Skip to contentSkip to conversation

linux-stable: backport usb related commit to 3.10 and 3.12?

PATCH Schrey 2 participants 4 messages Aug 23, 2014 – Sep 25, 2014 USB

lore
Share & tools
lore

Schrey

author
#0 lore raw

Greetings.

On an administered server, I noticed that there's a constant load around 0.70 when running kernels 3.10 and 3.12, even if the system is doing nothing, and in single user mode. Culprit seems to be the process 'khubd'.
No such effect when running kernels 3.4 or 3.14.

I reverse bisected linux-stable and found that the following commit fixes it for me, for both kernel 3.10 and 3.12:

[08d1dec6f4054e3613f32051d9b149d4203ce0d2] usb:hub set hub->change_bits when over-current happens

Could that commit be backported to the stable 3.10 and 3.12 series?

(Disclaimer: I'm not a programmer, but I'd be happy to help with any testing involved)

Regards
Ingmar

--------------------------------------------------------------------------

$ git diff ac5166bcdb43889a5bd837f5076b78049e1f8bca 08d1dec6f4054e3613f32051d9b149d4203ce0d2
drivers/usb/core/hub.c +2 -1
index c1422a0..babba88 100644
@@ -1147,7 +1147,8 @@ static void hub_activate(struct usb_hub *hub, enum hub_activation_type type)
                        /* Tell khubd to disconnect the device or
                         * check for a new connection
                         */
-                       if (udev || (portstatus & USB_PORT_STAT_CONNECTION))
+                       if (udev || (portstatus & USB_PORT_STAT_CONNECTION) ||
+                           (portstatus & USB_PORT_STAT_OVERCURRENT))
                                set_bit(port1, hub->change_bits);
 
                } else if (portstatus & USB_PORT_STAT_ENABLE) {
Signature · 1 line
this space for rent
author
#1 lore raw

Hi,

resending; multiple systems are and have been working fine for me, with that commit backported to kernel 3.10 and 3.12.

2014-08-23:

Quoted 44 lines

Greetings.

On an administered server, I noticed that there's a constant load around 0.70 when running kernels 3.10 and 3.12, even if the system is doing nothing, and in single user mode. Culprit seems to be the process 'khubd'.
No such effect when running kernels 3.4 or 3.14.

I reverse bisected linux-stable and found that the following commit fixes it for me, for both kernel 3.10 and 3.12:

[08d1dec6f4054e3613f32051d9b149d4203ce0d2] usb:hub set hub->change_bits when over-current happens

Could that commit be backported to the stable 3.10 and 3.12 series?

(Disclaimer: I'm not a programmer, but I'd be happy to help with any testing involved)

Regards
Ingmar

--------------------------------------------------------------------------

$ git diff ac5166bcdb43889a5bd837f5076b78049e1f8bca 08d1dec6f4054e3613f32051d9b149d4203ce0d2
drivers/usb/core/hub.c +1 -1
index c1422a0..babba88 100644
@@ -1147,7 +1147,8 @@ static void hub_activate(struct usb_hub *hub, 
 enum hub_activation_type type)
                        /* Tell khubd to disconnect the device or
                         * check for a new connection
                         */
-                       if (udev || (portstatus & 
 USB_PORT_STAT_CONNECTION))
+                       if (udev || (portstatus & 

USB_PORT_STAT_CONNECTION) ||
+ (portstatus & USB_PORT_STAT_OVERCURRENT))

                                set_bit(port1, hub->change_bits);

                } else if (portstatus & USB_PORT_STAT_ENABLE) {
Signature · 1 line
this space for rent
#2 lore raw
In reply to Schrey
Quoted 38 lines · Schrey

Greetings.

On an administered server, I noticed that there's a constant load around 0.70 when running kernels 3.10 and 3.12, even if the system is doing nothing, and in single user mode. Culprit seems to be the process 'khubd'.
No such effect when running kernels 3.4 or 3.14.

I reverse bisected linux-stable and found that the following commit fixes it for me, for both kernel 3.10 and 3.12:

[08d1dec6f4054e3613f32051d9b149d4203ce0d2] usb:hub set hub->change_bits when over-current happens

Could that commit be backported to the stable 3.10 and 3.12 series?

(Disclaimer: I'm not a programmer, but I'd be happy to help with any testing involved)

Regards
Ingmar

--------------------------------------------------------------------------

$ git diff ac5166bcdb43889a5bd837f5076b78049e1f8bca
08d1dec6f4054e3613f32051d9b149d4203ce0d2
drivers/usb/core/hub.c +1 -1
index c1422a0..babba88 100644
@@ -1147,7 +1147,8 @@ static void hub_activate(struct usb_hub *hub, enum
 hub_activation_type type)
                        /* Tell khubd to disconnect the device or
                         * check for a new connection
                         */
-                       if (udev || (portstatus & USB_PORT_STAT_CONNECTION))
+                       if (udev || (portstatus & USB_PORT_STAT_CONNECTION)
Quoted 6 lines
||
+                           (portstatus & USB_PORT_STAT_OVERCURRENT))
                                set_bit(port1, hub->change_bits);

                } else if (portstatus & USB_PORT_STAT_ENABLE) {

I think this could be eligible for backporting. I

#3 lore raw
Quoted 43 lines · Frans Klaver
Quoted 42 lines · Schrey

Greetings.

On an administered server, I noticed that there's a constant load around 0.70 when running kernels 3.10 and 3.12, even if the system is doing nothing, and in single user mode. Culprit seems to be the process 'khubd'.
No such effect when running kernels 3.4 or 3.14.

I reverse bisected linux-stable and found that the following commit fixes it for me, for both kernel 3.10 and 3.12:

[08d1dec6f4054e3613f32051d9b149d4203ce0d2] usb:hub set hub->change_bits when over-current happens

Could that commit be backported to the stable 3.10 and 3.12 series?

(Disclaimer: I'm not a programmer, but I'd be happy to help with any testing involved)

Regards
Ingmar

--------------------------------------------------------------------------

$ git diff ac5166bcdb43889a5bd837f5076b78049e1f8bca
08d1dec6f4054e3613f32051d9b149d4203ce0d2
drivers/usb/core/hub.c +1 -1
index c1422a0..babba88 100644
@@ -1147,7 +1147,8 @@ static void hub_activate(struct usb_hub *hub, enum
 hub_activation_type type)
                        /* Tell khubd to disconnect the device or
                         * check for a new connection
                         */
-                       if (udev || (portstatus & USB_PORT_STAT_CONNECTION))
+                       if (udev || (portstatus & USB_PORT_STAT_CONNECTION)
||
+                           (portstatus & USB_PORT_STAT_OVERCURRENT))
                                set_bit(port1, hub->change_bits);
Quoted 4 lines
Quoted 2 lines
                } else if (portstatus & USB_PORT_STAT_ENABLE) {

I think this could be eligible for backporting. I

I was going to say: It seems to tick the boxes in
Documentation/stable_kernel_rules.txt.

But then gmail thought it would be a good idea to send out the e-mail.

Frans