Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 29 Feb 2016 08:44:49 -0800
From:      Bryan Drewery <bdrewery@FreeBSD.org>
To:        Slawa Olhovchenkov <slw@zxy.spb.ru>, Brooks Davis <brooks@freebsd.org>
Cc:        src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-9@freebsd.org
Subject:   Re: svn commit: r296146 - stable/9/sys/dev/filemon
Message-ID:  <56D47581.60802@FreeBSD.org>
In-Reply-To: <20160228123724.GC11654@zxy.spb.ru>
References:  <201602272057.u1RKvYfF035338@repo.freebsd.org> <20160228102728.GA94639@zxy.spb.ru> <20160228122008.GA27425@spindle.one-eyed-alien.net> <20160228123724.GC11654@zxy.spb.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
On 2/28/16 4:37 AM, Slawa Olhovchenkov wrote:
> On Sun, Feb 28, 2016 at 12:20:09PM +0000, Brooks Davis wrote:
>=20
>> On Sun, Feb 28, 2016 at 01:27:28PM +0300, Slawa Olhovchenkov wrote:
>>> On Sat, Feb 27, 2016 at 08:57:34PM +0000, Bryan Drewery wrote:
>>>
>>>> Author: bdrewery
>>>> Date: Sat Feb 27 20:57:34 2016
>>>> New Revision: 296146
>>>> URL: https://svnweb.freebsd.org/changeset/base/296146
>>>>
>>>> Log:
>>>>   MFC r242088:
>>>>  =20
>>>>     Desupport pre-FreeBSD 7.1.
>>>>
>>>> Modified:
>>>>   stable/9/sys/dev/filemon/filemon.c
>>>> Directory Properties:
>>>>   stable/9/sys/   (props changed)
>>>>   stable/9/sys/dev/   (props changed)
>>>> -#if __FreeBSD_version < 701000
>>>
>>> Not specialy to this commit, just for information: what purpose to
>>> have in src tree, already divided to branches, pieces of code under
>>> #if?
>>>
>>> Of couse, question not about external maintained code.
>>
>> Filemon started as externally maintained code and initially it was
>> a bit ambiguous which direction changes would flow.
>=20
> I.e. historical behaviour.
>=20
>> Sometimes it's also useful to allow code in HEAD to just work when
>> copied to stable so people can see if their problem is fixed without
>> having to merge too much.
>=20
> Please, check me: in mostly cases this is equivalent of MFC in one
> commit (and have same complexity as MFC) but w/o actualy MFC.

Yes. For example, in the MFC of some filemon changes to stable/10 and
stable/9 I had to make this change directly in the MFC that I was not
expecting and did not get audited well in SVN:

> diff --git a/sys/dev/filemon/filemon_wrapper.c b/sys/dev/filemon/filemo=
n_wrapper.c
> index a20ece5..d08aa6f 100644
> --- a/sys/dev/filemon/filemon_wrapper.c
> +++ b/sys/dev/filemon/filemon_wrapper.c
> @@ -424,7 +424,7 @@ filemon_event_process_exit(void *arg __unused, stru=
ct proc *p)
>=20
>         if ((filemon =3D filemon_pid_check(p)) !=3D NULL) {
>                 len =3D snprintf(filemon->msgbufr, sizeof(filemon->msgb=
ufr),
> -                   "X %d %d %d\n", p->p_pid, p->p_xexit, p->p_xsig);
> +                   "X %d %d\n", p->p_pid, W_EXITCODE(p->p_xstat, 0));
>=20
>                 filemon_output(filemon, filemon->msgbufr, len);

It is possible to add both versions in head such that the MFC is safe
and easy, but it pollutes head and I think it generally should not be don=
e.


>=20
>> In general though, we should remove code like
>> this in things maintained entirely in tree.
>=20
> Thanks!
>=20


--=20
Regards,
Bryan Drewery



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?56D47581.60802>