On Thu, Feb 24, 2011 at 8:41 PM, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote: > On Thu, 24 Feb 2011, Michal Nazarewicz wrote: > >> > On Tue, Feb 22, 2011 at 8:49 PM, Alan Stern wrote: >> >> In principle, function drivers -- including f_mass_storage -- >> >> shouldn't depend on req->length being set to anything. A little >> >> auditing would be a good idea. >> >> On Thu, 24 Feb 2011 12:58:22 +0100, Maulik Mankad >> <mankad.maulik@xxxxxxxxx> wrote: >> > I didn't quite understand your last point. >> > >> > This test works for g_file_storage gadget since "fsg->ep0req->length = >> > 0" in fsg_setup() function in file_storage.c. >> > >> > Does setting "req->length= 0" in composite_setup() in f_mass_storage.c >> > have any unwanted side effect? >> > >> > Should this change be done in the f_mass_storage.c under "case >> > FSG_STATE_RESET" just before ep0_queue() is called (since we know that >> > RESET request does not have a data phase)? >> >> Alan point is that your patch is correct. Composite functions should not >> depend on value of req->length, so if your patch breaks some functions >> its the function that need to be fixed. What could be handy is trying >> to check in advance if any functions in fact depend on req->length. > > Exactly. Maulik, you determined that f_mass_storage works okay with > your patch. But what about the other function drivers? Did you check > to make sure the patch won't break any of them? > Ok, I have validated the following functions with the patch on top of 2.6.38-rc6. (1) g_multi composite gadget. Verified all the three functions(mass storage, serial, and rndis) (2) g_cdc composite gadget. (Verified serial and rndis) (3) g_serial gadget. (4) g_ether gadget. (5) g_zero gadget. The above gadgets function as expected. Regards, Maulik -- 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
