[PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[Tom Gundersen]

udev will only manage static nodes that exist at the time udev is started, so
creating static nodes later on will likely not behave as expected. In
particular, recreating the static nodes at run-time will reset any permissions
udev may have applied to the nodes at boot.

See <https://bugzilla.redhat.com/show_bug.cgi?id=1147248> and the discussion
following <http://permalink.gmane.org/gmane.comp.sysutils.systemd.devel/23795>.

Note that this requires (the yet to be released) systemd v217 or a backport
of systemd patch 8c94052ee543c3598a3c7b0c46688150aa2c6168.
---
 tools/static-nodes.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/static-nodes.c b/tools/static-nodes.c
index 0195390..4bdacc8 100644
--- a/tools/static-nodes.c
+++ b/tools/static-nodes.c
@@ -95,7 +95,7 @@ static int write_tmpfiles(FILE *out, char modname[], char devname[], char type,
 			return EXIT_FAILURE;
 	}
 
-	ret = fprintf(out, "%c /dev/%s 0600 - - - %u:%u\n",
+	ret = fprintf(out, "%c! /dev/%s 0600 - - - %u:%u\n",
 		      type, devname, maj, min);
 	if (ret < 0)
 		return EXIT_FAILURE;
-- 
2.1.0

Re: [PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[Lucas De Marchi]

On Mon, Oct 27, 2014 at 2:55 PM, Tom Gundersen <teg@jklm.no> wrote:
> udev will only manage static nodes that exist at the time udev is started, so
> creating static nodes later on will likely not behave as expected. In
> particular, recreating the static nodes at run-time will reset any permissions
> udev may have applied to the nodes at boot.
>
> See <https://bugzilla.redhat.com/show_bug.cgi?id=1147248> and the discussion
> following <http://permalink.gmane.org/gmane.comp.sysutils.systemd.devel/23795>.
>
> Note that this requires (the yet to be released) systemd v217 or a backport
> of systemd patch 8c94052ee543c3598a3c7b0c46688150aa2c6168.

Is this patch going to systemd-stable?

-- 
Lucas De Marchi

Re: [PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[Tom Gundersen]

On Tue, Oct 28, 2014 at 3:31 AM, Lucas De Marchi
<lucas.de.marchi@gmail.com> wrote:
> On Mon, Oct 27, 2014 at 2:55 PM, Tom Gundersen <teg@jklm.no> wrote:
>> udev will only manage static nodes that exist at the time udev is started, so
>> creating static nodes later on will likely not behave as expected. In
>> particular, recreating the static nodes at run-time will reset any permissions
>> udev may have applied to the nodes at boot.
>>
>> See <https://bugzilla.redhat.com/show_bug.cgi?id=1147248> and the discussion
>> following <http://permalink.gmane.org/gmane.comp.sysutils.systemd.devel/23795>.
>>
>> Note that this requires (the yet to be released) systemd v217 or a backport
>> of systemd patch 8c94052ee543c3598a3c7b0c46688150aa2c6168.
>
> Is this patch going to systemd-stable?

I think that makes sense, as the current status is broken in distros.
Not entirely sure though, as the --boot logic is new functionality.
Zbigniew, what do you think?

Cheers,

Tom

Re: [PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[Tom Gundersen]

On Tue, Oct 28, 2014 at 12:16 PM, Tom Gundersen <teg@jklm.no> wrote:
> On Tue, Oct 28, 2014 at 3:31 AM, Lucas De Marchi
> <lucas.de.marchi@gmail.com> wrote:
>> On Mon, Oct 27, 2014 at 2:55 PM, Tom Gundersen <teg@jklm.no> wrote:
>>> udev will only manage static nodes that exist at the time udev is started, so
>>> creating static nodes later on will likely not behave as expected. In
>>> particular, recreating the static nodes at run-time will reset any permissions
>>> udev may have applied to the nodes at boot.
>>>
>>> See <https://bugzilla.redhat.com/show_bug.cgi?id=1147248> and the discussion
>>> following <http://permalink.gmane.org/gmane.comp.sysutils.systemd.devel/23795>.
>>>
>>> Note that this requires (the yet to be released) systemd v217 or a backport
>>> of systemd patch 8c94052ee543c3598a3c7b0c46688150aa2c6168.
>>
>> Is this patch going to systemd-stable?
>
> I think that makes sense, as the current status is broken in distros.
> Not entirely sure though, as the --boot logic is new functionality.
> Zbigniew, what do you think?

Sorry, I misread the commit logs. The --boot logic is already in
stable, and I think my follow up patch should also go there.

Cheers,

Tom

Re: [PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[Zbigniew Jędrzejewski-Szmek]

On Tue, Oct 28, 2014 at 12:23:53PM +0100, Tom Gundersen wrote:
> On Tue, Oct 28, 2014 at 12:16 PM, Tom Gundersen <teg@jklm.no> wrote:
> > On Tue, Oct 28, 2014 at 3:31 AM, Lucas De Marchi
> > <lucas.de.marchi@gmail.com> wrote:
> >> On Mon, Oct 27, 2014 at 2:55 PM, Tom Gundersen <teg@jklm.no> wrote:
> >>> udev will only manage static nodes that exist at the time udev is started, so
> >>> creating static nodes later on will likely not behave as expected. In
> >>> particular, recreating the static nodes at run-time will reset any permissions
> >>> udev may have applied to the nodes at boot.
> >>>
> >>> See <https://bugzilla.redhat.com/show_bug.cgi?id=1147248> and the discussion
> >>> following <http://permalink.gmane.org/gmane.comp.sysutils.systemd.devel/23795>.
> >>>
> >>> Note that this requires (the yet to be released) systemd v217 or a backport
> >>> of systemd patch 8c94052ee543c3598a3c7b0c46688150aa2c6168.
> >>
> >> Is this patch going to systemd-stable?
> >
> > I think that makes sense, as the current status is broken in distros.
> > Not entirely sure though, as the --boot logic is new functionality.
> > Zbigniew, what do you think?
Yeah, I'll pull it into -stable.

(But the kmod patch doesn't really depend on your patch 'units:
tmpfiles-setup-dev - allow unsafe file creation to happen in /dev at
boot', just on the --boot functionality, which I think was done for F20,
so can be considered available.)

Zbyszek

> Sorry, I misread the commit logs. The --boot logic is already in
> stable, and I think my follow up patch should also go there.

Re: [PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[Tom Gundersen]

On Tue, Oct 28, 2014 at 2:21 PM, Zbigniew J=C4=99drzejewski-Szmek
<zbyszek@in.waw.pl> wrote:
> On Tue, Oct 28, 2014 at 12:23:53PM +0100, Tom Gundersen wrote:
>> On Tue, Oct 28, 2014 at 12:16 PM, Tom Gundersen <teg@jklm.no> wrote:
>> > On Tue, Oct 28, 2014 at 3:31 AM, Lucas De Marchi
>> > <lucas.de.marchi@gmail.com> wrote:
>> >> On Mon, Oct 27, 2014 at 2:55 PM, Tom Gundersen <teg@jklm.no> wrote:
>> >>> udev will only manage static nodes that exist at the time udev is st=
arted, so
>> >>> creating static nodes later on will likely not behave as expected. I=
n
>> >>> particular, recreating the static nodes at run-time will reset any p=
ermissions
>> >>> udev may have applied to the nodes at boot.
>> >>>
>> >>> See <https://bugzilla.redhat.com/show_bug.cgi?id=3D1147248> and the =
discussion
>> >>> following <http://permalink.gmane.org/gmane.comp.sysutils.systemd.de=
vel/23795>.
>> >>>
>> >>> Note that this requires (the yet to be released) systemd v217 or a b=
ackport
>> >>> of systemd patch 8c94052ee543c3598a3c7b0c46688150aa2c6168.
>> >>
>> >> Is this patch going to systemd-stable?
>> >
>> > I think that makes sense, as the current status is broken in distros.
>> > Not entirely sure though, as the --boot logic is new functionality.
>> > Zbigniew, what do you think?
> Yeah, I'll pull it into -stable.
>
> (But the kmod patch doesn't really depend on your patch 'units:
> tmpfiles-setup-dev - allow unsafe file creation to happen in /dev at
> boot', just on the --boot functionality, which I think was done for F20,
> so can be considered available.)

It is true that even F20 will parse the new kmod.conf file correctly,
but since --boot is not passed to tmpfiles, the entries will be
ignored. So you really want that patch as well.

Cheers,

Tom

Re: [PATCH] static-nodes: indicate that creation of static nodes should only happen at boot

[Lucas De Marchi]

On Tue, Oct 28, 2014 at 12:45 PM, Tom Gundersen <teg@jklm.no> wrote:
> On Tue, Oct 28, 2014 at 2:21 PM, Zbigniew J=C4=99drzejewski-Szmek
> <zbyszek@in.waw.pl> wrote:
>> On Tue, Oct 28, 2014 at 12:23:53PM +0100, Tom Gundersen wrote:
>>> On Tue, Oct 28, 2014 at 12:16 PM, Tom Gundersen <teg@jklm.no> wrote:
>>> > On Tue, Oct 28, 2014 at 3:31 AM, Lucas De Marchi
>>> > <lucas.de.marchi@gmail.com> wrote:
>>> >> On Mon, Oct 27, 2014 at 2:55 PM, Tom Gundersen <teg@jklm.no> wrote:
>>> >>> udev will only manage static nodes that exist at the time udev is s=
tarted, so
>>> >>> creating static nodes later on will likely not behave as expected. =
In
>>> >>> particular, recreating the static nodes at run-time will reset any =
permissions
>>> >>> udev may have applied to the nodes at boot.
>>> >>>
>>> >>> See <https://bugzilla.redhat.com/show_bug.cgi?id=3D1147248> and the=
 discussion
>>> >>> following <http://permalink.gmane.org/gmane.comp.sysutils.systemd.d=
evel/23795>.
>>> >>>
>>> >>> Note that this requires (the yet to be released) systemd v217 or a =
backport
>>> >>> of systemd patch 8c94052ee543c3598a3c7b0c46688150aa2c6168.
>>> >>
>>> >> Is this patch going to systemd-stable?
>>> >
>>> > I think that makes sense, as the current status is broken in distros.
>>> > Not entirely sure though, as the --boot logic is new functionality.
>>> > Zbigniew, what do you think?
>> Yeah, I'll pull it into -stable.
>>
>> (But the kmod patch doesn't really depend on your patch 'units:
>> tmpfiles-setup-dev - allow unsafe file creation to happen in /dev at
>> boot', just on the --boot functionality, which I think was done for F20,
>> so can be considered available.)
>
> It is true that even F20 will parse the new kmod.conf file correctly,
> but since --boot is not passed to tmpfiles, the entries will be
> ignored. So you really want that patch as well.

Yeah, as far as I can see you need that as well.

Applied.

--=20
Lucas De Marchi