Hi, On Thu, Jul 02, 2015 at 01:58:44PM +0200, Gregory CLEMENT wrote: > Hi Felipe, > > thanks for you prompt feedback. > > On 02/07/2015 13:45, Felipe Balbi wrote: > > Hi, > > > > On Thu, Jul 02, 2015 at 12:51:54PM +0200, Gregory CLEMENT wrote: > >> Hello, > >> > >> When I use configs to configure the mass storage function for the > >> gadget, and when the device is plugged under Windows, then the > >> partition that I expose is well managed, but I also see 7 other gadget > >> in the device manager with error. > >> > >> This seven bogus gadget seems to be the 7 other LUN that are not > >> used. Indeed if I apply this dirty patch: > >> > >> diff --git a/drivers/usb/gadget/function/f_mass_storage.c b/drivers/usb/gadget/function/f_mass_storage.c > >> index 3cc109f..2b4ae98 100644 > >> --- a/drivers/usb/gadget/function/f_mass_storage.c > >> +++ b/drivers/usb/gadget/function/f_mass_storage.c > >> @@ -3511,7 +3511,8 @@ static struct usb_function_instance *fsg_alloc_inst(void) > >> rc = PTR_ERR(opts->common); > >> goto release_opts; > >> } > >> - rc = fsg_common_set_nluns(opts->common, FSG_MAX_LUNS); > >> +// rc = fsg_common_set_nluns(opts->common, FSG_MAX_LUNS); > >> + rc = fsg_common_set_nluns(opts->common, 1); > >> if (rc) > >> goto release_opts; > >> > >> Then there is no more gadget error under Windows. As the value of > >> FSG_MAX_LUNS is 8 and in my configuration I only use one partition, > >> then it makes sens that I see 7 bogus gadgets. > >> > >> I also saw that in the legacy driver, it was possible to modify the > >> number of LUN using the module parameter file_count. > > > > This has been reported. Michal was working on a fix, but the patch > > hasn't been applied yet. > > Oh sorry to not have finding the thread about it. > > > > >> There is no such parameter when using configfs. > >> > >> What would be the correct way to fix it? > > > > Making sure you don't lie to the other side of the cable :-) > > > >> I thought about enumerate how many LUN where configured through > >> configfs and then setting the exact number using fsg_common_set_nluns. > > > > yeah, that should be the way. > > > >> There is no problem at all under Linux so maybe it is jut the way how > >> the gadget is exposed through USB. > > > > no, we shouldn't really lie, this really needs fixing. > > Ok so don't hesitate to ask me to test or if you need any help. sure thing, here's the thread if you want to chip in: http://www.spinics.net/lists/linux-usb/msg126284.html -- balbi
Attachment:
signature.asc
Description: Digital signature
