Re: Check device use before detaching?
On Mon, Nov 18, 2013 at 08:19:38PM +, Luke-Jr wrote: > My software does not interface with keyboards under any condition. > > Some specific devices I'm working with include: > FTDI: Avalon, Butterfly Labs BitFORCE & BitFORCE SC, The Chili > CP210X: Bitfountain Block Erupter Emerald/Sapphire > CDC/ACM: Big Picture Mining, BTCFPGA ModMiner > PL2303: Icarus So you somehow got into a goldrush here! :) The software will probably be an daemon that will guarantee that thos gold-diggers are working 24/7? If so, I think the manufacturers of the devices did choose the wrong interface then. Regards, Michael -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0| Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917- | -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Monday, November 18, 2013 9:25:49 PM Greg KH wrote: > Ok, this is making no sense. As you really aren't saying exactly what > you are trying to do, I'm going to just give up and say you can't do it. As I said before, I want to use devices only if they aren't already in use. With hidapi's hidraw implementation, I'm pretty sure this works fine. But when failing over to hidapi's libusb implementation it always needs to interrupt other software since there is no way to determine if it's in use or not. Luke -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Mon, Nov 18, 2013 at 08:38:39PM +, Luke-Jr wrote: > On Monday, November 18, 2013 8:29:24 PM Greg KH wrote: > > On Mon, Nov 18, 2013 at 08:19:38PM +, Luke-Jr wrote: > > > My software does not interface with keyboards under any condition. > > > > > > Some specific devices I'm working with include: > > > FTDI: Avalon, Butterfly Labs BitFORCE & BitFORCE SC, The Chili > > > CP210X: Bitfountain Block Erupter Emerald/Sapphire > > > CDC/ACM: Big Picture Mining, BTCFPGA ModMiner > > > PL2303: Icarus > > > > How can you control them through libusb? Why not just use the tty > > device the kernel exports for you? > > Right, I got distracted writing this part. I'm only really using libusb as a > fallback for HID devices (since hidapi has support already)... > HID: NanoFury and HashBuster nano Ok, this is making no sense. As you really aren't saying exactly what you are trying to do, I'm going to just give up and say you can't do it. good luck, greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Monday, November 18, 2013 8:29:24 PM Greg KH wrote: > On Mon, Nov 18, 2013 at 08:19:38PM +, Luke-Jr wrote: > > My software does not interface with keyboards under any condition. > > > > Some specific devices I'm working with include: > > FTDI: Avalon, Butterfly Labs BitFORCE & BitFORCE SC, The Chili > > CP210X: Bitfountain Block Erupter Emerald/Sapphire > > CDC/ACM: Big Picture Mining, BTCFPGA ModMiner > > PL2303: Icarus > > How can you control them through libusb? Why not just use the tty > device the kernel exports for you? Right, I got distracted writing this part. I'm only really using libusb as a fallback for HID devices (since hidapi has support already)... HID: NanoFury and HashBuster nano > > > And why not just use the kernel interfaces to the device? That way you > > > know it works properly with anything else that has opened it at the > > > time. > > > > I do, when permissions are set such that I can. Unfortunately, the > > current Gentoo-stable eudev does not give any groups access to hidraw > > devices, so it fails over to libusb. > > Then fix that by using console-kit to set the permissions of the devices > that the local user has access to. Often there is no local user on these systems. Luke -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Mon, Nov 18, 2013 at 08:19:38PM +, Luke-Jr wrote: > My software does not interface with keyboards under any condition. > > Some specific devices I'm working with include: > FTDI: Avalon, Butterfly Labs BitFORCE & BitFORCE SC, The Chili > CP210X: Bitfountain Block Erupter Emerald/Sapphire > CDC/ACM: Big Picture Mining, BTCFPGA ModMiner > PL2303: Icarus How can you control them through libusb? Why not just use the tty device the kernel exports for you? > > And why not just use the kernel interfaces to the device? That way you > > know it works properly with anything else that has opened it at the > > time. > > I do, when permissions are set such that I can. Unfortunately, the current > Gentoo-stable eudev does not give any groups access to hidraw devices, so it > fails over to libusb. Then fix that by using console-kit to set the permissions of the devices that the local user has access to. best of luck, greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Monday, November 18, 2013 1:37:23 PM Greg KH wrote: > On Mon, Nov 18, 2013 at 07:10:45AM +, Luke-Jr wrote: > > On Monday, November 18, 2013 6:07:37 AM Greg KH wrote: > > > > I want my software to ignore devices that are already in use by other > > > > software. The "other software" might be accessing it via the kernel > > > > drivers, or perhaps libusb. My software (in this case) would be using > > > > libusb, and need to detach the kernel drivers to claim the interface > > > > - even if the kernel driver isn't currently in use. > > > > > > What types of devices? > > > > Just tty/serial devices (HID, VCOM, and CDC/ACM) > > HID devices are not tty/serial devices, so I'm confused as to what you > want to do with them. Ok, but they are still opened and closed. > > > And no, there is no "general" way to determine this, sorry, it all > > > depends on the type of device, and for some of them, you really can't > > > tell (think about a keyboard, when is it "in use"?) > > > > When X has it open? > > X opens all input devices all the time, it never closes them. Yes, I expect a keyboard would always be considered in use when X is running - perhaps even when it isn't, if it's used by virtual consoles. > > > > So, I want to detach the driver so I can use it myself, but only if > > > > some other software isn't already using it. > > > > > > What will you do with the device? > > > > Send it work to process, and get results back. > > What type of work can you send a keyboard? My software does not interface with keyboards under any condition. Some specific devices I'm working with include: FTDI: Avalon, Butterfly Labs BitFORCE & BitFORCE SC, The Chili CP210X: Bitfountain Block Erupter Emerald/Sapphire CDC/ACM: Big Picture Mining, BTCFPGA ModMiner PL2303: Icarus > And why not just use the kernel interfaces to the device? That way you > know it works properly with anything else that has opened it at the > time. I do, when permissions are set such that I can. Unfortunately, the current Gentoo-stable eudev does not give any groups access to hidraw devices, so it fails over to libusb. Luke -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Mon, 18 Nov 2013, Luke-Jr wrote: > Is there a way to tell if a USB device is in use or not before requesting the > kernel detach drivers? > > I'd define "in use" as either an interface claimed on usbfs or tty/etc > provided by a kernel driver being opened. And "not in use" including "kernel > driver attached, but not interfacing device to any software". There is a form of the detach-driver call in usbfs that will not do anything if the current driver is usbfs. But there is no way to tell if a tty or other device file is open, or if the device is interfaced to any software. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Mon, Nov 18, 2013 at 07:10:45AM +, Luke-Jr wrote: > On Monday, November 18, 2013 6:07:37 AM Greg KH wrote: > > > I want my software to ignore devices that are already in use by other > > > software. The "other software" might be accessing it via the kernel > > > drivers, or perhaps libusb. My software (in this case) would be using > > > libusb, and need to detach the kernel drivers to claim the interface - > > > even if the kernel driver isn't currently in use. > > > > What types of devices? > > Just tty/serial devices (HID, VCOM, and CDC/ACM) HID devices are not tty/serial devices, so I'm confused as to what you want to do with them. > > And no, there is no "general" way to determine this, sorry, it all > > depends on the type of device, and for some of them, you really can't > > tell (think about a keyboard, when is it "in use"?) > > When X has it open? X opens all input devices all the time, it never closes them. > > > So, I want to detach the driver so I can use it myself, but only if some > > > other software isn't already using it. > > > > What will you do with the device? > > Send it work to process, and get results back. What type of work can you send a keyboard? And why not just use the kernel interfaces to the device? That way you know it works properly with anything else that has opened it at the time. greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Sun, 2013-11-17 at 20:40 -0800, Greg KH wrote: > On Mon, Nov 18, 2013 at 01:44:19AM +, Luke-Jr wrote: > > Is there a way to tell if a USB device is in use or not before requesting > > the > > kernel detach drivers? > > Not really, sorry. Nor should there be. The concept is a race condition by design. If you really need this, you'll need to implement a "yield unused device" interface to the kernel. > > I'd define "in use" as either an interface claimed on usbfs or tty/etc > > provided by a kernel driver being opened. And "not in use" including > > "kernel > > driver attached, but not interfacing device to any software". > > > > Thoughts? > > You might want to find a "way", but think about a user removing a device > at any point in time. Drivers have to handle this already, so it really > doesn't make sense to have this type of functionality if it can't be > used by anyone. Well no. The drivers certainly must handle it. That doesn't mean you do no damage in user space. To do it cleanly you'd add per device counters. and perhaps a method to the driver interface. Regards Oliver -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Monday, November 18, 2013 6:07:37 AM Greg KH wrote: > > I want my software to ignore devices that are already in use by other > > software. The "other software" might be accessing it via the kernel > > drivers, or perhaps libusb. My software (in this case) would be using > > libusb, and need to detach the kernel drivers to claim the interface - > > even if the kernel driver isn't currently in use. > > What types of devices? Just tty/serial devices (HID, VCOM, and CDC/ACM) > And no, there is no "general" way to determine this, sorry, it all > depends on the type of device, and for some of them, you really can't > tell (think about a keyboard, when is it "in use"?) When X has it open? > > So, I want to detach the driver so I can use it myself, but only if some > > other software isn't already using it. > > What will you do with the device? Send it work to process, and get results back. > Why not just have the user tell you which device you want to claim? They can do that, but users like automatic stuff by default :/ Luke -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Mon, Nov 18, 2013 at 05:35:53AM +, Luke-Jr wrote: > On Monday, November 18, 2013 4:40:55 AM Greg KH wrote: > > On Mon, Nov 18, 2013 at 01:44:19AM +, Luke-Jr wrote: > > > Is there a way to tell if a USB device is in use or not before requesting > > > the kernel detach drivers? > > > > Not really, sorry. > > > > > I'd define "in use" as either an interface claimed on usbfs or tty/etc > > > provided by a kernel driver being opened. And "not in use" including > > > "kernel driver attached, but not interfacing device to any software". > > > > > > Thoughts? > > > > You might want to find a "way", but think about a user removing a device > > at any point in time. Drivers have to handle this already, so it really > > doesn't make sense to have this type of functionality if it can't be > > used by anyone. > > > > What problem are you trying to solve? > > I want my software to ignore devices that are already in use by other > software. The "other software" might be accessing it via the kernel drivers, > or perhaps libusb. My software (in this case) would be using libusb, and need > to detach the kernel drivers to claim the interface - even if the kernel > driver isn't currently in use. What types of devices? And no, there is no "general" way to determine this, sorry, it all depends on the type of device, and for some of them, you really can't tell (think about a keyboard, when is it "in use"?) > So, I want to detach the driver so I can use it myself, but only if some > other > software isn't already using it. What will you do with the device? Why not just have the user tell you which device you want to claim? greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Monday, November 18, 2013 4:40:55 AM Greg KH wrote: > On Mon, Nov 18, 2013 at 01:44:19AM +, Luke-Jr wrote: > > Is there a way to tell if a USB device is in use or not before requesting > > the kernel detach drivers? > > Not really, sorry. > > > I'd define "in use" as either an interface claimed on usbfs or tty/etc > > provided by a kernel driver being opened. And "not in use" including > > "kernel driver attached, but not interfacing device to any software". > > > > Thoughts? > > You might want to find a "way", but think about a user removing a device > at any point in time. Drivers have to handle this already, so it really > doesn't make sense to have this type of functionality if it can't be > used by anyone. > > What problem are you trying to solve? I want my software to ignore devices that are already in use by other software. The "other software" might be accessing it via the kernel drivers, or perhaps libusb. My software (in this case) would be using libusb, and need to detach the kernel drivers to claim the interface - even if the kernel driver isn't currently in use. So, I want to detach the driver so I can use it myself, but only if some other software isn't already using it. Luke -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Check device use before detaching?
On Mon, Nov 18, 2013 at 01:44:19AM +, Luke-Jr wrote: > Is there a way to tell if a USB device is in use or not before requesting the > kernel detach drivers? Not really, sorry. > I'd define "in use" as either an interface claimed on usbfs or tty/etc > provided by a kernel driver being opened. And "not in use" including "kernel > driver attached, but not interfacing device to any software". > > Thoughts? You might want to find a "way", but think about a user removing a device at any point in time. Drivers have to handle this already, so it really doesn't make sense to have this type of functionality if it can't be used by anyone. What problem are you trying to solve? greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
