On Thu Mar 7, 2024 at 2:03 AM EET, Adam Alves wrote: > My PC would hang on almost every shutdown/suspend until I started > testing this patch and so far in the past week I haven’t experienced > any problems anymore. > > I suspect that the root cause on my specific board is that after the > ACPI command to put the device to S3 or S5, some firmware > application/driver will try to use the TPM chip expecting it to be in > Locality 0 as expected by TCG PC Client Platform Firmware Profile > Version 1.06 Revision 52 (3.1.1 – Pre-OS Environment) and then when it > fails to do so it simply halts the whole system. > > This issue might be related to the following bug: > https://bugzilla.kernel.org/show_bug.cgi?id=217890 > > Enable a user to configure the kernel > through “tpm.locality_on_suspend=1” boot parameter so that the locality > is set before suspend/shutdown in order to diagnose whether or not the > board is one of the buggy ones that require this workaround. Since this > bug is related to the board/platform instead of the specific TPM chip, > call dmi_check_system on the tpm_init function so that this setting is > automatically enabled for boards specified in code (ASUS TUF GAMING > B460M-PLUS already included) – automatic configuration only works in > case CONFIG_DMI is set though, since dmi_check_system is a non-op when > CONFIG_DMI is not set. > > In case “tpm.locality_on_suspend=0” (the default) don't change any > behavior thus preserving current functionality of any other board > except ASUSTeK COMPUTER INC. TUF GAMING B460M-PLUS and possibly future > boards as we successfully diagnose other boards with the same issue > fixed by using “tpm.locality_on_suspend=1”. > > Signed-off-by: Adam Alves <adamoa@xxxxxxxxx> > --- > drivers/char/tpm/tpm-chip.c | 9 ++++++++ > drivers/char/tpm/tpm-interface.c | 36 +++++++++++++++++++++++++++++++- > drivers/char/tpm/tpm.h | 1 + > include/linux/tpm.h | 1 + > 4 files changed, 46 insertions(+), 1 deletion(-) > > diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c > index 42b1062e33cd..8fdf7a137a94 100644 > --- a/drivers/char/tpm/tpm-chip.c > +++ b/drivers/char/tpm/tpm-chip.c > @@ -139,6 +139,9 @@ void tpm_chip_stop(struct tpm_chip *chip) > { > tpm_go_idle(chip); > tpm_relinquish_locality(chip); > + // If locality is to be preserved, we need to make sure it is Locality 0. If you put that kind C++ comment you should also check out https://www.kernel.org/doc/html/latest/process/coding-style.html Other stuff that I said in my earlier response still applies. BR, Jarkko
