<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content="text/html; charset=utf-8" http-equiv=Content-Type><!-- flashmail style begin -->
<STYLE type=text/css>
body {border-width:0;margin:0}
img {border:0;margin:0;padding:0}
</STYLE>
<BASE target=_blank><!-- flashmail style end -->
<META name=GENERATOR content="MSHTML 11.00.10586.494"></HEAD>
<BODY
style="BORDER-LEFT-WIDTH: 0px; FONT-SIZE: 10.5pt; FONT-FAMILY: times new roman; BORDER-RIGHT-WIDTH: 0px; BORDER-BOTTOM-WIDTH: 0px; COLOR: #000000; MARGIN: 12px; LINE-HEIGHT: 1.5; BORDER-TOP-WIDTH: 0px"
marginheight="0" marginwidth="0">
<DIV>Hello:</DIV>
<DIV> </DIV>
<DIV style="TEXT-INDENT: 2em">The previous pic 'temp.jpg' is snipped from the
tutor16's slides <A
href="http://www.litmus-rt.org/tutor16/slides.pdf">http://www.litmus-rt.org/tutor16/slides.pdf</A></DIV>
<DIV style="TEXT-INDENT: 2em">The 'temp_2.jpg' and 'temp_3.jpg' are snipped from
my platform. You can see the gaps in different period are variant and
unpredictable.</DIV>
<DIV style="TEXT-INDENT: 2em">I run the xubuntu img on virtualbox, and the
hardware information is listed as follows:</DIV>
<DIV style="TEXT-INDENT: 2em">H/W path
Device Class
Description<BR>===================================================<BR>
system VirtualBox
()<BR>/0
bus
VirtualBox<BR>/0/0
memory 128KiB
BIOS<BR>/0/1
memory 3952MiB System
memory<BR>/0/2
processor Intel(R) Core(TM) i5-3317U CPU @
1.70GHz<BR>/0/100
bridge 440FX - 82441FX PMC
[Natoma]<BR>/0/100/1
bridge 82371SB PIIX3 ISA [Natoma/Triton
II]<BR>/0/100/1.1
storage 82371AB/EB/MB PIIX4
IDE<BR>/0/100/2
display VirtualBox Graphics
Adapter<BR>/0/100/3
eth0 network
82540EM Gigabit Ethernet
Controller<BR>/0/100/4
generic VirtualBox Guest
Service<BR>/0/100/5
multimedia 82801AA AC'97 Audio
Controller<BR>/0/100/6
bus KeyLargo/Intrepid
USB<BR>/0/100/7
bridge 82371AB/EB/MB PIIX4
ACPI<BR>/0/100/d
storage 82801HM/HEM (ICH8M/ICH8M-E) SATA
Control<BR>/0/3
scsi1 storage
<BR>/0/3/0.0.0 /dev/cdrom
disk DVD
reader<BR>/0/4
scsi2 storage
<BR>/0/4/0.0.0 /dev/sda
disk 34GB VBOX
HARDDISK<BR>/0/4/0.0.0/1 /dev/sda1
volume 29GiB EXT4
volume<BR>/0/4/0.0.0/2 /dev/sda2
volume 2099MiB Extended
partition<BR>/0/4/0.0.0/2/5 /dev/sda5
volume 2097MiB Linux swap / Solaris
partition<BR></DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: Verdana; COLOR: #c0c0c0"
align=left>2016-07-19
<HR id=SignNameHR
style="BORDER-TOP: #c0c0c0 1px solid; HEIGHT: 1px; BORDER-RIGHT: 0px; WIDTH: 122px; BORDER-BOTTOM: 0px; BORDER-LEFT: 0px"
align=left>
<SPAN id=_FlashSignName>shijunjie92</SPAN> </DIV>
<HR
style="BORDER-TOP: #c0c0c0 1px solid; HEIGHT: 1px; BORDER-RIGHT: 0px; BORDER-BOTTOM: 0px; BORDER-LEFT: 0px">
<BLOCKQUOTE id=ntes-flashmail-quote
style="FONT-SIZE: 10pt; FONT-FAMILY: Verdana; PADDING-LEFT: 0px; MARGIN-LEFT: 0px">
<DIV><STRONG>发件人：</STRONG>Björn Brandenburg <bbb@mpi-sws.org></DIV>
<DIV><STRONG>发送时间：</STRONG>2016-07-19 10:43</DIV>
<DIV><STRONG>主题：</STRONG>Re: [LITMUS^RT] The gap between the release time and
the real beginning of execution</DIV>
<DIV><STRONG>收件人：</STRONG>"litmus-dev"<litmus-dev@lists.litmus-rt.org></DIV>
<DIV><STRONG>抄送：</STRONG></DIV>
<DIV> </DIV>
<DIV><BR>
<DIV>
<BLOCKQUOTE type="cite">
<DIV>On 19 Jul 2016, at 11:13, shijunjie92 <<A
href="mailto:shijunjie92@gmail.com">shijunjie92@gmail.com</A>>
wrote:</DIV>
<DIV><BR>
<DIV
style="FONT-SIZE: 14px; FONT-FAMILY: 'times new roman'; WHITE-SPACE: normal; WORD-SPACING: 0px; TEXT-TRANSFORM: none; FONT-WEIGHT: normal; FONT-STYLE: normal; LETTER-SPACING: normal; TEXT-INDENT: 2em; font-variant-caps: normal; -webkit-text-stroke-width: 0px">When
I use the feature trace tool to draw the Gantt chart of the cpus' execution.
I notice that there is a gap between the release time and the beginning of
the real execution. I have traced the code of 'rtspin.c' but found nothing.
Can anyone explain the reason of the random gap or how to reduce
the gap? (sometimes the gap is 2 ms)</DIV></DIV></BLOCKQUOTE></DIV><BR>
<DIV>That’s release latency. Either your platform’s timers are extremely
imprecise (this is usually the case under virtualization), or you have
extremely long non-preemptive sections, or there is some extreme overhead. In
mainline LITMUS^RT (with a normal config), I would rule out the last two
issues.</DIV>
<DIV><BR></DIV>
<DIV>What hardware platform are you running your test on? This looks like a
trace from a VM to me.</DIV>
<DIV><BR></DIV>
<DIV>- Björn</DIV>
<DIV><BR></DIV></DIV></BLOCKQUOTE></BODY></HTML>