[PATCH BlueZ 0/1] Fix failure to register bt transport

[PATCH BlueZ 0/1] Fix failure to register bt transport

[Pete Beardmore]

Regurgitating the single commit message:

On initial logind update, the internal 'active' state is FALSE. Where
querying the current state (`sd_uid_get_state`) returns 'active' we do
NOT want to short-circuit the update to the internal state value and
ultimate transport initialisation.

---

I am guessing that this was a typo in the initial implementation
[obexd: Unregister profiles when the user is inactive](https://github.com/bluez/bluez/commit/765356e80262a20c359ec7722a5590232442d0d3)
which is where I landed having bisected for the error:
```
obexd[188706]: obexd/src/server.c:obex_server_init() No transport driver registered
obexd[188706]: obex_server_init failed
```
in my distro (ArchLinux)'s latest 5.83 version. I went from working bt
file transfer to not working post upgrade. Dropping back to 5.82
resolved the issue and I went from there.

I haven't seen any other complaints about this yet, which leads me to
speculate that IFF one upgrades, reboots (yuck!) and the systemd USER
SPACE obexd service is thus loaded prior to login, then the mechanism
that the implementation was aiming for (new seat detection) kicks in
and the internal active state is correctly set and the bt transport
initialised. This just didn't happen in my scenario.

Pete Beardmore (1):
  obexd: Fix failure to register bt transport

 obexd/src/logind.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

-- 
2.50.0

[PATCH BlueZ 1/1] obexd: Fix failure to register bt transport

[Pete Beardmore]

On initial logind update, the internal 'active' state is FALSE. Where
querying the current state (`sd_uid_get_state`) returns 'active' we do
NOT want to short-circuit the update to the internal state value and
ultimate transport initialisation.
---
 obexd/src/logind.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/obexd/src/logind.c b/obexd/src/logind.c
index b4279b209..e681dd39f 100644
--- a/obexd/src/logind.c
+++ b/obexd/src/logind.c
@@ -68,7 +68,7 @@ static int update(void)
 		return res;
 
 	if (state_is_active) {
-		if (!active)
+		if (active)
 			return 0;
 	} else {
 		res = sd_uid_get_seats(uid, 1, NULL);
-- 
2.50.0

RE: Fix failure to register bt transport

[bluez.test.bot]

[-- Attachment #1: Type: text/plain, Size: 1261 bytes --]

This is automated email and please do not reply to this email!

Dear submitter,

Thank you for submitting the patches to the linux bluetooth mailing list.
This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=975328

---Test result---

Test Summary:
CheckPatch                    PENDING   0.28 seconds
GitLint                       PENDING   0.29 seconds
BuildEll                      PASS      20.72 seconds
BluezMake                     PASS      3037.97 seconds
MakeCheck                     PASS      19.96 seconds
MakeDistcheck                 PASS      202.42 seconds
CheckValgrind                 PASS      279.52 seconds
CheckSmatch                   PASS      307.19 seconds
bluezmakeextell               PASS      129.87 seconds
IncrementalBuild              PENDING   0.26 seconds
ScanBuild                     PASS      928.37 seconds

Details
##############################
Test: CheckPatch - PENDING
Desc: Run checkpatch.pl script
Output:

##############################
Test: GitLint - PENDING
Desc: Run gitlint
Output:

##############################
Test: IncrementalBuild - PENDING
Desc: Incremental build with the patches in the series
Output:



---
Regards,
Linux Bluetooth

Re: [PATCH BlueZ 1/1] obexd: Fix failure to register bt transport

[Pauli Virtanen]

ti, 2025-06-24 kello 15:23 +0100, Pete Beardmore kirjoitti:
> On initial logind update, the internal 'active' state is FALSE. Where
> querying the current state (`sd_uid_get_state`) returns 'active' we do
> NOT want to short-circuit the update to the internal state value and
> ultimate transport initialisation.
> ---
>  obexd/src/logind.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/obexd/src/logind.c b/obexd/src/logind.c
> index b4279b209..e681dd39f 100644
> --- a/obexd/src/logind.c
> +++ b/obexd/src/logind.c
> @@ -68,7 +68,7 @@ static int update(void)
>  		return res;
>  
>  	if (state_is_active) {
> -		if (!active)
> +		if (active)
>  			return 0;
>  	} else {
>  		res = sd_uid_get_seats(uid, 1, NULL);

Just above this line there appears to be

	state_is_active = g_strcmp0(state, "active");

which looks like it's inverted too.

What this function probably should do is to make it simpler and
something like

	state_is_active = (g_strcmp0(state, "active") == 0);

	/* if active but no seat, consider as not active */
	if (state_is_active && sd_uid_get_seats(uid, 1, NULL) == 0)
		state_is_active = FALSE;

	if (state_is_active == active)
		return;

	active = state_is_active;
	emit callbacks;

-- 
Pauli Virtanen