Re: [linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Oliver Neukum
Am Freitag, 1. Juli 2005 16:28 schrieb Duncan Sands: > > > but do we want to guarantee that the altsetting is always zero > > > when probe is called (likewise for configuration)?  After all, drivers > > > can be responsible for setting the values they want. > > > > I think we do.  Drivers should f

Re: [linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Duncan Sands
Hi Alan, > ? You sound confused. seems so :) > Configurations aren't changed as easily as altsettings. They can't be > changed at all by drivers, normally. Remember, each configuration > contains a set of interfaces, and drivers bind to interfaces. Change the > config and you've got a com

Re: [linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Alan Stern
On Fri, 1 Jul 2005, Duncan Sands wrote: > > > but do we want to guarantee that the altsetting is always zero > > > when probe is called (likewise for configuration)? After all, drivers > > > can be responsible for setting the values they want. > > > > I think we do. Drivers should find their de

Re: [linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Duncan Sands
> > but do we want to guarantee that the altsetting is always zero > > when probe is called (likewise for configuration)? After all, drivers > > can be responsible for setting the values they want. > > I think we do. Drivers should find their devices initially in the > default state. That's wh

Re: [linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Alan Stern
On Fri, 1 Jul 2005, Duncan Sands wrote: > Hi Alan, > > > > Suppose my driver's probe method is called for a device, > > > some initialisation is done (eg: I change the altsetting), > > > but finally the probe fails. Will the device state be > > > reset in any way before other drivers are probed?

Re: [linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Duncan Sands
Hi Alan, > > Suppose my driver's probe method is called for a device, > > some initialisation is done (eg: I change the altsetting), > > but finally the probe fails. Will the device state be > > reset in any way before other drivers are probed? > > No. Do you think selecting altsetting 0 will r

Re: [linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Alan Stern
On Fri, 1 Jul 2005, Duncan Sands wrote: > Suppose my driver's probe method is called for a device, > some initialisation is done (eg: I change the altsetting), > but finally the probe fails. Will the device state be > reset in any way before other drivers are probed? No. Do you think selecting

[linux-usb-devel] resetting altsetting on probe failure

2005-07-01 Thread Duncan Sands
Suppose my driver's probe method is called for a device, some initialisation is done (eg: I change the altsetting), but finally the probe fails. Will the device state be reset in any way before other drivers are probed? Thanks a lot, Duncan.