Dear Lava users,
I'm looking for a way to measure, with Lava, time measurements between some power modes. Let's say the time between two messages :
- Standby exit trigger
- Actual standby exit from the kernel
The constraint is that we can't rely on kernel timestamps (frozen during standby) and the power process implies not only kernel but also boot stages. We can only rely on the time spent between two specific messages in the console. Basically, if it was done without Lava, we'd use a tool like grabserial.
Do you have any clue on what can be done?
Best regards,
On Fri, 30 Aug 2019 at 13:58, Denis HUMEAU denis.humeau@st.com wrote:
Dear Lava users,
I’m looking for a way to measure, with Lava, time measurements between some power modes. Let’s say the time between two messages :
Standby exit triggerActual standby exit from the kernelThe constraint is that we can’t rely on kernel timestamps (frozen during standby) and the power process implies not only kernel but also boot stages. We can only rely on the time spent between two specific messages in the console.
Do you mean that when you run dmesg you get the proper timestamps of events (even after they happened)?
Basically, if it was done without Lava, we’d use a tool like grabserial.
Do you have any clue on what can be done?
With android we did (probably) similar exercise with determining boot stages. All messages were taken either from dmesg or from logcat and parsed with the script on the target. Would that approach work for you?
milosz
Best regards,
Lava-users mailing list Lava-users@lists.lavasoftware.org https://lists.lavasoftware.org/mailman/listinfo/lava-users
Hi Milosz,
No, i can't rely on dmesg timestamps.
Your approach on Android may be interesting. Can you share something so that I can take a look?
Best regards,
Denis
-----Original Message----- From: Milosz Wasilewski milosz.wasilewski@linaro.org Sent: vendredi 30 août 2019 15:59 To: Denis HUMEAU denis.humeau@st.com Cc: Lava-users@lists.lavasoftware.org Subject: Re: [Lava-users] Power time measurements
On Fri, 30 Aug 2019 at 13:58, Denis HUMEAU denis.humeau@st.com wrote:
Dear Lava users,
I’m looking for a way to measure, with Lava, time measurements between some power modes. Let’s say the time between two messages :
Standby exit triggerActual standby exit from the kernelThe constraint is that we can’t rely on kernel timestamps (frozen during standby) and the power process implies not only kernel but also boot stages. We can only rely on the time spent between two specific messages in the console.
Do you mean that when you run dmesg you get the proper timestamps of events (even after they happened)?
Basically, if it was done without Lava, we’d use a tool like grabserial.
Do you have any clue on what can be done?
With android we did (probably) similar exercise with determining boot stages. All messages were taken either from dmesg or from logcat and parsed with the script on the target. Would that approach work for you?
milosz
Best regards,
Lava-users mailing list Lava-users@lists.lavasoftware.org https://lists.lavasoftware.org/mailman/listinfo/lava-users
On Tue, 3 Sep 2019 at 10:53, Denis HUMEAU denis.humeau@st.com wrote:
Hi Milosz,
No, i can't rely on dmesg timestamps.
If you were to do the measurement without LAVA, how would that work?
Your approach on Android may be interesting. Can you share something so that I can take a look?
Sure: https://github.com/Linaro/test-definitions/blob/master/automated/android/boo...
milosz
Best regards,
Denis
-----Original Message----- From: Milosz Wasilewski milosz.wasilewski@linaro.org Sent: vendredi 30 août 2019 15:59 To: Denis HUMEAU denis.humeau@st.com Cc: Lava-users@lists.lavasoftware.org Subject: Re: [Lava-users] Power time measurements
On Fri, 30 Aug 2019 at 13:58, Denis HUMEAU denis.humeau@st.com wrote:
Dear Lava users,
I’m looking for a way to measure, with Lava, time measurements between some power modes. Let’s say the time between two messages :
Standby exit triggerActual standby exit from the kernelThe constraint is that we can’t rely on kernel timestamps (frozen during standby) and the power process implies not only kernel but also boot stages. We can only rely on the time spent between two specific messages in the console.
Do you mean that when you run dmesg you get the proper timestamps of events (even after they happened)?
Basically, if it was done without Lava, we’d use a tool like grabserial.
Do you have any clue on what can be done?
With android we did (probably) similar exercise with determining boot stages. All messages were taken either from dmesg or from logcat and parsed with the script on the target. Would that approach work for you?
milosz
Best regards,
Lava-users mailing list Lava-users@lists.lavasoftware.org https://lists.lavasoftware.org/mailman/listinfo/lava-users
lava-users@lists.lavasoftware.org
-
Denis HUMEAU -
Milosz Wasilewski