cgroup: Add note about IO distribution availability.

content/posts/cgroup v2 by hand.adoc

@@ -23,6 +23,7 @@ toc: true
 :url-nice: https://manpages.debian.org/buster/coreutils/nice.1.en.html
 :url-htop: https://htop.dev/
 :url-ionice: https://manpages.debian.org/buster/util-linux/ionice.1.en.html
+:url-linux-git: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
 
 We have Control Group v2 since 2016 but I had trouble finding good documentation
 on how to use it. Most tutorials and blog posts only cover v1 or are specific to
@@ -177,7 +178,7 @@ echo "$$" > cgroup.procs
 
 This control group will increase the IO priority and limit the write speed to 2
 pass:[<abbr title="Mebibyte, 1024 Kibibyte">MiB</abbr>] a second using the io
-controllerfootnote:[link:{url-kernel-doc}#io[Kernel documentation on Control
+controllerfootnote:io[link:{url-kernel-doc}#io[Kernel documentation on Control
 Group v2, section Controllers → IO]]. IO limits are set per device. You need to
 specify the major and minor device numbers of the _device_ (not partition) you
 want to limit (in my case it is “8:0” for `/dev/sda`). Run `lsblk` or `cat
@@ -208,8 +209,14 @@ bypassing caches.]. You will see that the speed is around 2 MiB a second.
 Kernel 5.14 introduced
 **blkio.prio.class**footnote:[link:{url-kernel-doc-14}#io-priority[Kernel
 documentation on Control Group v2, section “IO Priority”]] that controls the IO
-priority. It seems to work like link:{url-ionice}[ionice]. I could not test it
-yet, since I run kernel 5.10.
+priority. It seems to work similar to link:{url-ionice}[ionice].
+
+[IMPORTANT]
+Weight based distribution (*io.weight*) is available only if cfq-iosched is in
+use and absolute bandwidth or IOPS limit distribution (*io.max*) is not
+available for blk-mq devices.footnote:io[] The CFQ scheduler was removed in
+kernel 5.0.footnote:[link:{url-linux-git}/commit/?id=f382fb0[git commit: “block:
+remove legacy IO schedulers”]]
 
 == Controlling process numbers