Thank you Roshan for the patch and thank you Michal for the discussion! Appearently I am not as familiar with USB as you guys are. But what I understood from the discussion is, that if I apply the patch from Roshan to f_fs.c and composite.c, the CV Test should not fail? Thanks! On Tue, Dec 24, 2013 at 5:14 PM, Michal Nazarewicz <mina86@xxxxxxxxxx> wrote: > > >>This is a bug in user space. FUNCTIONFS_DISABLE event means the host > >>has chosen a configuration with given function disabled. User-space > >>should clean up it's state, as any pending requests of any >> kind are > >>now invalid. It should not close all of the ep files. > > On Tue, Dec 24 2013, roshan.jhalani@xxxxxxxxxxxx wrote: > > Then, I would say this bug exists in both Android 4.2 and Android 4.4 > > version. But we are facing issue in Android 4.4 because kernel uses > > FunctionFs instead of f_adb.c. > > That may be the case. > > > Do you think this should be ONLY fixed in user space or we can avoid > > all ep releases in kernel? > > I think this should be fixed in user space since kernel space behaves as > it always has and as advertised[1]. Changing the behaviour would change > the ABI and risk situations that when a user space daemon crashes, > gadget's USB configuration becomes dysfunctional for the host. > > Perhaps it would make sense in some cases for the FunctionFS to hang on > even when all files are closed, but I don't think the issue you've > described justifies such a change, since it's easily fixed in user space > (a quick hack is to simply dup() ep0 and never close it). > > [1] See Documentation/usb/functionfs.txt: > > When all files are closed the function disables itself. > > Perhaps the documentation should be made clearer here though. > > -- > Best regards, _ _ > .o. | Liege of Serenely Enlightened Majesty of o' \,=./ `o > ..o | Computer Science, Michał “mina86” Nazarewicz (o o) > ooo +--<mpn@xxxxxxxxxx>--<xmpp:mina86@xxxxxxxxxx>--ooO--(_)--Ooo-- > > -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
