From: Udipto Goswami <quic_ugoswami@quicinc.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Felipe Balbi <balbi@kernel.org>, John Keeping <john@metanate.com>,
<linux-usb@vger.kernel.org>,
Pratham Pratap <quic_ppratap@quicinc.com>,
Pavankumar Kondeti <quic_pkondeti@quicinc.com>,
Jack Pham <quic_jackp@quicinc.com>
Subject: Re: [PATCH v8] usb: f_fs: Fix use-after-free for epfile
Date: Tue, 11 Jan 2022 17:16:43 +0530 [thread overview]
Message-ID: <95a13a8d-44b8-bb38-1485-d9cc623ef1aa@quicinc.com> (raw)
In-Reply-To: <YdwtUkq4hTT2oxh0@kroah.com>
Hi Greg,
On 10-01-2022 06:27 pm, Greg Kroah-Hartman wrote:
> On Mon, Jan 10, 2022 at 06:22:48PM +0530, Udipto Goswami wrote:
>> Hi Greg,
>>
>> On 07-01-2022 03:34 pm, Greg Kroah-Hartman wrote:
>>> On Fri, Jan 07, 2022 at 01:52:05PM +0530, Udipto Goswami wrote:
>>>> Hi Greg,
>>>>
>>>> On 06-01-2022 08:03 pm, Greg Kroah-Hartman wrote:
>>>>> On Wed, Jan 05, 2022 at 07:35:26PM +0530, Udipto Goswami wrote:
>>>>>> Consider a case where ffs_func_eps_disable is called from
>>>>>> ffs_func_disable as part of composition switch and at the
>>>>>> same time ffs_epfile_release get called from userspace.
>>>>>> ffs_epfile_release will free up the read buffer and call
>>>>>> ffs_data_closed which in turn destroys ffs->epfiles and
>>>>>> mark it as NULL. While this was happening the driver has
>>>>>> already initialized the local epfile in ffs_func_eps_disable
>>>>>> which is now freed and waiting to acquire the spinlock. Once
>>>>>> spinlock is acquired the driver proceeds with the stale value
>>>>>> of epfile and tries to free the already freed read buffer
>>>>>> causing use-after-free.
>>>>>>
>>>>>> Following is the illustration of the race:
>>>>>>
>>>>>> CPU1 CPU2
>>>>>>
>>>>>> ffs_func_eps_disable
>>>>>> epfiles (local copy)
>>>>>> ffs_epfile_release
>>>>>> ffs_data_closed
>>>>>> if (last file closed)
>>>>>> ffs_data_reset
>>>>>> ffs_data_clear
>>>>>> ffs_epfiles_destroy
>>>>>> spin_lock
>>>>>> dereference epfiles
>>>>>>
>>>>>> Fix this races by taking epfiles local copy & assigning it under
>>>>>> spinlock and if epfiles(local) is null then update it in ffs->epfiles
>>>>>> then finally destroy it.
>>>>>> Extending the scope further from the race, protecting the ep related
>>>>>> structures, and concurrent accesses.
>>>>>>
>>>>>> Fixes: a9e6f83c2df (usb: gadget: f_fs: stop sleeping in
>>>>>> ffs_func_eps_disable)
>>>>> No need to line-wrap this line, the scripts will complain about it :(
>>>> checkpatch didn't give any error for this though.
>>> It's not a checkpatch error, it will show up in the scripts we run on
>>> commits in our trees. linux-next will report a problem with this, and
>>> so I have the same scripts as well. They were posted to the
>>> users@kernel.org mailing list a year or so ago if you are curious. It's
>>> not a big deal.
>>>
>>>>>> Reviewed-by: John Keeping <john@metanate.com>
>>>>>> Signed-off-by: Pratham Pratap <quic_ppratap@quicinc.com>
>>>>>> Co-developed-by: Udipto Goswami <quic_ugoswami@quicinc.com>
>>>>>> Signed-off-by: Udipto Goswami <quic_ugoswami@quicinc.com>
>>>>>> ---
>>>>>> v8: Fixed compilation errors from previous version.
>>>>>>
>>>>>> drivers/usb/gadget/function/f_fs.c | 60 ++++++++++++++++++++++++++++----------
>>>>>> 1 file changed, 45 insertions(+), 15 deletions(-)
>>>>>>
>>>>>> diff --git a/drivers/usb/gadget/function/f_fs.c b/drivers/usb/gadget/function/f_fs.c
>>>>>> index 3c584da..541a4af 100644
>>>>>> --- a/drivers/usb/gadget/function/f_fs.c
>>>>>> +++ b/drivers/usb/gadget/function/f_fs.c
>>>>>> @@ -1711,16 +1711,24 @@ static void ffs_data_put(struct ffs_data *ffs)
>>>>>> static void ffs_data_closed(struct ffs_data *ffs)
>>>>>> {
>>>>>> + struct ffs_epfile *epfiles;
>>>>>> + unsigned long flags;
>>>>>> +
>>>>>> ENTER();
>>>>>> if (atomic_dec_and_test(&ffs->opened)) {
>>>>>> if (ffs->no_disconnect) {
>>>>>> ffs->state = FFS_DEACTIVATED;
>>>>>> - if (ffs->epfiles) {
>>>>>> - ffs_epfiles_destroy(ffs->epfiles,
>>>>>> - ffs->eps_count);
>>>>>> - ffs->epfiles = NULL;
>>>>>> - }
>>>>>> + spin_lock_irqsave(&ffs->eps_lock, flags);
>>>>>> + epfiles = ffs->epfiles;
>>>>>> + ffs->epfiles = NULL;
>>>>>> + spin_unlock_irqrestore(&ffs->eps_lock,
>>>>>> + flags);
>>>>>> +
>>>>>> + if (epfiles)
>>>>>> + ffs_epfiles_destroy(epfiles,
>>>>>> + ffs->eps_count);
>>>>>> +
>>>>>> if (ffs->setup_state == FFS_SETUP_PENDING)
>>>>>> __ffs_ep0_stall(ffs);
>>>>>> } else {
>>>>>> @@ -1767,14 +1775,27 @@ static struct ffs_data *ffs_data_new(const char *dev_name)
>>>>>> static void ffs_data_clear(struct ffs_data *ffs)
>>>>>> {
>>>>>> + struct ffs_epfile *epfiles;
>>>>>> + unsigned long flags;
>>>>>> +
>>>>>> ENTER();
>>>>>> ffs_closed(ffs);
>>>>>> BUG_ON(ffs->gadget);
>>>>>> - if (ffs->epfiles)
>>>>>> - ffs_epfiles_destroy(ffs->epfiles, ffs->eps_count);
>>>>>> + spin_lock_irqsave(&ffs->eps_lock, flags);
>>>>>> + epfiles = ffs->epfiles;
>>>>>> + ffs->epfiles = NULL;
>>>>>> + spin_unlock_irqrestore(&ffs->eps_lock, flags);
>>>>>> +
>>>>>> + /*
>>>>>> + * potential race possible between ffs_func_eps_disable
>>>>>> + * & ffs_epfile_release therefore maintaining a local
>>>>>> + * copy of epfile will save us from use-after-free.
>>>>>> + */
>>>>>> + if (epfiles)
>>>>>> + ffs_epfiles_destroy(epfiles, ffs->eps_count);
>>>>>> if (ffs->ffs_eventfd)
>>>>>> eventfd_ctx_put(ffs->ffs_eventfd);
>>>>>> @@ -1790,7 +1811,6 @@ static void ffs_data_reset(struct ffs_data *ffs)
>>>>>> ffs_data_clear(ffs);
>>>>>> - ffs->epfiles = NULL;
>>>>>> ffs->raw_descs_data = NULL;
>>>>>> ffs->raw_descs = NULL;
>>>>>> ffs->raw_strings = NULL;
>>>>>> @@ -1870,6 +1890,7 @@ static int ffs_epfiles_create(struct ffs_data *ffs)
>>>>>> {
>>>>>> struct ffs_epfile *epfile, *epfiles;
>>>>>> unsigned i, count;
>>>>>> + unsigned long flags;
>>>>>> ENTER();
>>>>>> @@ -1895,7 +1916,9 @@ static int ffs_epfiles_create(struct ffs_data *ffs)
>>>>>> }
>>>>>> }
>>>>>> + spin_lock_irqsave(&ffs->eps_lock, flags);
>>>>>> ffs->epfiles = epfiles;
>>>>>> + spin_unlock_irqrestore(&ffs->eps_lock, flags);
>>>>> Why is this lock needed when you set this value? What is that
>>>>> protecting?
>>>> Was making it uniform, protection ffs->epfiles all over. Here intention is
>>>> to protect the operation of epfiles getting assigned to ffs->epfiles so that
>>>> we protect the ffs->epfiles instance at the time of creation as well.
>>> But it is not needed here, so no need to add it.
>> we haven't encountered this race although, if ffs_func_set_alt() which is
>> asynchronously scheduled ffs_reset_work() ->
>> ffs_data_reset()->ffs_data_clear() this can take place while user space just
>> opened ep0.
>>
>> in this case the epfiles_destroy() if took place just after this operation
>> of ffs->epfiles = epfiles, the epfiles_create, the ffs->epfiles will be
>> freed. It will be better if we protect this case as well.
> What exactly are you protecting here? You are only ever setting the
> value, never looking at the result. If you were looking to see if this
> was a value before setting it, it would make more sense, but as-is, I
> can't figure out what this lock here is protecting.
Yah you are right on this. was trying to protect ffs->epfiles only but
it will be not needed here. Will push another version.
>
> confused,
>
> greg k-h
prev parent reply other threads:[~2022-01-11 11:46 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-01-05 14:05 [PATCH v8] usb: f_fs: Fix use-after-free for epfile Udipto Goswami
2022-01-06 14:33 ` Greg Kroah-Hartman
2022-01-07 8:22 ` Udipto Goswami
2022-01-07 10:04 ` Greg Kroah-Hartman
2022-01-10 12:52 ` Udipto Goswami
2022-01-10 12:57 ` Greg Kroah-Hartman
2022-01-11 11:46 ` Udipto Goswami [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=95a13a8d-44b8-bb38-1485-d9cc623ef1aa@quicinc.com \
--to=quic_ugoswami@quicinc.com \
--cc=balbi@kernel.org \
--cc=gregkh@linuxfoundation.org \
--cc=john@metanate.com \
--cc=linux-usb@vger.kernel.org \
--cc=quic_jackp@quicinc.com \
--cc=quic_pkondeti@quicinc.com \
--cc=quic_ppratap@quicinc.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).