[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <12a593bc-68bc-ac03-0307-a65a0c064af3@suse.de>
Date: Mon, 18 Oct 2021 22:42:37 +0800
From: Coly Li <colyli@...e.de>
To: Arnd Bergmann <arnd@...nel.org>
Cc: Arnd Bergmann <arnd@...db.de>, Jens Axboe <axboe@...nel.dk>,
Kent Overstreet <kmo@...erainc.com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Christoph Hellwig <hch@....de>
Subject: Re: [PATCH] [v2] bcache: hide variable-sized types from uapi header
check
On 10/18/21 10:39 PM, Arnd Bergmann wrote:
> On Mon, Oct 18, 2021 at 4:20 PM Coly Li <colyli@...e.de> wrote:
>> IMHO, remove bcache related header from uapi check might be better
>> solution. So far only bcache-tools uses this header with its own copy,
>> no application includes the header(s) so far. It makes sense to exclude
>> bcache.h from upai headers check.
> Should we just move it to include/linux/ and out of the uapi headers entirely
> then? It sounds like it's not actually an ABI but just the definition of the
> data layout that is not included by anything from user space.
>
> We are a bit inconsistent here already, e.g. btrfs has all its structures
> in uapi, but ext4 does not.
I am quite open for this idea. It is in uapi directory before I maintain
bcache. I just though the header fines on-media format should go into
include/uapi/, but if this is not the restricted rule, it is fine for me
to move this header to drivers/md/bcache/.
Coly Li
Powered by blists - more mailing lists