From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bhavesh Davda Subject: Re: cyclictest overflow instance tracking -- irc continued Date: Tue, 16 Oct 2012 09:14:59 -0700 Message-ID: References: <507CAF63.7080703@am.sony.com> <490E56CB-B2C1-4B58-A9FA-C3D37BD84F52@vmware.com> Mime-Version: 1.0 Content-Type: text/plain; charset=Windows-1252 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Frank Rowand , "linux-rt-users@vger.kernel.org" To: John Kacur Return-path: Received: from smtp-outbound-2.vmware.com ([208.91.2.13]:56420 "EHLO smtp-outbound-2.vmware.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754799Ab2JPQPA convert rfc822-to-8bit (ORCPT ); Tue, 16 Oct 2012 12:15:00 -0400 In-Reply-To: <490E56CB-B2C1-4B58-A9FA-C3D37BD84F52@vmware.com> Content-Language: en-US Sender: linux-rt-users-owner@vger.kernel.org List-ID: Also, just to confirm: where is the GIT tree located? I was sync'ed to https://github.com/clrkwllms/rt-tests.git But I also see a tree at http://git.kernel.org/pub/scm/linux/kernel/git= /clrkwllms/rt-tests.git Thanks On Oct 16, 2012, at 9:11 AM, Bhavesh Davda wrote: > Sorry about that John, Frank! Don't know how that happened=85 >=20 > Please look for two separate patches against the right ToT: one for w= hitespace cleanup, and the other with this feature, in the next few hou= rs=85 >=20 > Thanks >=20 > - Bhavesh >=20 > On Oct 16, 2012, at 3:56 AM, John Kacur wrote: >=20 >>=20 >>=20 >> On Mon, 15 Oct 2012, Frank Rowand wrote: >>=20 >> Thanks for your further review Frank. Bhavesh - I think the right th= ing to=20 >> do here, is to ask you to respin your patch. Can you make sure you a= re=20 >> applying and testing against the latest. (0.84). Also, please put an= y=20 >> white space clean-ups in a separate patch to make it easier to read = the=20 >> important patch. Oh, and send these in a new thread. If people could= =20 >> preface their mail with [PATCH RT-TESTS] that would also help. >>=20 >> Thank You. >>=20 >> John Kacur >>=20 >>>=20 >>> 16:13 jkacur frowand_bot: would you like a Reviewed-by: credit for = the cyclictest histogram overflow instance tracking patch? >>> 16:13 jkacur can I just add that or is the reviewer supposed to req= uest it? >>> 17:02 frowand jkacur: sure, you can add my reviewed-by (at least fo= r the first version, i didn't really look at the second version in deta= il) >>>=20 >>> For the first version, it was a drive-by review, just obvious stuff= =2E >>>=20 >>> Now that I look at the second version a little closer (but still no= t in detail), it >>> looks like he did some white space clean up, which should probably = be a separate patch. >>>=20 >>> It also looks like he created his patch against an old version of c= yclictest, >>> because he has: >>>=20 >>> @@ -154,6 +155,7 @@ struct thread_stat { >>> long redmax; >>> long cycleofmax; >>> long hist_overflow; >>> + long num_outliers; >>> }; >>>=20 >>>=20 >>> but struct thread_stat currently has another field: >>>=20 >>> long redmax; >>> long cycleofmax; >>> long hist_overflow; >>> unsigned long now_after_next; >>> }; >>>=20 >>> and also the line numbers in the patch are off. >>>=20 >>>> from #linux-rt: >>>>=20 >>>> 17:07 jkacur frowand: hmm, the patch doesn't appear to be working = for me >>>> 17:07 jkacur I'll look at it more tomorrow >>>> 17:08 jkacur Thread 0: 18446744073709551615 00000 00001 00002 0000= 3 00004 00005 00006 00007 00008 # 02841 others >>>> 17:08 jkacur is an example of broken output >>>=20 >>> 18446744073709551615 is 0xffffffffffffffff (or -1LL) >>>=20 >>> It looks like in timerthread(), the line: >>>=20 >>> + stat->outliers[stat->num_outliers++] =3D stat->cycles - 1; >>>=20 >>> is picking up stat->cycles =3D=3D zero. Looking at the current ver= sion of cyclictest, for the >>> first event cycles should already be one at this point. But given = that the line numbers in >>> the patch are not consistent with the current cyclictest, I don't k= now where that chunk >>> actually got inserted. >>>=20 >>>=20 >>> -Frank >>>=20 >>>=20 >=20 > -- > Bhavesh Davda > bhavesh@vmware.com >=20 >=20 >=20 >=20 >=20 >=20 -- Bhavesh Davda bhavesh@vmware.com -- To unsubscribe from this list: send the line "unsubscribe linux-rt-user= s" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html