lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJq09z5SEqCxC5jQ7mPS+rqr-U-eXgWH=mjcpHpWT7UF_7twxA@mail.gmail.com>
Date: Sat, 13 Jan 2024 11:24:27 -0300
From: Luiz Angelo Daros de Luca <luizluca@...il.com>
To: Christian Marangi <ansuelsmth@...il.com>
Cc: linus.walleij@...aro.org, alsi@...g-olufsen.dk, andrew@...n.ch, 
	f.fainelli@...il.com, olteanv@...il.com, davem@...emloft.net, 
	edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com, 
	arinc.unal@...nc9.com, netdev@...r.kernel.org
Subject: Re: [RFC net-next 0/2] net: dsa: realtek: fix LED support for rtl8366rb

> I'll try to get who scheduled the work and the stack from there. It
> might already pinpoint the cause. Checking if the work is pending
> during netdev trigger activation might also help. I'll also try to
> flush (or cancel) the work before activating the new trigger. I just
> don't know if I can flush (and that way, blocking) inside
> led_set_brightness().
>

Hi Christian,

I got it. It was actually from the netdev trigger. During activation we have:

        /* Check if hw control is active by default on the LED.
         * Init already enabled mode in hw control.
         */
        if (supports_hw_control(led_cdev)) {
                dev = led_cdev->hw_control_get_device(led_cdev);
                if (dev) {
                        const char *name = dev_name(dev);

                        set_device_name(trigger_data, name, strlen(name));
                        trigger_data->hw_control = true;

                        rc = led_cdev->hw_control_get(led_cdev, &mode);
                        if (!rc)
                                trigger_data->mode = mode;
                }
        }

The set_device_name calls set_baseline_state() that, at this point,
will start to monitor the device using sw control
(trigger_data->hw_control is only set afterwards). In
set_baseline_state(), it will call led_set_brightness in most
codepaths (all of them if trigger_data->hw_control is false). With
link down (and other situations), it will call
led_set_brightness(led_cdev, LED_OFF). If that led_set_brightness
takes some time to be processed, it will happen after the hw control
was configured, undoing what it previously just did.

Is there any good reason to call set_device_name before the led mode
and hw_control are defined? Will this break anything?

diff --git a/drivers/leds/trigger/ledtrig-netdev.c
b/drivers/leds/trigger/ledtrig-netdev.c
index d76214fa9ad8..6f72d55c187a 100644
--- a/drivers/leds/trigger/ledtrig-netdev.c
+++ b/drivers/leds/trigger/ledtrig-netdev.c
@@ -572,12 +572,13 @@ static int netdev_trig_activate(struct
led_classdev *led_cdev)
               if (dev) {
                       const char *name = dev_name(dev);

-                       set_device_name(trigger_data, name, strlen(name));
                       trigger_data->hw_control = true;

                       rc = led_cdev->hw_control_get(led_cdev, &mode);
                       if (!rc)
                               trigger_data->mode = mode;
+
+                       set_device_name(trigger_data, name, strlen(name));
               }
       }

With this patch, it will not undo the trigger setting in hardware
anymore. However, it now calls the hw_control_set 3 times during
activation:

1) set_device_name
2) register_netdevice_notifier on NETDEV_REGISTER
3) register_netdevice_notifier on NETDEV_UP

Anyway, calling it multiple times doesn't break anything.

Regards,

Luiz

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ