* [PATCH libgpiod] dbus: client: tests: fix the way we wait for simulated chips to appear
@ 2025-02-03 17:20 Bartosz Golaszewski
2025-02-05 13:35 ` Bartosz Golaszewski
0 siblings, 1 reply; 2+ messages in thread
From: Bartosz Golaszewski @ 2025-02-03 17:20 UTC (permalink / raw)
To: Vincent Fazio, Kent Gibson, Linus Walleij, Erik Schilling,
Phil Howard, Viresh Kumar, Koichiro Den
Cc: linux-gpio, Bartosz Golaszewski
From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
The org.freedesktop.DBus.Peer.Ping() method isn't correct for checking
if a given object is exposed by a service. It only lets us know if the
service is available no matter the object on which it's called (in fact:
it will not return an error even if it's called on a nonexistent object).
While this has worked for most part by accident, if the timing isn't
right, we may start calling methods in chips which are not yet fully
registered. We should try to read the chip's property instead and keep
on waiting for as long as org.freedesktop.DBus.Properties.Get() returns
an error.
While at it: fix the counter increment as its syntax is invalid.
Fixes: a5ab76da1e0a ("dbus: add the D-Bus daemon, command-line client and tests")
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
---
dbus/client/gpiocli-test.bash | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/dbus/client/gpiocli-test.bash b/dbus/client/gpiocli-test.bash
index f210183..1d2337d 100755
--- a/dbus/client/gpiocli-test.bash
+++ b/dbus/client/gpiocli-test.bash
@@ -14,11 +14,11 @@ wait_for_sim() {
while true
do
- gdbus call --system --dest io.gpiod1 \
- --object-path /io/gpiod1/chips/"$1" \
- --method org.freedesktop.DBus.Peer.Ping > /dev/null 2>&1 && break
+ gdbus call --system --dest io.gpiod1 --object-path /io/gpiod1/chips/"$1" \
+ --method org.freedesktop.DBus.Properties.Get \
+ io.gpiod1.Chip Label > /dev/null 2>&1 && break
sleep 0.01
- COUNTER=$($COUNTER - 1)
+ COUNTER=$(expr $COUNTER - 1)
if [ "$COUNTER" -eq 0 ]
then
fail "error waiting for the GPIO sim chip to be exported on the bus"
---
base-commit: d6457b28e29a8edadcb619d389878ea99cd4bab4
change-id: 20250203-fix-dbus-wait-for-chip-49e36ad58ee2
Best regards,
--
Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH libgpiod] dbus: client: tests: fix the way we wait for simulated chips to appear
2025-02-03 17:20 [PATCH libgpiod] dbus: client: tests: fix the way we wait for simulated chips to appear Bartosz Golaszewski
@ 2025-02-05 13:35 ` Bartosz Golaszewski
0 siblings, 0 replies; 2+ messages in thread
From: Bartosz Golaszewski @ 2025-02-05 13:35 UTC (permalink / raw)
To: Vincent Fazio, Kent Gibson, Linus Walleij, Erik Schilling,
Phil Howard, Viresh Kumar, Koichiro Den, Bartosz Golaszewski
Cc: Bartosz Golaszewski, linux-gpio
From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
On Mon, 03 Feb 2025 18:20:27 +0100, Bartosz Golaszewski wrote:
> The org.freedesktop.DBus.Peer.Ping() method isn't correct for checking
> if a given object is exposed by a service. It only lets us know if the
> service is available no matter the object on which it's called (in fact:
> it will not return an error even if it's called on a nonexistent object).
>
> While this has worked for most part by accident, if the timing isn't
> right, we may start calling methods in chips which are not yet fully
> registered. We should try to read the chip's property instead and keep
> on waiting for as long as org.freedesktop.DBus.Properties.Get() returns
> an error.
>
> [...]
Applied, thanks!
[1/1] dbus: client: tests: fix the way we wait for simulated chips to appear
commit: f52874b19beab4d0a1cfc076f7a0eb7d6df5d1e1
Best regards,
--
Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2025-02-05 13:35 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-02-03 17:20 [PATCH libgpiod] dbus: client: tests: fix the way we wait for simulated chips to appear Bartosz Golaszewski
2025-02-05 13:35 ` Bartosz Golaszewski
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).