Date: Mon, 7 Feb 2022 11:04:37 -0700 From: Warner Losh <imp@bsdimp.com> To: Konstantin Belousov <kostikbel@gmail.com> Cc: Sebastian Huber <sebastian.huber@embedded-brains.de>, FreeBSD Hackers <freebsd-hackers@freebsd.org> Subject: Re: ntp_init() looks like a nop Message-ID: <CANCZdfoOVekuyZkkvdgzRQPYhgyOWjaqLagD0HNMFPquQAt1hw@mail.gmail.com> In-Reply-To: <YgE4RaZVZhOlhJHW@kib.kiev.ua> References: <ac806cf6-a17f-77ba-5633-0c95eda1f8cf@embedded-brains.de> <YgE4RaZVZhOlhJHW@kib.kiev.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
--000000000000b0ea4d05d7716fe2
Content-Type: text/plain; charset="UTF-8"
On Mon, Feb 7, 2022 at 8:20 AM Konstantin Belousov <kostikbel@gmail.com>
wrote:
> On Mon, Feb 07, 2022 at 04:04:56PM +0100, Sebastian Huber wrote:
> > Hello,
> >
> > I review currently the kern_ntptime.c module since I would like to use
> it in
> > RTEMS. I have a question to ntp_init():
> >
> > /*
> > * ntp_init() - initialize variables and structures
> > *
> > * This routine must be called after the kernel variables hz and tick
> > * are set or changed and before the next tick interrupt. In this
> > * particular implementation, these values are assumed set elsewhere in
> > * the kernel. The design allows the clock frequency and tick interval
> > * to be changed while the system is running. So, this routine should
> > * probably be integrated with the code that does that.
> > */
> > static void
> > ntp_init(void)
> > {
> >
> > /*
> > * The following variables are initialized only at startup. Only
> > * those structures not cleared by the compiler need to be
> > * initialized, and these only in the simulator. In the actual
> > * kernel, any nonzero values here will quickly evaporate.
> > */
> > L_CLR(time_offset);
> > L_CLR(time_freq);
> > #ifdef PPS_SYNC
> > pps_tf[0].tv_sec = pps_tf[0].tv_nsec = 0;
> > pps_tf[1].tv_sec = pps_tf[1].tv_nsec = 0;
> > pps_tf[2].tv_sec = pps_tf[2].tv_nsec = 0;
> > pps_fcount = 0;
> > L_CLR(pps_freq);
> > #endif /* PPS_SYNC */
> > }
> >
> > SYSINIT(ntpclocks, SI_SUB_CLOCKS, SI_ORDER_MIDDLE, ntp_init, NULL);
> >
> > The ntp_init() function sets a couple of global variables to zero. These
> > variables should be in the .bss section. Are they not already cleared
> during
> > the kernel loading?
> This is nop indeed, be it linked in kernel, or loaded as a module.
>
I just looked at the posted patch, and I concur.
Warner
--000000000000b0ea4d05d7716fe2
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div dir=3D"ltr"><div dir=3D"ltr"><br></div><br><div class=3D"gmail_quote">=
<div dir=3D"ltr" class=3D"gmail_attr">On Mon, Feb 7, 2022 at 8:20 AM Konsta=
ntin Belousov <<a href=3D"mailto:kostikbel@gmail.com">kostikbel@gmail.co=
m</a>> wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin=
:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"=
>On Mon, Feb 07, 2022 at 04:04:56PM +0100, Sebastian Huber wrote:<br>
> Hello,<br>
> <br>
> I review currently the kern_ntptime.c module since I would like to use=
it in<br>
> RTEMS. I have a question to ntp_init():<br>
> <br>
> /*<br>
>=C2=A0 * ntp_init() - initialize variables and structures<br>
>=C2=A0 *<br>
>=C2=A0 * This routine must be called after the kernel variables hz and =
tick<br>
>=C2=A0 * are set or changed and before the next tick interrupt. In this=
<br>
>=C2=A0 * particular implementation, these values are assumed set elsewh=
ere in<br>
>=C2=A0 * the kernel. The design allows the clock frequency and tick int=
erval<br>
>=C2=A0 * to be changed while the system is running. So, this routine sh=
ould<br>
>=C2=A0 * probably be integrated with the code that does that.<br>
>=C2=A0 */<br>
> static void<br>
> ntp_init(void)<br>
> {<br>
> <br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0/*<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 * The following variables are initialized o=
nly at startup. Only<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 * those structures not cleared by the compi=
ler need to be<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 * initialized, and these only in the simula=
tor. In the actual<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 * kernel, any nonzero values here will quic=
kly evaporate.<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0 */<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0L_CLR(time_offset);<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0L_CLR(time_freq);<br>
> #ifdef PPS_SYNC<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0pps_tf[0].tv_sec =3D pps_tf[0].tv_nsec =3D 0=
;<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0pps_tf[1].tv_sec =3D pps_tf[1].tv_nsec =3D 0=
;<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0pps_tf[2].tv_sec =3D pps_tf[2].tv_nsec =3D 0=
;<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0pps_fcount =3D 0;<br>
>=C2=A0 =C2=A0 =C2=A0 =C2=A0L_CLR(pps_freq);<br>
> #endif /* PPS_SYNC */ <br>
> }<br>
> <br>
> SYSINIT(ntpclocks, SI_SUB_CLOCKS, SI_ORDER_MIDDLE, ntp_init, NULL);<br=
>
> <br>
> The ntp_init() function sets a couple of global variables to zero. The=
se<br>
> variables should be in the .bss section. Are they not already cleared =
during<br>
> the kernel loading?<br>
This is nop indeed, be it linked in kernel, or loaded as a module.<br></blo=
ckquote><div><br></div><div>I just looked at the posted patch, and I concur=
.</div><div><br></div><div>Warner=C2=A0</div></div></div>
--000000000000b0ea4d05d7716fe2--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfoOVekuyZkkvdgzRQPYhgyOWjaqLagD0HNMFPquQAt1hw>