[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <56A57D0C.6030308@collabora.co.uk>
Date: Sun, 24 Jan 2016 22:40:28 -0300
From: Emilio López <emilio.lopez@...labora.co.uk>
To: Bjørn Mork <bjorn@...k.no>
Cc: gregkh@...uxfoundation.org, stern@...land.harvard.edu,
kborer@...il.com, k.opasiak@...sung.com, reillyg@...omium.org,
keescook@...omium.org, linux-api@...r.kernel.org,
linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org,
jorgelo@...omium.org, dan.carpenter@...cle.com
Subject: Re: [PATCH v2] usb: devio: Add ioctl to disallow detaching kernel USB
drivers.
Hi Bjørn,
El 22/01/16 a las 06:41, Bjørn Mork escribió:
> Emilio López <emilio.lopez@...labora.co.uk> writes:
>
>> diff --git a/drivers/usb/core/devio.c b/drivers/usb/core/devio.c
>> index 38ae877c..bf40aa6 100644
>> --- a/drivers/usb/core/devio.c
>> +++ b/drivers/usb/core/devio.c
>> @@ -77,6 +77,8 @@ struct usb_dev_state {
>> unsigned long ifclaimed;
>> u32 secid;
>> u32 disabled_bulk_eps;
>> + bool privileges_dropped;
>> + unsigned long interface_allowed_mask;
>> };
>>
>> struct async {
>> @@ -641,6 +643,14 @@ static int claimintf(struct usb_dev_state *ps, unsigned int ifnum)
>> if (test_bit(ifnum, &ps->ifclaimed))
>> return 0;
>>
>> + if (ps->privileges_dropped) {
>> + if (ifnum >= 8*sizeof(ps->interface_allowed_mask))
>> + return -EINVAL;
>
>
> I don't think you need this runtime test. You can just make sure that
> sizeof(ps->interface_allowed_mask) == sizeof(ps->ifclaimed) at build
> time.
>
> I do find this variable and arbitrary limit a bit confusing, but that's
> not your fault - I guess it is an indication that ifnums > 31 are rare
> :)
>
>
>> diff --git a/include/uapi/linux/usbdevice_fs.h b/include/uapi/linux/usbdevice_fs.h
>> index 019ba1e..9abcb34 100644
>> --- a/include/uapi/linux/usbdevice_fs.h
>> +++ b/include/uapi/linux/usbdevice_fs.h
>> @@ -154,6 +154,10 @@ struct usbdevfs_streams {
>> unsigned char eps[0];
>> };
>>
>> +struct usbdevfs_drop_privs {
>> + unsigned long interface_allowed_mask;
>> +};
>> +
>
> "unsigned long" isn't a very good choice here, is it?
I went with a type matching ifclaimed on struct usb_dev_state to keep
the limit the same, but I guess it's not the best idea for an ioctl. I
can switch it to __u32, keeping the runtime check above as is, or use
__u64. Which one would you prefer?
Thanks for the review!
Emilio
Powered by blists - more mailing lists