timeout when run "netcfg" or "getprop init.svc.bootanim" after android booted up
Bug #1001112 reported by
Yongqin Liu
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
LAVA Dispatcher |
Fix Released
|
High
|
Yongqin Liu |
Bug Description
There are pretty some cases that it is timed out when run "netcfg" or "getprop init.svc.bootanim" command after android booted up and displayed the console prompt.
From the log and by logging into the board's conmux connection, I found that the timeout seems caused by the suspend function.
So will try make disable_suspend the first thing after android booted up, and see if this can decrease the timeout rate.
Related branches
lp:~liuyq0307/lava-dispatcher/timeout_wait_home_screen
- Michael Hudson-Doyle (community): Approve
- Yongqin Liu: Approve (test with staging)
- Zygmunt Krynicki (community): Needs Information
-
Diff: 34 lines (+7/-5)1 file modifiedlava_dispatcher/client/base.py (+7/-5)
Changed in lava-dispatcher: | |
assignee: | nobody → Yongqin Liu (liuyq0307) |
importance: | Undecided → High |
status: | New → In Progress |
milestone: | none → 2012.05 |
Changed in lava-dispatcher: | |
milestone: | 2012.05 → 2012.06 |
status: | In Progress → Fix Committed |
Changed in lava-dispatcher: | |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
And another reason seems to the wrong usage of pexpect.TIMEOUT.
Below is the related talk with mwhudson:
<liuyq> mwhudson, in method wait_home_screen, it is now using "self.run(cmd, response= ['stopped' , pexpect.TIMEOUT], timeout=5)" to catch timeout g()." is not caught
<liuyq> mwhudson, but in fact, the "TIMEOUT: Timeout exceeded in read_nonblockin
<liuyq> mwhudson, so I consider how about trying catch TIMEOUT sentence there
<liuyq> mwhudson, not catch the timeout in the response
<liuyq> mwhudson, how do you think about that?
<mwhudson> liuyq: oh heh
<mwhudson> liuyq: that sounds like a bug actually
<mwhudson> liuyq: if you pass pexpect.TIMEOUT in response the code still waits for the prompt
<mwhudson> and that probably won't appear
<mwhudson> so it will raise an exception
<mwhudson> liuyq: can you file a bug on lava-dispatcher about this?
<mwhudson> liuyq: in the mean time, just catch pexpect.TIMEOUT I guess
<liuyq> mwhudson, sorry, why do you say "and that probably won't appear", the prompt “root@linaro” should appear
<mwhudson> liuyq: but it only waits one more second for the root@linaro to appear
<mwhudson> liuyq: if the timeout is 5 seconds and nothing has appeared, the root@linaro _probably_ won't appear in that 1 more second