LKML Archive on
help / color / mirror / Atom feed
From: "Marek Behún" <>
To: Andrew Lunn <>
Cc: Ansuel Smith <>,
	Vivien Didelot <>,
	Florian Fainelli <>,
	Vladimir Oltean <>,
	"David S. Miller" <>,
	Jakub Kicinski <>,
	Rob Herring <>,
	Jonathan Corbet <>, Pavel Machek <>,
	John Crispin <>,,,,,
Subject: Re: [RFC PATCH v2 1/5] leds: trigger: add API for HW offloading of triggers
Date: Mon, 8 Nov 2021 21:11:10 +0100	[thread overview]
Message-ID: <20211108211110.4ad78e41@thinkpad> (raw)
In-Reply-To: <>

On Mon, 8 Nov 2021 20:53:36 +0100
Andrew Lunn <> wrote:

> > I guess I will have to work on this again ASAP or we will end up with
> > solution that I don't like.
> > 
> > Nonetheless, what is your opinion about offloading netdev trigger vs
> > introducing another trigger?  
> It is a solution that fits the general pattern, do it in software, and
> offload it if possible.
> However, i'm not sure the software solution actually works very well.
> At least for switches. The two DSA drivers which implement
> get_stats64() simply copy the cached statistics. The XRS700X updates
> its cached values every 3000ms. The ar9331 is the same. Those are the
> only two switch drivers which implement get_stats64 and none implement
> get_stats. There was also was an issue that get_stats64() cannot
> perform blocking calls. I don't remember if that was fixed, but if
> not, get_stats64() is going to be pretty useless on switches.
> We also need to handle drivers which don't actually implement
> dev_get_stats(). That probably means only supporting offloads, all
> modes which cannot be offloaded need to be rejected. This is pretty
> much the same case of software control of the LEDs is not possible.
> Unfortunately, dev_get_stats() does not return -EOPNOTSUPP, you need
> to look at dev->netdev_ops->ndo_get_stats64 and
> dev->netdev_ops->ndo_get_stats.
> Are you working on Marvell switches? Have you implemented
> get_stats64() for mv88e6xxx? How often do you poll the hardware for
> the stats?
> Given this, i think we need to bias the API so that it very likely
> ends up offloading, if offloading is available.

I am working with Marvell PHYs and Marvell switches. I am aware of the
problem that SW netdev does not work for switches because we don't have
uptodate data.

It seems to me that yes, if the user wants to blink the LEDs on
activity on switch port, the netdev trigger should only work in offload
mode and only on LEDs that are connected to switch pins, unless we
implement a mechanism to get statistics at lest 10 times a second
(which could maybe be done on marvell switches by reading the registers
via ethernet frames instead of MDIO).

Marvell switches don't seem to support rx only / tx only activity
blinking, only rx/tx. I think this could be solved by making rx/tx
sysfs files for netdev trigger behave so that writing to one would also
write to another.

But since currently netdev trigger does not work for these switches, I
think making it so that it works at least to some fashion (in HW
supported modes) would still be better that current situation.

I will try to look into this, maybe work together with Ansuel.


  parent reply	other threads:[~2021-11-08 20:11 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-08  0:24 [RFC PATCH v2 0/5] Adds support for PHY LEDs with offload triggers Ansuel Smith
2021-11-08  0:24 ` [RFC PATCH v2 1/5] leds: trigger: add API for HW offloading of triggers Ansuel Smith
2021-11-08  2:29   ` Randy Dunlap
2021-11-08 14:04   ` Andrew Lunn
2021-11-08 15:16     ` Ansuel Smith
2021-11-08 16:13       ` Marek Behún
2021-11-08 16:46         ` Ansuel Smith
2021-11-08 17:35           ` Marek Behún
2021-11-08 17:58             ` Ansuel Smith
2021-11-08 18:41               ` Marek Behún
2021-11-08 19:08                 ` Ansuel Smith
2021-11-08 19:17                   ` Marek Behún
2021-11-08 17:46         ` Andrew Lunn
2021-11-08 17:56           ` Marek Behún
2021-11-08 19:53             ` Andrew Lunn
2021-11-08 20:03               ` Vladimir Oltean
2021-11-08 20:11               ` Marek Behún [this message]
2021-11-08 20:15                 ` Ansuel Smith
2021-11-08 17:37       ` Andrew Lunn
2021-11-08  0:24 ` [RFC PATCH v2 2/5] leds: add function to configure offload leds Ansuel Smith
2021-11-08  2:22   ` Randy Dunlap
2021-11-08  0:24 ` [RFC PATCH v2 3/5] leds: trigger: add offload-phy-activity trigger Ansuel Smith
2021-11-08  2:24   ` Randy Dunlap
2021-11-08 14:17   ` Andrew Lunn
2021-11-08 15:19     ` Ansuel Smith
2021-11-08  0:24 ` [RFC PATCH v2 4/5] net: dsa: qca8k: add LEDs support Ansuel Smith
2021-11-08  2:26   ` Randy Dunlap
2021-11-08  0:25 ` [RFC PATCH v2 5/5] dt-bindings: net: dsa: qca8k: add LEDs definition example Ansuel Smith

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211108211110.4ad78e41@thinkpad \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).