[FE training-materials-updates] Boot slides: improve "quiet" measurements

[Michael Opdenacker]

Repository : git://git.free-electrons.com/training-materials.git

On branch  : master
Link       : http://git.free-electrons.com/training-materials/commit/?id=839d67e86f237d43ec315672f24083e17a605df9

>---------------------------------------------------------------

commit 839d67e86f237d43ec315672f24083e17a605df9
Author: Michael Opdenacker <michael.opdenacker at free-electrons.com>
Date:   Mon Apr 21 15:41:11 2014 +0200

    Boot slides: improve "quiet" measurements
    
    - Only count kernel time
      I was previously counting bootloader time too,
      and it was huge (copying the kernel from a slow USB drive)
      Would have giving the impression that the Atmel kernel
      was taking 20 s to boot!
    
    Signed-off-by: Michael Opdenacker <michael.opdenacker at free-electrons.com>


>---------------------------------------------------------------

839d67e86f237d43ec315672f24083e17a605df9
 .../boottime-filesystems/boottime-filesystems.tex  |    8 +++++---
 slides/boottime-kernel/boottime-kernel.tex         |   12 ++++++------
 2 files changed, 11 insertions(+), 9 deletions(-)

diff --git a/slides/boottime-filesystems/boottime-filesystems.tex b/slides/boottime-filesystems/boottime-filesystems.tex
index 0dd33bc..e47cfeb 100644
--- a/slides/boottime-filesystems/boottime-filesystems.tex
+++ b/slides/boottime-filesystems/boottime-filesystems.tex
@@ -70,9 +70,11 @@ For raw flash storage
       belongs to each file.
 \item Need to use the \code{CONFIG_JFFS2_SUMMARY} kernel option
       to store such information in flash. This dramatically reduces
-      mount time (benchmark on ARM: from 16 s to 0.8 s for a 128 MB partition).
-\item Rather poor read and write performance (compared to YAFFS2 and
-      UBIFS)
+      mount time.
+\item Benchmark on ARM:\\
+      from 16 s to 0.8 s for a 128 MB partition.
+\item Rather poor read and write performance,\\
+      compared to YAFFS2 and UBIFS.
 \end{itemize}
 \end{frame}
 
diff --git a/slides/boottime-kernel/boottime-kernel.tex b/slides/boottime-kernel/boottime-kernel.tex
index dd13798..3582cbb 100644
--- a/slides/boottime-kernel/boottime-kernel.tex
+++ b/slides/boottime-kernel/boottime-kernel.tex
@@ -145,20 +145,20 @@ Conclusion: don't use LZO for now.
 \begin{frame}
 \frametitle{Turning off console output}
 \begin{itemize}
-\item Console output is actually taking a lot of time. You
-      probably don't need it in production. It can be disabled by
+\item Console output is actually taking a lot of time (very slow device).
+      Probably not needed in production. Disable it by
       passing the \code{quiet} argument on the kernel command line.
 \item You will still be able to use \code{dmesg} to get the kernel
       messages.
-\item Time to start the \code{init} program, on Atmel
-      SAMA5D3 Xplained (ARM), Linux 3.10:
+\item Time between starting the kernel and starting the \code{init}
+      program, on Atmel SAMA5D3 Xplained (ARM), Linux 3.10:
       \newline\newline
     \begin{tabular}{| l || c | c | c |}
     \hline
     & Time & Diff \\
     \hline
-    Without \code{quiet} & 20.380 s & \\
-    With \code{quiet} & 19.308 s & -1.072 s\\
+    Without \code{quiet} & 2.352 s & \\
+    With \code{quiet} & 1.285 s & -1.067 s\\
     \hline
     \end{tabular}
       \newline