From mboxrd@z Thu Jan 1 00:00:00 1970 From: balbi@ti.com (Felipe Balbi) Date: Wed, 16 Jul 2014 13:57:32 -0500 Subject: Function Profiler broken on today's linux-next In-Reply-To: <20140716145442.15f25ec3@gandalf.local.home> References: <20140716162328.GJ1365@saruman.home> <20140716122921.71b118a9@gandalf.local.home> <20140716164141.GA3016@saruman.home> <20140716132413.1ede8e1d@gandalf.local.home> <20140716182944.GK3016@saruman.home> <20140716184152.GM3016@saruman.home> <20140716145442.15f25ec3@gandalf.local.home> Message-ID: <20140716185732.GN3016@saruman.home> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Jul 16, 2014 at 02:54:42PM -0400, Steven Rostedt wrote: > On Wed, 16 Jul 2014 13:41:52 -0500 > Felipe Balbi wrote: > > > Hi, > > > > On Wed, Jul 16, 2014 at 01:29:44PM -0500, Felipe Balbi wrote: > > > On Wed, Jul 16, 2014 at 01:24:13PM -0400, Steven Rostedt wrote: > > > > On Wed, 16 Jul 2014 11:41:42 -0500 > > > > Felipe Balbi wrote: > > > > > > > > > > > > > .config attached. It's actually an ARM platform, I can help out with > > > > > testing anything you need. > > > > > > > > In that case, can you see if it works under my repo? > > > > > > > > git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git > > > > branch: ftrace/core > > > > > > works fine, thanks. FWIW > > > > > > Tested-by: Felipe Balbi > > > > Actually it seems like it's not counting time properly if I run my test > > multiple times in a row: > > > > # /root/testusb -t 1 -c 100000 -s 2048 -a > > # head -3 trace_stat/function0 > > Function Hit Time Avg s^2 > > -------- --- ---- --- --- > > thread_interrupt 174876 166120839 us 949.935 us 3048498 us > > > > # /root/testusb -t 1 -c 100000 -s 2048 -a > > # head -3 trace_stat/function0 > > Function Hit Time Avg s^2 > > -------- --- ---- --- --- > > dwc3_thread_interrupt 286796 967267300 us 4190545199 us 3241109 us > > > > # /root/testusb -t 1 -c 100000 -s 2048 -a > > # head -3 trace_stat/function0 > > Function Hit Time Avg s^2 > > -------- --- ---- --- --- > > dwc3_thread_interrupt 401346 642632072 us 1601.192 us 1194316 us > > > > and so on... > > > > Did it use to work? In other words, is this a regression, or just some > another bug? It used to work at least on v3.12... It seems like it's not always either. After I rebooted, then it took a little effort to trigger it again. But if I just run my test in a loop, eventually it'll start miscalculating. -- balbi -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: Digital signature URL: