bbr-doc: update quick-start guide: no need for fq qdisc in Linux v4.13

Documentation/bbr-quick-start.md

@@ -27,7 +27,7 @@ typeset -x ZONE="us-west1-a"          # A GCE Zone
 ```
 
 Next, we can create a VM to build the kernel with BBR. This will create a
-high-cpu instance with SSD disks to compile our kernel:
+Google Cloud instance to compile our kernel:
 
 ```
 gcloud compute \
@@ -63,8 +63,13 @@ sudo apt-get upgrade
 
 ## Obtain kernel sources with TCP BBR
 
-Since BBR was only recently contributed to Linux, we'll need to compile a
-development kernel that includes this feature.
+TCP BBR is in Linux v4.9 and beyond. However, we recommend compiling from the
+latest sources, from the networking development branch. In particular, in the
+`davem/net-next` networking development branch (and Linux v4.13-rc1 and
+beyond) there is new support for 
+[TCP-level pacing](https://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git/commit/?id=218af599fa635b107cfe10acf3249c4dfe5e4123).
+This means that there is no longer a requirement to install the "fq" qdisc to
+use BBR. Any qdisc will do.
 
 For this guide, we'll grab the Linux networking development branch
 `davem/net-next` from `git.kernel.org`.
@@ -115,16 +120,6 @@ make -j`nproc`
 make -j`nproc` modules
 ```
 
-## Configure the machine
-
-On your GCE instance, configure the system, changing the default qdisc to fq,
-and default TCP congestion control to BBR:
-
-```
-sudo bash -c 'echo "net.core.default_qdisc=fq" >> /etc/sysctl.conf'
-sudo bash -c 'echo "net.ipv4.tcp_congestion_control=bbr" >> /etc/sysctl.conf'
-```
-
 ##  Install the kernel and reboot
 
 On your GCE instance, install the newly-compiled kernel and reboot:
@@ -150,32 +145,29 @@ build timestamp matching the kernel you compiled above:
 Linux bbrtest1 4.8.0-rc7+ #1 SMP Thu Sep 29 20:06:31 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
 ```
 
-Finally, on your GCE instance, confirm that the fq qdisc is installed and that
-BBR is being used:
+Finally, on your GCE instance, confirm that BBR is being used:
 
 ```
-tc qdisc show
 sysctl net.ipv4.tcp_congestion_control
 ```
 
 Enjoy!
 
 ## Further reading
 
-If you already have a kernel config for GCE, then you can enable BBR and
-FQ. On your GCE instance, check that if you run:
+If you already have a kernel config for GCE, then you can just enable BBR,
+rebuild, and reboot. On your GCE instance, check that if you run:
 
 ```
 cd /usr/src/net-next
-egrep '(CONFIG_TCP_CONG_BBR|CONFIG_NET_SCH_FQ)=' .config
+egrep '(_BBR)' .config
 ```
 
 then you see exactly the following lines:
 
 ```
 CONFIG_TCP_CONG_BBR=y
-CONFIG_NET_SCH_FQ=y
-
+CONFIG_DEFAULT_BBR=y
 ```
 
 If you want to create your own .config, then just remember to include those two

Documentation/config.gce

@@ -811,7 +811,7 @@ CONFIG_TCP_CONG_CUBIC=y
 # CONFIG_TCP_CONG_DCTCP is not set
 # CONFIG_TCP_CONG_CDG is not set
 CONFIG_TCP_CONG_BBR=y
-CONFIG_DEFAULT_CUBIC=y
+CONFIG_DEFAULT_BBR=y
 # CONFIG_DEFAULT_RENO is not set
 CONFIG_DEFAULT_TCP_CONG="cubic"
 CONFIG_TCP_MD5SIG=y