diff mbox series

systemd: check SERIAL_CONSOLES for serial-getty-generator PACKAGECONFIG

Message ID 20250313052729.1809525-1-Qi.Chen@windriver.com
State New
Headers show
Series systemd: check SERIAL_CONSOLES for serial-getty-generator PACKAGECONFIG | expand

Commit Message

Chen, Qi March 13, 2025, 5:27 a.m. UTC 
From: Chen Qi <Qi.Chen@windriver.com>

If SERIAL_CONSOLES is empty, then systemd-serialgetty recipe does basically
nothing, and making systemd RDEPENDS on it does not help.

So a better default behavior is that if SERIAL_CONSOLES is empty, then
we put serial-getty-generator in PACKAGECONFIG which let systemd to
generate corresponding service instances.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
---
 meta/recipes-core/systemd/systemd_257.3.bb | 1 +
 1 file changed, 1 insertion(+)

Comments

Ross Burton March 17, 2025, 5:47 p.m. UTC | #1 
On 13 Mar 2025, at 05:27, Chen Qi via lists.openembedded.org <Qi.Chen=windriver.com@lists.openembedded.org> wrote:
> So a better default behavior is that if SERIAL_CONSOLES is empty, then
> we put serial-getty-generator in PACKAGECONFIG which let systemd to
> generate corresponding service instances.

Whilst this seems like a good idea, it would also make the systemd recipe machine-specific.  Considering systemd provides libsystemd and libudev, this is not a good idea.

Ross
Chen, Qi March 18, 2025, 3:08 a.m. UTC | #2 
Hi Ross,

Let me provide some background of the problem I was trying to solve here.
Some BSP layers don't set SERIAL_CONSOLES, and they have worked for a long time with previous systemd. Now systemd-getty-generator is deleted by default, and they don't get any serial console.

How about using the following solution:
Do not delete anything when serial-getty-generator is in PACKAGECONFIG.

I suspect deleting systemd-getty-generator will have potential negative impact on getty@.service. I'm not sure though.

Regards,
Qi

-----Original Message-----
From: Ross Burton <Ross.Burton@arm.com> 
Sent: Tuesday, March 18, 2025 1:47 AM
To: Chen, Qi <Qi.Chen@windriver.com>
Cc: openembedded-core@lists.openembedded.org
Subject: Re: [OE-core][PATCH] systemd: check SERIAL_CONSOLES for serial-getty-generator PACKAGECONFIG

On 13 Mar 2025, at 05:27, Chen Qi via lists.openembedded.org <Qi.Chen=windriver.com@lists.openembedded.org> wrote:
> So a better default behavior is that if SERIAL_CONSOLES is empty, then 
> we put serial-getty-generator in PACKAGECONFIG which let systemd to 
> generate corresponding service instances.

Whilst this seems like a good idea, it would also make the systemd recipe machine-specific.  Considering systemd provides libsystemd and libudev, this is not a good idea.

Ross
Chen, Qi March 19, 2025, 6:37 a.m. UTC | #3 
On 3/18/25 11:08, Chen Qi via lists.openembedded.org wrote:
> Hi Ross,
>
> Let me provide some background of the problem I was trying to solve here.
> Some BSP layers don't set SERIAL_CONSOLES, and they have worked for a long time with previous systemd. Now systemd-getty-generator is deleted by default, and they don't get any serial console.
>
> How about using the following solution:
> Do not delete anything when serial-getty-generator is in PACKAGECONFIG.
>
> I suspect deleting systemd-getty-generator will have potential negative impact ongetty@.service. I'm not sure though.

OK. I just checked. According to 
https://www.freedesktop.org/software/systemd/man/latest/systemd-getty-generator.html, 
this affects |serial-getty@.service|, |console-getty.service and 
||container-getty@.service.
|

Anyway, I'll send out V2 using the above solution.

Please help review when convenient.

Regards,
Qi

>
> Regards,
> Qi
>
> -----Original Message-----
> From: Ross Burton<Ross.Burton@arm.com>  
> Sent: Tuesday, March 18, 2025 1:47 AM
> To: Chen, Qi<Qi.Chen@windriver.com>
> Cc:openembedded-core@lists.openembedded.org
> Subject: Re: [OE-core][PATCH] systemd: check SERIAL_CONSOLES for serial-getty-generator PACKAGECONFIG
>
> On 13 Mar 2025, at 05:27, Chen Qi via lists.openembedded.org<Qi.Chen=windriver.com@lists.openembedded.org>  wrote:
>> So a better default behavior is that if SERIAL_CONSOLES is empty, then
>> we put serial-getty-generator in PACKAGECONFIG which let systemd to
>> generate corresponding service instances.
> Whilst this seems like a good idea, it would also make the systemd recipe machine-specific.  Considering systemd provides libsystemd and libudev, this is not a good idea.
>
> Ross
>
> -=-=-=-=-=-=-=-=-=-=-=-
> Links: You receive all messages sent to this group.
> View/Reply Online (#213133):https://lists.openembedded.org/g/openembedded-core/message/213133
> Mute This Topic:https://lists.openembedded.org/mt/111674751/3618072
> Group Owner:openembedded-core+owner@lists.openembedded.org
> Unsubscribe:https://lists.openembedded.org/g/openembedded-core/unsub  [Qi.Chen@windriver.com]
> -=-=-=-=-=-=-=-=-=-=-=-
>
diff mbox series

Patch

diff --git a/meta/recipes-core/systemd/systemd_257.3.bb b/meta/recipes-core/systemd/systemd_257.3.bb
index 64fb8fe69a..f1f6055637 100644
--- a/meta/recipes-core/systemd/systemd_257.3.bb
+++ b/meta/recipes-core/systemd/systemd_257.3.bb
@@ -92,6 +92,7 @@  PACKAGECONFIG ??= " \
     quotacheck \
     randomseed \
     resolved \
+    ${@'' if d.getVar('SERIAL_CONSOLES') else 'serial-getty-generator'} \
     set-time-epoch \
     sysusers \
     timedated \