[opensuse-kernel] Enable CONFIG_USB_SUSPEND?
Playing with powertop, I got the following suggestion: Suggestion: Enable the CONFIG_USB_SUSPEND kernel configuration option. This option will automatically disable UHCI USB when not in use, and may save approximately 1 Watt of power. Is this something we should do in our kernel? Or what is the risk for this? Will it break automatic access to hot-pluggable devices? Andreas -- Andreas Jaeger, Director Platform/openSUSE, aj@suse.de SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
Am Mittwoch 01 August 2007 schrieb Andreas Jaeger:
Playing with powertop, I got the following suggestion:
Suggestion: Enable the CONFIG_USB_SUSPEND kernel configuration option. This option will automatically disable UHCI USB when not in use, and may save approximately 1 Watt of power.
Is this something we should do in our kernel? Or what is the risk for this? Will it break automatic access to hot-pluggable devices?
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Oliver Neukum
Am Mittwoch 01 August 2007 schrieb Andreas Jaeger:
Playing with powertop, I got the following suggestion:
Suggestion: Enable the CONFIG_USB_SUSPEND kernel configuration option. This option will automatically disable UHCI USB when not in use, and may save approximately 1 Watt of power.
Is this something we should do in our kernel? Or what is the risk for this? Will it break automatic access to hot-pluggable devices?
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist.
So, what is the best way to move forward? Enable it directly after 10.3 is out so that people using alpha releases can report problems? Andreas -- Andreas Jaeger, Director Platform/openSUSE, aj@suse.de SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
On Wed, Aug 01, 2007 at 08:36:20PM +0200, Oliver Neukum wrote:
Am Mittwoch 01 August 2007 schrieb Andreas Jaeger:
Playing with powertop, I got the following suggestion:
Suggestion: Enable the CONFIG_USB_SUSPEND kernel configuration option. This option will automatically disable UHCI USB when not in use, and may save approximately 1 Watt of power.
Is this something we should do in our kernel? Or what is the risk for this? Will it break automatic access to hot-pluggable devices?
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist.
Yeah, I'm a bit leary of adding this to our kernel just yet, it's still not quite stable on all devices because other operating systems do not do this and hence, a lot of devices are broken with it :( So far, the majority seem to be scanners and some printers. thanks, greg k-h -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Am Mittwoch 01 August 2007 schrieb Greg KH:
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist.
Yeah, I'm a bit leary of adding this to our kernel just yet, it's still not quite stable on all devices because other operating systems do not do this and hence, a lot of devices are broken with it :(
Yes, on the other hand, something needs to be done about the case of not using USB. I'd activate it only for the hub driver. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
On Wed, Aug 01, 2007 at 09:19:24PM +0200, Oliver Neukum wrote:
Am Mittwoch 01 August 2007 schrieb Greg KH:
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist.
Yeah, I'm a bit leary of adding this to our kernel just yet, it's still not quite stable on all devices because other operating systems do not do this and hence, a lot of devices are broken with it :(
Yes, on the other hand, something needs to be done about the case of not using USB. I'd activate it only for the hub driver.
How about the other way around? We should have a way to dynamically add devices to the quirk list. That would probably help the majority of the people who run into crappy hardware, and also allow people to easily test things without having to rebuild their whole kernel. Anyone want to make up a patch to do this? thanks, greg k-h -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Am Mittwoch 01 August 2007 schrieb Greg KH:
Yes, on the other hand, something needs to be done about the case of not using USB. I'd activate it only for the hub driver.
How about the other way around? We should have a way to dynamically add devices to the quirk list. That would probably help the majority of the people who run into crappy hardware, and also allow people to easily test things without having to rebuild their whole kernel.
Do you consider $DEVICE/power/level insufficient? The newest version of SANE comes with an udev rule that uses it to disable autosuspend for scanners. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
On Thu, Aug 02, 2007 at 07:50:27AM +0200, Oliver Neukum wrote:
Am Mittwoch 01 August 2007 schrieb Greg KH:
Yes, on the other hand, something needs to be done about the case of not using USB. I'd activate it only for the hub driver.
How about the other way around? ?We should have a way to dynamically add devices to the quirk list. ?That would probably help the majority of the people who run into crappy hardware, and also allow people to easily test things without having to rebuild their whole kernel.
Do you consider $DEVICE/power/level insufficient?
Ah, no that works, but it isn't an automated way for users to handle it easily. Perhaps a simple YAST module to add new device ids to a udev rule to handle this automatically?
The newest version of SANE comes with an udev rule that uses it to disable autosuspend for scanners.
Do we ship this latest version of SANE with this rule in 10.3? thanks, greg k-h -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Am Donnerstag 02 August 2007 schrieb Greg KH:
On Thu, Aug 02, 2007 at 07:50:27AM +0200, Oliver Neukum wrote:
Am Mittwoch 01 August 2007 schrieb Greg KH:
Yes, on the other hand, something needs to be done about the case of not using USB. I'd activate it only for the hub driver.
How about the other way around? ?We should have a way to dynamically add devices to the quirk list. ?That would probably help the majority of the people who run into crappy hardware, and also allow people to easily test things without having to rebuild their whole kernel.
Do you consider $DEVICE/power/level insufficient?
Ah, no that works, but it isn't an automated way for users to handle it easily. Perhaps a simple YAST module to add new device ids to a udev rule to handle this automatically?
Actually, it would be nice to have a GUI for runtime power management with an option to make a setting permanent for a device. It would be nice if this weren't specific to USB.
The newest version of SANE comes with an udev rule that uses it to disable autosuspend for scanners.
Do we ship this latest version of SANE with this rule in 10.3?
No. Newest meaning added 2007-07-29 to SANE's CVS. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Hi, On Wed, 1 Aug 2007, Oliver Neukum wrote:
Am Mittwoch 01 August 2007 schrieb Greg KH:
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist.
Yeah, I'm a bit leary of adding this to our kernel just yet, it's still not quite stable on all devices because other operating systems do not do this and hence, a lot of devices are broken with it :(
Yes, on the other hand, something needs to be done about the case of not using USB. I'd activate it only for the hub driver.
That (or something to that effect) would be nice, so that I don't have to unload the USB modules all the time to save some more power. Ciao, Michael. -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
On Thu, Aug 02, 2007 at 11:29:17AM +0200, Michael Matz wrote:
Hi,
On Wed, 1 Aug 2007, Oliver Neukum wrote:
Am Mittwoch 01 August 2007 schrieb Greg KH:
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist.
Yeah, I'm a bit leary of adding this to our kernel just yet, it's still not quite stable on all devices because other operating systems do not do this and hence, a lot of devices are broken with it :(
Yes, on the other hand, something needs to be done about the case of not using USB. I'd activate it only for the hub driver.
That (or something to that effect) would be nice, so that I don't have to unload the USB modules all the time to save some more power.
Ok, now that it seems the SANE package will handle keeping all scanners from autosuspending, I don't have an objection to enabling this option in the kernel. I'll update the quirk table with the latest from upstream also. Oliver, any objection from you about this? thanks, greg k-h -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Am Donnerstag 02 August 2007 schrieb Greg KH:
That (or something to that effect) would be nice, so that I don't have to unload the USB modules all the time to save some more power.
Ok, now that it seems the SANE package will handle keeping all scanners from autosuspending, I don't have an objection to enabling this option in the kernel. I'll update the quirk table with the latest from upstream also.
Oliver, any objection from you about this?
No objection. I'll keep gathering quirks, not only for scanners. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
On Wednesday 01 August 2007 20:36:20 Oliver Neukum wrote:
It will _not_ break hotplug. But it'll break all non-compliant devices unless they are in the list list of quirky devices.However, we know that list is incomplete. But we don't know how incomplete. As a compromise, we might change the default to enable it only for devices with drivers in the kernel. That would be a kind of whitelist.
If that is done there should be definitely a way to add more devices to the white list at runtime. Perhaps with a sysfs interface? Perhaps it could be also done with hal hardware profiles? e.g. users could report what devices work and some list is slowly updated and perhaps pushed out in updates. -Andi -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Hi, has CONFIG_USB_SUSPEND now been enabled in the kernel? Or was that not the conclusion? Thanks, Andreas -- Andreas Jaeger, Director Platform/openSUSE, aj@suse.de SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
On Fri, Aug 03, 2007 at 07:56:09PM +0200, Andreas Jaeger wrote:
Hi,
has CONFIG_USB_SUSPEND now been enabled in the kernel? Or was that not the conclusion?
That was the conclusion, however, based on the current conversation on the linux-usb-devel list, with the Ubuntu and Red Hat developers telling us of all of the problems that were caused when they did it, I think we should hold off for now. thanks, greg k-h -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
On Fri, Aug 03, 2007 at 06:54:20PM -0700, Greg KH wrote:
That was the conclusion, however, based on the current conversation on the linux-usb-devel list, with the Ubuntu and Red Hat developers telling us of all of the problems that were caused when they did it, I think we should hold off for now.
My usb printer stops working too. A few seconds after loading usblp the device goes away and lsusb doesn't even show it, then it appears again after a few second because lsusb triggers a scan, usblp finds it and then in a few seconds it's gone again. Initially I set support_autosuspend=0 in usblp.c to work around it. If you've patches to test let me know. -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Am Dienstag 14 August 2007 schrieb Andrea Arcangeli:
On Fri, Aug 03, 2007 at 06:54:20PM -0700, Greg KH wrote:
That was the conclusion, however, based on the current conversation on the linux-usb-devel list, with the Ubuntu and Red Hat developers telling us of all of the problems that were caused when they did it, I think we should hold off for now.
My usb printer stops working too. A few seconds after loading usblp the device goes away and lsusb doesn't even show it, then it appears again after a few second because lsusb triggers a scan, usblp finds it and then in a few seconds it's gone again. Initially I set support_autosuspend=0 in usblp.c to work around it. If you've patches to test let me know.
Independant of everything else, your printer must be blacklisted. Please give me vendor:product. Apart from that I'd like to switch off autosuspend for usblp in 10.3. On the list we are currently discussing defaults. I'd like to wait a few days on results and incorporate them. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
On Tue, Aug 14, 2007 at 06:07:04PM +0200, Oliver Neukum wrote:
Am Dienstag 14 August 2007 schrieb Andrea Arcangeli:
On Fri, Aug 03, 2007 at 06:54:20PM -0700, Greg KH wrote:
That was the conclusion, however, based on the current conversation on the linux-usb-devel list, with the Ubuntu and Red Hat developers telling us of all of the problems that were caused when they did it, I think we should hold off for now.
My usb printer stops working too. A few seconds after loading usblp the device goes away and lsusb doesn't even show it, then it appears again after a few second because lsusb triggers a scan, usblp finds it and then in a few seconds it's gone again. Initially I set support_autosuspend=0 in usblp.c to work around it. If you've patches to test let me know.
Independant of everything else, your printer must be blacklisted. Please give me vendor:product.
04e8:326c Bus 002 Device 007: ID 04e8:326c Samsung Electronics Co., Ltd Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 1.10 bDeviceClass 0 (Defined at Interface level) bDeviceSubClass 0 bDeviceProtocol 0 bMaxPacketSize0 16 idVendor 0x04e8 Samsung Electronics Co., Ltd idProduct 0x326c bcdDevice 1.00 iManufacturer 1 Samsung iProduct 2 Samsung ML-2010 iSerial 3 3A21BKBL908052Z. bNumConfigurations 1 Configuration Descriptor: bLength 9 bDescriptorType 2 wTotalLength 32 bNumInterfaces 1 bConfigurationValue 1 iConfiguration 0 bmAttributes 0xc0 Self Powered MaxPower 0mA Interface Descriptor: bLength 9 bDescriptorType 4 bInterfaceNumber 0 bAlternateSetting 0 bNumEndpoints 2 bInterfaceClass 7 Printer bInterfaceSubClass 1 Printer bInterfaceProtocol 2 Bidirectional iInterface 0 Endpoint Descriptor: bLength 7 bDescriptorType 5 bEndpointAddress 0x02 EP 2 OUT bmAttributes 2 Transfer Type Bulk Synch Type None Usage Type Data wMaxPacketSize 0x0040 1x 64 bytes bInterval 0 Endpoint Descriptor: bLength 7 bDescriptorType 5 bEndpointAddress 0x83 EP 3 IN bmAttributes 2 Transfer Type Bulk Synch Type None Usage Type Data wMaxPacketSize 0x0040 1x 64 bytes bInterval 0 (btw, this printer works really good for me, it's just this feature that seems broken, if you're sure the usb driver is not to blame) Thanks. -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Am Dienstag 14 August 2007 schrieb Andrea Arcangeli:
(btw, this printer works really good for me, it's just this feature that seems broken, if you're sure the usb driver is not to blame)
Yes, the symptoms are typical and another printer of that vendor is known broken in that regard. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Am Freitag 03 August 2007 schrieb Andreas Jaeger:
Hi,
has CONFIG_USB_SUSPEND now been enabled in the kernel? Or was that not the conclusion?
I've deactivated it specifically for printers as it is quite unlikely that a computer is on battery and connected to a printer, while we had bug reports. A user space work around is in place for SANE, so I think we have the two areas complaints came from covered. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
participants (6)
-
Andi Kleen -
Andrea Arcangeli -
Andreas Jaeger -
Greg KH -
Michael Matz -
Oliver Neukum