Re: [PATCH] tpm, tpm_tis: Acquire locality in tpm_tis_gen_interrupt() and tpm_get_timeouts()

[Lukasz Majczak <lma@semihalf.com>]

Hi Laurent,

I think your case is different, as the trace shows the issue occurs in
the place already guarded by tpm_chip_start()/stop() - tpm2_probe().
Can you paste more dmesg log? (With tpm prefix?). I believe
tpm_tis_status() might return something different than 0xff here.
Please ensure you have applied Jarkko's patch that logs status
(https://patchwork.kernel.org/project/linux-integrity/patch/20210202222150.120664-1-jarkko@kernel.org/)

Best regards
Lukasz

wt., 16 lut 2021 o 12:02 Laurent Bigonville <bigon@debian.org> napisał(a):
>
> Le 16/02/21 à 09:17, Jarkko Sakkinen a écrit :
> > From: Lukasz Majczak <lma@semihalf.com>
> >
> > This is shown with Samsung Chromebook Pro (Caroline) with TPM 1.2
> > (SLB 9670):
> >
> > [    4.324298] TPM returned invalid status
> > [    4.324806] WARNING: CPU: 2 PID: 1 at drivers/char/tpm/tpm_tis_core.c:275 tpm_tis_status+0x86/0x8f
> >
> > Background
> > ==========
> >
> > TCG PC Client Platform TPM Profile (PTP) Specification, paragraph 6.1 FIFO
> > Interface Locality Usage per Register, Table 39 Register Behavior Based on
> > Locality Setting for FIFO - a read attempt to TPM_STS_x Registers returns
> > 0xFF in case of lack of locality. The described situation manifests itself
> > with the following warning trace:
> >
> > The fix
> > =======
> >
> > Add the proper decorations to tpm_tis_gen_interrupt() and
> > tpm_get_timeouts().
>
> I tried that patch (alone on the top of the HEAD of Linus master) and I
> still get the same trace in dmesg
>