Re: [PATCH] usb: gadget: f_fs: Fix kernel panic for SuperSpeed

[Jim Lin]

On 2016å04æ28æ 20:21, Felipe Balbi wrote:
> > I also attach git log of system/core/adb/usb_linux_client.cpp of Android
> > N for your reference.
> > "
> > Author: Badhri Jagan Sridharan <Badhri@xxxxxxxxxx>
> > Date:   Mon Oct 5 13:04:03 2015 -0700
> >
> >       adbd: Add os descriptor support for adb.
> >
> >       Eventhough windows does not rely on extended os
> >       descriptor for adbd, when android usb device is
> >       configures as a composite device such as mtp+adb,
> >       windows discards the extended os descriptor even
> >       if one of the USB function fails to send
> >       the extended compat descriptor. This results in automatic
> >       install of MTP driverto fail when Android device is in
> >       "File Transfer" mode with adb enabled.
> >
> > https://msdn.microsoft.com/en-us/library/windows/hardware/gg463179.aspx
> > "
> Okay, cool. Can you check that you're limitting your controller's speed
> to high-speed ?
Let's focus on original patch.
Could you help to explain why we need below d->Reserved1 checking?
Now the question is that

https://msdn.microsoft.com/en-us/library/windows/hardware/gg463179.aspx

Page 7 of OS_Desc_CompatID.doc
defines reserved field to be 1 and
below code will think that os_desc is invalid because d->Reserved1 is 1.


In f_fs.c
"
static int __ffs_data_do_os_desc(enum ffs_os_desc_type type,
                 struct usb_os_desc_header *h, void *data,
                 unsigned len, void *priv)
{
    struct ffs_data *ffs = priv;
    u8 length;

    ENTER();

    switch (type) {
    case FFS_OS_DESC_EXT_COMPAT: {
        struct usb_ext_compat_desc *d = data;
        int i;

        if (len < sizeof(*d) ||
            d->bFirstInterfaceNumber >= ffs->interfaces_count ||
            d->Reserved1)
            return -EINVAL;
"

--nvpublic