[Bug 1859713] [NEW] ARM v8.3a pauth not working

[Bug 1859713] [NEW] ARM v8.3a pauth not working

[Adrien Grassein]

Public bug reported:

Host: Ubuntu 19.10 - x86_64 machine
QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

ARMV8.3 pauth is not working well.

With a test code containing two pauth instructions:
    - paciasp that sign LR with A key and sp as context;
    - autiasp that verify the signature.

Test:
    - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

Expected:
    - autiasp places an invalid pointer in LR

Result:
    - autiasp successfully auth the pointer and places 0x0400660 in LR.

Further explanations:
    Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
    With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
    With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
    Values of top_bit and bottom_bit are strictly the same and it should not.

** Affects: qemu
     Importance: Undecided
         Status: New


** Tags: arm

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  New

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions

[Bug 1859713] Re: ARM v8.3a pauth not working

[Richard Henderson]

** Changed in: qemu
       Status: New => Confirmed

** Changed in: qemu
     Assignee: (unassigned) => Richard Henderson (rth)

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  Confirmed

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions

[Bug 1859713] Re: ARM v8.3a pauth not working

[Vincent Dehors]

Hi,

Here is a patch for this bug. The sbox function was using "b+=16"
instead of "b+=4".

Also, you check test vector using :

```c
    uint64_t P = 0xfb623599da6e8127ull;
    uint64_t T = 0x477d469dec0b8762ull;
    uint64_t w0 = 0x84be85ce9804e94bull;
    uint64_t k0 = 0xec2802d4e0a488e9ull;
    ARMPACKey key = { .hi = w0, .lo = k0 };
    uint64_t C5 = pauth_computepac(P, T, key);
    /* C5 should be 0xc003b93999b33765 */
```

** Patch added: "0001-target-arm-Fix-PAuth-sbox-functions.patch"
   https://bugs.launchpad.net/qemu/+bug/1859713/+attachment/5320937/+files/0001-target-arm-Fix-PAuth-sbox-functions.patch

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  Confirmed

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions

[Bug 1859713] Re: ARM v8.3a pauth not working

[Richard Henderson]

Ooof.  Good catch on the sbox error.

That said, how did you test pauth_computepac?
I still do not get the C5 result above, but 0x99d88f4472f3be39.

The following test case sets up the parameters.

** Patch added: "test case"
   https://bugs.launchpad.net/qemu/+bug/1859713/+attachment/5320967/+files/0001-tests-tcg-aarch64-Add-pauth-3.patch

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  Confirmed

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions

[Bug 1859713] Re: ARM v8.3a pauth not working

[Richard Henderson]

Oops again.  The test case has the parts of the key the wrong way around.
I'll submit the pair of patches to the mailing list.

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  Confirmed

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions

[Bug 1859713] Re: ARM v8.3a pauth not working

[Richard Henderson]

Now upstream as commit de0b1bae6461f67243282555475f88b2384a1eb9.

** Changed in: qemu
       Status: Confirmed => Fix Committed

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  Fix Committed

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions

[Bug 1859713] Re: ARM v8.3a pauth not working

[Laurent Vivier]

** Changed in: qemu
       Status: Fix Committed => Fix Released

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  Fix Released

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions

[Bug 1859713] Re: ARM v8.3a pauth not working

[Philippe Mathieu-Daudé]

Apparently this fixed bug is the official CVE-2020-10702:
https://security-tracker.debian.org/tracker/CVE-2020-10702

** CVE added: https://cve.mitre.org/cgi-bin/cvename.cgi?name=2020-10702

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1859713

Title:
  ARM v8.3a pauth not working

Status in QEMU:
  Fix Released

Bug description:
  Host: Ubuntu 19.10 - x86_64 machine
  QEMU version: 3a63b24a1bbf166e6f455fe43a6bbd8dea413d92 (master)

  ARMV8.3 pauth is not working well.

  With a test code containing two pauth instructions:
      - paciasp that sign LR with A key and sp as context;
      - autiasp that verify the signature.

  Test:
      - Run the program and corrupt LR just before autiasp (ex 0x3e00000400660 instead of 0x3e000000400664)

  Expected:
      - autiasp places an invalid pointer in LR

  Result:
      - autiasp successfully auth the pointer and places 0x0400660 in LR.

  Further explanations:
      Adding traces in qemu code shows that "pauth_computepac" is not robust enough against truncating.
      With 0x31000000400664 as input of pauth_auth, we obtain "0x55b1d65b2c138e14" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      With 0x310040008743ec as input of pauth (with same key), we obtain "0x55b1d65b2c138ef4" for PAC, "0x30" for bot_bit and "0x38" for top_bit.
      Values of top_bit and bottom_bit are strictly the same and it should not.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1859713/+subscriptions