From: Tang Yizhou <yizhou.tang@shopee.com>

The current text states that peers with a lower latency target are
throttled, which is the opposite of the actual behavior. In fact,
blk-iolatency throttles peer groups with a higher latency target in order
to protect the more latency-sensitive group.

In addition, peer groups without a configured latency target are also
throttled, as they are treated as lower priority compared to groups with
explicit latency requirements.

Update the documentation to reflect the correct throttling behavior.

Signed-off-by: Tang Yizhou <yizhou.tang@shopee.com>
---
 Documentation/admin-guide/cgroup-v2.rst | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/Documentation/admin-guide/cgroup-v2.rst b/Documentation/admin-guide/cgroup-v2.rst
index 7f5b59d95fce..d6f7ef08b67d 100644
--- a/Documentation/admin-guide/cgroup-v2.rst
+++ b/Documentation/admin-guide/cgroup-v2.rst
@@ -2238,8 +2238,9 @@ IO Latency
 
 This is a cgroup v2 controller for IO workload protection.  You provide a group
 with a latency target, and if the average latency exceeds that target the
-controller will throttle any peers that have a lower latency target than the
-protected workload.
+controller will throttle any peers that have a higher latency target than the
+protected workload, as well as peers that do not have a latency target
+configured.
 
 The limits are only applied at the peer level in the hierarchy.  This means that
 in the diagram below, only groups A, B, and C will influence each other, and
@@ -2265,8 +2266,9 @@ How IO Latency Throttling Works
 
 io.latency is work conserving; so as long as everybody is meeting their latency
 target the controller doesn't do anything.  Once a group starts missing its
-target it begins throttling any peer group that has a higher target than itself.
-This throttling takes 2 forms:
+target it begins throttling any peer group that has a higher target than itself,
+as well as any peer group without a latency target. This throttling takes 2
+forms:
 
 - Queue depth throttling.  This is the number of outstanding IO's a group is
   allowed to have.  We will clamp down relatively quickly, starting at no limit
-- 
2.43.0

On Wed, Jan 14, 2026 at 07:08:37PM +0800, Tang Yizhou <yizhou.tang@shopee.com> wrote:
> From: Tang Yizhou <yizhou.tang@shopee.com>
> 
> The current text states that peers with a lower latency target are
> throttled, which is the opposite of the actual behavior. In fact,
> blk-iolatency throttles peer groups with a higher latency target in order
> to protect the more latency-sensitive group.
> 
> In addition, peer groups without a configured latency target are also
> throttled, as they are treated as lower priority compared to groups with
> explicit latency requirements.
> 
> Update the documentation to reflect the correct throttling behavior.
> 
> Signed-off-by: Tang Yizhou <yizhou.tang@shopee.com>
> ---
>  Documentation/admin-guide/cgroup-v2.rst | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)

Not a big deal but it could've been confusing.


Acked-by: Michal Koutný <mkoutny@suse.com>