Woken by "Interrupt" after IO state #2066
-
|
Hi, I would like to understand whether this could be due to a change in system behavior or some information loss during trace collection. Please help. Thank you!
|
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
|
This is a change in the data we collect which makes wakeup tracking more accurate. We used to say "woken up by thread X" on older versions of Android even if that was not true and we were woken up by an interrupt happening on the CPU the thread was scheduled on. On newer Android versions, we made a change to our collection/processing of the ftrace event to actually track whether the wakeup came from an interrupt and expose that information in the UI. |
Beta Was this translation helpful? Give feedback.
-
|
This sounds cool! Could you please tell me the location of the code changes and the commit link mentioned above? |
Beta Was this translation helpful? Give feedback.
-
|
The on device change was in https://source.chromium.org/chromium/_/chromium/external/github.com/google/perfetto/+/e3822f239667add9627c1a3e8064a9ce1a6e4254 The trace_processor/UI side change was in https://source.chromium.org/chromium/_/chromium/external/github.com/google/perfetto/+/4dc08287cdb7125ab1afbc98efb982568196a3ae |
Beta Was this translation helpful? Give feedback.
This is a change in the data we collect which makes wakeup tracking more accurate. We used to say "woken up by thread X" on older versions of Android even if that was not true and we were woken up by an interrupt happening on the CPU the thread was scheduled on. On newer Android versions, we made a change to our collection/processing of the ftrace event to actually track whether the wakeup came from an interrupt and expose that information in the UI.