...nel.org
Subject: Re: [PATCH v2 1/5] block: annotate bdev_disk_changed() deprecation
with a symbol namespace
On Fri, May 26, 2023 at 12:33:32AM -0700, Luis Chamberlain wrote:
> This ensures no other users pop up by mistake easily and provides
> us a with an easy vehicle to do the same with other routines should
> we need it later.
I don't see how this is related to the rest of the seris. I also don't
think it's a good idea. The APIs isn't deprecated per se. It just
should not be called by drivers. The right thing would be an interface
like
EXPORT_SYMBOL_GPL_FOR(bdev_disk_changed, loop.ko, CONFIG_BLK_DEV_LOOP);
EXPORT_SYMBOL_GPL_FOR(bdev_disk_changed, dasd_mod.ko, CONFIG_DASD);
with the modulo code enforcing that no one but the module this is
explicitly exorted for can use the symbol.
Powered by blists - more mailing lists