Re: Debugging errors with Dell XPS 9560 TPM

[Alex Guzman <alex@guzman.io>]

On Fri, 2020-02-14 at 16:04 -0500, James Bottomley wrote:
> On Fri, 2020-02-14 at 14:02 -0700, Jerry Snitselaar wrote:
> > On Fri Feb 14 20, James Bottomley wrote:
> > > On Fri, 2020-02-14 at 10:32 -0800, Alex Guzman wrote:
> > > > Looks like someone had a look on the bug tracker
> > > > (https://bugzilla.kernel.org/show_bug.cgi?id=206275#c6)
> > > > and they figure it's definitely a regression in the kernel and
> > > > should
> > > > be reverted or rectified. They advised me to come ping here
> > > > once
> > > > more.
> > > 
> > > Reading the bugzilla, I don't get *what* needs to be
> > > reverted.  The
> > > commit 4d6ebc4c4950595414722dfadd0b361f5a05d37e isn't present in
> > > upstream, so what kernel is it present in, or what is the full
> > > commit message so we can find the upstream commit?
> > > 
> > > James
> > > 
> > > 
> > > > - Alex
> > > > 
> > > > On Sat, Feb 1, 2020 at 4:19 PM Alex Guzman <alex@guzman.io>
> > > > wrote:
> > > > > Hey there! I reported a bug on the bug tracker a bit ago but
> > > > > haven't seen any movement, so I figured I'd drop in here. My
> > > > > XPS 9560 has been having issues with its TPM, and all
> > > > > commands
> > > > > will fail with error 1 when operating on the TPM device. I
> > > > > managed to bisect it back to commit
> > > > > 4d6ebc4c4950595414722dfadd0b361f5a05d37e (tpm: fix
> > > > > invalid locking in NONBLOCKING mode) though it began failing
> > > > > with error 14 (bad address) at that point.
> > > > > 
> > > > > I reported the bug at
> > > > > https://bugzilla.kernel.org/show_bug.cgi?id=206275 and
> > > > > attached
> > > > > some dmesg logs from boot there. Does anyone have any
> > > > > suggestions for additional debugging or such to figure out
> > > > > what's happening here?
> > > > > 
> > > > > - Alex
> > 
> > d23d12484307 | 2019-12-17 | tpm: fix invalid locking in NONBLOCKING
> > mode (Tadeusz Struk)
> > 
> > There is a commit that is a fix to this commit:
> > 
> > a430e67d9a2c | 2020-01-08 | tpm: Handle negative priv->response_len
> > in tpm_common_read() (Tadeusz Struk)
> 
> Yes, I suspected it might be that ... in which case upstream should
> have the fix, can we verify that 5.6-rc1 works just fine?
> 
> James
> 

I just tested with 5.6_rc1. The behavior is still present:


ERROR:tcti:src/tss2-tcti/tcti-device.c:290:tcti_device_receive() Failed
to read response from fd 3, got errno 1: Operation not permitted 
ERROR:esys:src/tss2-
esys/api/Esys_GetCapability.c:307:Esys_GetCapability_Finish() Received
a non-TPM Error 
ERROR:esys:src/tss2-
esys/api/Esys_GetCapability.c:107:Esys_GetCapability() Esys Finish
ErrorCode (0x000a000a) 
ERROR: Esys_GetCapability(0xA000A) - tcti:IO failure
ERROR: Unable to run tpm2_getcap