[PATCH v2 01/12] smb: smbdirect: add smbdirect_pdu.h with protocol definitions

[Namjae Jeon]

On Sun, Jun 1, 2025 at 12:49 PM Steve French <smfrench at gmail.com> wrote:
>
> Moving to use common headers is something I did multiple times with ksmbd and cifs.ko already, not a bad first step
This is not simply moved to the common header. He splits the header
into smaller pieces for some unknown reason.
Can you explain why he has split it into smbdirect_socket.h?
We doesn't need to do it now if he's thinking of creating a
smbdirect_socket.c file later.
>
> Thanks,
>
> Steve
>
> On Sat, May 31, 2025, 7:01 PM Namjae Jeon <linkinjeon at kernel.org> wrote:
>>
>> On Sun, Jun 1, 2025 at 8:23 AM Steve French <smfrench at gmail.com> wrote:
>> >
>> > I do like the small, relatively safe steps he is doing these in.
>> Small is okay, but I wonder when he will send the rest.
>> What if he just separates it like this and doesn't send the rest of
>> patches later?
>> I've never seen a case where the headers are separated first,
>> And send the main if it's implemented later. This is not a personal repository.
>>
>> Thanks.
>> >
>> > Thanks,
>> >
>> > Steve
>> >
>> > On Fri, May 30, 2025, 5:29 PM Namjae Jeon <linkinjeon at kernel.org> wrote:
>> >>
>> >> On Sat, May 31, 2025 at 4:03 AM Stefan Metzmacher <metze at samba.org> wrote:
>> >> >
>> >> > Am 29.05.25 um 01:28 schrieb Namjae Jeon:
>> >> > > On Thu, May 29, 2025 at 1:02 AM Stefan Metzmacher <metze at samba.org> wrote:
>> >> > >>
>> >> > >> This is just a start moving into a common smbdirect layer.
>> >> > >>
>> >> > >> It will be used in the next commits...
>> >> > >>
>> >> > >> Cc: Steve French <smfrench at gmail.com>
>> >> > >> Cc: Tom Talpey <tom at talpey.com>
>> >> > >> Cc: Long Li <longli at microsoft.com>
>> >> > >> Cc: Namjae Jeon <linkinjeon at kernel.org>
>> >> > >> Cc: Hyunchul Lee <hyc.lee at gmail.com>
>> >> > >> Cc: Meetakshi Setiya <meetakshisetiyaoss at gmail.com>
>> >> > >> Cc: linux-cifs at vger.kernel.org
>> >> > >> Cc: samba-technical at lists.samba.org
>> >> > >> Signed-off-by: Stefan Metzmacher <metze at samba.org>
>> >> > >> ---
>> >> > >>   fs/smb/common/smbdirect/smbdirect_pdu.h | 55 +++++++++++++++++++++++++
>> >> > >>   1 file changed, 55 insertions(+)
>> >> > >>   create mode 100644 fs/smb/common/smbdirect/smbdirect_pdu.h
>> >> > >>
>> >> > >> diff --git a/fs/smb/common/smbdirect/smbdirect_pdu.h b/fs/smb/common/smbdirect/smbdirect_pdu.h
>> >> > >> new file mode 100644
>> >> > >> index 000000000000..ae9fdb05ce23
>> >> > >> --- /dev/null
>> >> > >> +++ b/fs/smb/common/smbdirect/smbdirect_pdu.h
>> >> > >> @@ -0,0 +1,55 @@
>> >> > >> +/* SPDX-License-Identifier: GPL-2.0-or-later */
>> >> > >> +/*
>> >> > >> + *   Copyright (c) 2017 Stefan Metzmacher
>> >> > > Isn't it 2025? It looks like a typo.
>> >> >
>> >> > I took it from here:
>> >> > https://git.samba.org/?p=metze/linux/smbdirect.git;a=blob;f=smbdirect_private.h;hb=284ad8ea768c06e3cc70d6f2754929a6abbd2719
>> >> >
>> >> > > And why do you split the existing one into multiple header
>> >> > > files(smbdirect_pdu.h, smbdirect_socket.h, smbdirect.h)?
>> >> >
>> >> > In the end smbdirect.h will be the only header used outside
>> >> > of fs/smb/common/smbdirect, it will be the public api, to be used
>> >> > by the smb layer.
>> >> >
>> >> > smbdirect_pdu.h holds protocol definitions, while smbdirect_socket.h
>> >> > will be some kind of internal header that holds structures shared between multiple .c files.
>> >> >
>> >> > But we'll see I think this is a start in the correct direction.
>> >> When will you send the patches for multiple .c files?
>> >> I'm not sure if this is the right direction when I check only this patch-set.
>> >> I don't prefer to change the headers like this in advance without a body.
>> >> When you're ready, how about sending the patches including the body all at once?
>> >> >
>> >> > I try to focus on doing tiny steps avoiding doing to much at the same time
>> >> > or even try to avoid thinking about the next step already...
>> >> >
>> >> > metze