[Buildroot] [PATCH 1/1] package/libglib2: fix dbus system socket location
jean-pierre cartal
jpcartal at free.fr
Tue Apr 20 07:09:44 UTC 2021
Hello Thomas,
After further checking, I realized that because of historical reasons
we've been mounting a tmpfs directory on top on /var as part of the boot
process, hence loosing the /var/run -> /run link in the process.
Sorry for the erroneous patch.
Regards
Le 19/04/2021 à 23:43, Jean-pierre Cartal a écrit :
> Hello Thomas,
>
> We recently moved to the new 2021.02.x LTS version from previous
> 2020.02.x LTS version.
>
> While testing this new branch we discovered that we could not connect
> to dbus anymore when using libglib2 because the dbus socket location
> changed from /var/run to /run.
>
> I'll double check tomorrow but I'm quite sure that those are 2
> distinct directories and not a symlink in our builds.
>
> Regards
>
> Le 19 avril 2021 23:28:10 GMT+02:00, Thomas Petazzoni
> <thomas.petazzoni at bootlin.com> a écrit :
>
> Hello Jean-Pierre,
>
> +Norbert in Cc, since he changed the DBus socket location in commit
> 6b9a75a5bd1b1a559b7352003588d1461de8c4d1.
>
> On Thu, 15 Apr 2021 16:42:26 +0200
> Jean-pierre Cartal <jpcartal at free.fr> wrote:
>
> Fix dbus system socket location moved from
> /var/run/dbus/system_bus_socket to /run/dbus/system_bus_socket
> Signed-off-by: Jean-pierre Cartal <jpcartal at free.fr>
> ------------------------------------------------------------------------
> package/libglib2/0005-dbus-system-socket-path.patch | 11
> +++++++++++ 1 file changed, 11 insertions(+) create mode
> 100644 package/libglib2/0005-dbus-system-socket-path.patch
> diff --git
> a/package/libglib2/0005-dbus-system-socket-path.patch
> b/package/libglib2/0005-dbus-system-socket-path.patch new file
> mode 100644 index 0000000000..bb2ec33dc6 --- /dev/null +++
> b/package/libglib2/0005-dbus-system-socket-path.patch @@ -0,0
> +1,11 @@ +diff -Naur libglib2-2.66.8.orig/gio/gdbusaddress.c
> libglib2-2.66.8/gio/gdbusaddress.c +---
> libglib2-2.66.8.orig/gio/gdbusaddress.c 2021-03-18
> 14:47:48.256693000 +0100 ++++
> libglib2-2.66.8/gio/gdbusaddress.c 2021-04-15
> 16:14:28.519296584 +0200 +@@ -1331,7 +1331,7 @@ + + if (ret ==
> NULL) + { +- ret = g_strdup
> ("unix:path=/var/run/dbus/system_bus_socket"); ++ ret =
> g_strdup ("unix:path=/run/dbus/system_bus_socket"); + } + break;
>
>
> On the form, all patches in Buildroot need to have a proper commit
> description and Signed-off-by line.
>
> But on the actual content, I am wondering if this is the right
> direction to take. Do we want to keep this non-upstreamable patch
> around forever ?
>
> According to commit 6b9a75a5bd1b1a559b7352003588d1461de8c4d1, /var/run
> is supposed to be a symlink to /run, so your patch should be a no-op.
> Both our systemd and sysv skeletons ensure this symlink exist.
>
> Could you provide more details about the issue ?
>
> Thomas
>
>
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
More information about the buildroot
mailing list