Re: [PATCH v7 08/10] tpm: Proxy driver for supporting multiple emulated TPMs

On Sat, Mar 12, 2016 at 06:27:13PM -0500, Stefan Berger wrote:

   Jarkko Sakkinen [off-list ref] wrote on 03/12/2016
   01:51:54 PM:

   >
   > On Fri, Mar 11, 2016 at 09:51:03PM -0500, Stefan Berger wrote:
   > > This patch implements a proxy driver for supporting multiple emulated
   TPMs
   > > in a system.
   > >
   > > The driver implements a device /dev/vtpmx that is used to created
   > > a client device pair /dev/tpmX (e.g., /dev/tpm10) and a server side
   that
   > > is accessed using a file descriptor returned by an ioctl.
   > > The device /dev/tpmX is the usual TPM device created by the core TPM
   > > driver. Applications or kernel subsystems can send TPM commands to it
   > > and the corresponding server-side file descriptor receives these
   > > commands and delivers them to an emulated TPM.
   >
   > With my test script [1] running on QEMU and TPM 2.0 simulator running on
   the
   > host side I get this:
   >
   > $ python tpm2-simulator-vtpm --host=10.0.2.2
   > cmd
   > 80 01 00 00 00 0c 00 00 01 44 00 00
   > rsp
   > 80 01 00 00 00 0a 00 00 00 00
   > Traceback (most recent call last):
   >   File "tpm2-simulator-vtpm", line 85, in <module>
   >     main()
   >   File "tpm2-simulator-vtpm", line 80, in main
   >     resp = client.send_cmd(stream)
   >   File "/home/tpmdd/tpm2-scripts/tpm2.py", line 454, in send_cmd
   >     rsp = self.simulator.send_cmd(cmd)
   >   File "/home/tpmdd/tpm2-scripts/tpm2.py", line 436, in send_cmd
   >     raise SimulatorError("Empty response")
   > tpm2.SimulatorError: Empty response
   >
   > However, the process does not exit before the read call expires:
   >
   > $ python tpm2-list-handles
   > Traceback (most recent call last):
   >   File "tpm2-list-handles", line 61, in <module>
   >     main()
   >   File "tpm2-list-handles", line 51, in main
   >     handles += client.get_cap(tpm2.TPM2_CAP_HANDLES,
   tpm2.HR_LOADED_SESSION)
   >   File "/home/tpmdd/tpm2-scripts/tpm2.py", line 782, in get_cap
   >     next_handles, more_data = self.__get_cap_cnt(cap, pt, 1)
   >   File "/home/tpmdd/tpm2-scripts/tpm2.py", line 766, in __get_cap_cnt
   >     rsp = self.send_cmd(cmd)[10:]
   >   File "/home/tpmdd/tpm2-scripts/tpm2.py", line 458, in send_cmd
   >     rsp = f.read()
   > IOError: [Errno 62] Timer expired
   >
   > The server side stays stuck unti this happens.

   This fix should solve the problem:

   diff --git a/drivers/char/tpm/tpm_vtpm_proxy.c
   b/drivers/char/tpm/tpm_vtpm_proxy.c
   index d73944e..01e5070 100644
   --- a/drivers/char/tpm/tpm_vtpm_proxy.c
   +++ b/drivers/char/tpm/tpm_vtpm_proxy.c
   @@ -456,10 +456,10 @@ err_delete_proxy_dev:
     */
    static void vtpm_proxy_delete_device(struct proxy_dev *proxy_dev)
    {
   -       tpm_chip_unregister(proxy_dev->chip);
   -
           vtpm_proxy_fops_undo_open(proxy_dev);

   +      tpm_chip_unregister(proxy_dev->chip);
   +
           vtpm_proxy_delete_proxy_dev(proxy_dev);
    }

   Can you let me know whether this gets it working for you? I'd prepare a
   v9.

So is the deadlock such that:

* tpm_chip_unregister() tries to write lock ops_sem.
* tpm_transmit() holds read lock to ops_sem.

This takes two minutes if no timeouts are calculated.

And is the effect of moving vtpm_proxy_fops_undo_open() upwards such
that vtpm_proxy_tpm_req_canceled() starts returning true, which in
effect breaks the loop in tpm_transmit()?

Yeah, the fix as a code change is very simple but I had to use perf
probe to verify this so maybe a comment there would be in place to
tell why tpm_chip_unregister() must be called last (write lock).

      Stefan

/Jarkko

------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785111&iu=/4140