From: Richard Purdie <richard.purdie@linuxfoundation.org>
To: Kang Kai <kai.kang@windriver.com>
Cc: bitbake-devel@lists.openembedded.org, Zhenfeng.Zhao@windriver.com
Subject: Re: [PATCH 2/2] cooker: treat termination by disk monitor as failure
Date: Tue, 24 Jul 2012 09:14:40 +0100 [thread overview]
Message-ID: <1343117680.21788.114.camel@ted> (raw)
In-Reply-To: <0c6e89f1ac8abee8be17f821bda9b193b1761f6c.1343113295.git.kai.kang@windriver.com>
On Tue, 2012-07-24 at 15:04 +0800, Kang Kai wrote:
> [Yocto #2168]
>
> Check the return value of function diskMonitor.check() in runqueue then
> pass it to cooker. The cooker treats that runqueue terminated by disk
> monitor as failure, then increase the build failure count if the return
> value is true.
>
> The UIs could check the failure count accordingly, and this will helpful
> for UIs such as Hob.
>
> Signed-off-by: Kang Kai <kai.kang@windriver.com>
> ---
> bitbake/lib/bb/cooker.py | 14 ++++++++++++--
> bitbake/lib/bb/event.py | 5 +++--
> bitbake/lib/bb/runqueue.py | 9 +++++----
> 3 files changed, 20 insertions(+), 8 deletions(-)
I have to say I don't really like this code. I agree that the UI needs
to know about diskmonitor failures but I'm not sure about changing the
BuiuldCompleted event in this way. I think we should add the exit status
to the BuildCompleted event as an extra field (leaving the failures
count alone as it means something specific right now).
The UI is going to want more information about what failed (which disk
was full, how full was it, where is it mounted?). The best approach
therefore could be a specific DiskFull event which gets fired and the UI
can act upon containing much more information.
Cheers,
Richard
> diff --git a/bitbake/lib/bb/cooker.py b/bitbake/lib/bb/cooker.py
> index 361bc88..58a10f7 100644
> --- a/bitbake/lib/bb/cooker.py
> +++ b/bitbake/lib/bb/cooker.py
> @@ -1095,7 +1095,7 @@ class BBCooker:
> rq.finish_runqueue(False)
> failures = 0
> try:
> - retval = rq.execute_runqueue()
> + retval, termedbydm = rq.execute_runqueue()
> except runqueue.TaskFailure as exc:
> failures += len(exc.args)
> retval = False
> @@ -1103,6 +1103,11 @@ class BBCooker:
> self.command.finishAsyncCommand()
> return False
>
> + if termedbydm:
> + bb.event.fire(bb.event.BuildCompleted(len(rq.rqdata.runq_fnid), buildname, item, failures, bymonitor=True), self.configuration.event_data)
> + self.command.finishAsyncCommand()
> + return False
> +
> if not retval:
> bb.event.fire(bb.event.BuildCompleted(len(rq.rqdata.runq_fnid), buildname, item, failures), self.configuration.event_data)
> self.command.finishAsyncCommand()
> @@ -1135,7 +1140,7 @@ class BBCooker:
> rq.finish_runqueue(False)
> failures = 0
> try:
> - retval = rq.execute_runqueue()
> + retval, termedbydm = rq.execute_runqueue()
> except runqueue.TaskFailure as exc:
> failures += len(exc.args)
> retval = False
> @@ -1143,6 +1148,11 @@ class BBCooker:
> self.command.finishAsyncCommand()
> return False
>
> + if termedbydm:
> + bb.event.fire(bb.event.BuildCompleted(len(rq.rqdata.runq_fnid), buildname, targets, failures, bymonitor=True), self.configuration.data)
> + self.command.finishAsyncCommand()
> + return False
> +
> if not retval:
> bb.event.fire(bb.event.BuildCompleted(len(rq.rqdata.runq_fnid), buildname, targets, failures), self.configuration.data)
> self.command.finishAsyncCommand()
> diff --git a/bitbake/lib/bb/event.py b/bitbake/lib/bb/event.py
> index 1116c0a..d719105 100644
> --- a/bitbake/lib/bb/event.py
> +++ b/bitbake/lib/bb/event.py
> @@ -305,14 +305,15 @@ class BuildStarted(BuildBase, OperationStarted):
>
> class BuildCompleted(BuildBase, OperationCompleted):
> """bbmake build run completed"""
> - def __init__(self, total, n, p, failures = 0):
> + def __init__(self, total, n, p, failures = 0, bymonitor = False):
> + if bymonitor:
> + failures += 1
> if not failures:
> OperationCompleted.__init__(self, total, "Building Succeeded")
> else:
> OperationCompleted.__init__(self, total, "Building Failed")
> BuildBase.__init__(self, n, p, failures)
>
> -
> class NoProvider(Event):
> """No Provider for an Event"""
>
> diff --git a/bitbake/lib/bb/runqueue.py b/bitbake/lib/bb/runqueue.py
> index 0a8c723..efe3623 100644
> --- a/bitbake/lib/bb/runqueue.py
> +++ b/bitbake/lib/bb/runqueue.py
> @@ -867,6 +867,7 @@ class RunQueue:
> """
>
> retval = 0.5
> + termedbydm = False
>
> if self.state is runQueuePrepare:
> self.rqexe = RunQueueExecuteDummy(self)
> @@ -882,7 +883,7 @@ class RunQueue:
> self.rqexe = RunQueueExecuteScenequeue(self)
>
> if self.state in [runQueueSceneRun, runQueueRunning, runQueueCleanUp]:
> - self.dm.check(self)
> + termedbydm = self.dm.check(self)
>
> if self.state is runQueueSceneRun:
> retval = self.rqexe.execute()
> @@ -914,14 +915,14 @@ class RunQueue:
>
> if self.state is runQueueComplete:
> # All done
> - return False
> + return False, termedbydm
>
> if self.state is runQueueChildProcess:
> print("Child process, eeek, shouldn't happen!")
> - return False
> + return False, termedbydm
>
> # Loop
> - return retval
> + return retval, termedbydm
>
> def finish_runqueue(self, now = False):
> if not self.rqexe:
next prev parent reply other threads:[~2012-07-24 8:26 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-07-24 7:04 [PATCH 0/2] treat runqueue terminated by disk monitor as failure Kang Kai
2012-07-24 7:04 ` [PATCH 1/2] monitordisk: update return value Kang Kai
2012-07-24 7:04 ` [PATCH 2/2] cooker: treat termination by disk monitor as failure Kang Kai
2012-07-24 8:14 ` Richard Purdie [this message]
2012-07-24 23:24 ` Kang Kai
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1343117680.21788.114.camel@ted \
--to=richard.purdie@linuxfoundation.org \
--cc=Zhenfeng.Zhao@windriver.com \
--cc=bitbake-devel@lists.openembedded.org \
--cc=kai.kang@windriver.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox