>-----Original Message----- >From: Jason Gunthorpe [mailto:jgg@xxxxxxxx] >Sent: Thursday, December 21, 2017 12:27 PM >To: Shaikh, Azhar <azhar.shaikh@xxxxxxxxx> >Cc: jarkko.sakkinen@xxxxxxxxxxxxxxx; javierm@xxxxxxxxxx; >peterhuewe@xxxxxx; linux-security-module@xxxxxxxxxxxxxxx; linux- >integrity@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; tpmdd- >devel@xxxxxxxxxxxxxxxxxxxxx >Subject: Re: [PATCH] tpm: Fix the driver cleanup code > >On Thu, Dec 21, 2017 at 12:17:02PM -0800, Azhar Shaikh wrote: >> Commit 3c1701339284353c41 ("tpm: Keep CLKRUN enabled throughout the >> duration of transmit_cmd()") added code which accessed >> chip->ops, even after it was set to NULL in tpm_del_char_device(), >> called from tpm_chip_unregister() in error / driver exit paths. >> So fix this code. >> >> Fixes: 3c1701339284353c41 ("tpm: Keep CLKRUN enabled throughout the >> duration of transmit_cmd()") >> >> Suggested-by: Javier Martinez Canillas <javierm@xxxxxxxxxx> >> Suggested-by: Jason Gunthorpe <jgg@xxxxxxxx> >> Signed-off-by: Azhar Shaikh <azhar.shaikh@xxxxxxxxx> >> drivers/char/tpm/tpm-chip.c | 5 +++++ >> drivers/char/tpm/tpm.h | 1 + >> drivers/char/tpm/tpm_tis.c | 8 ++------ >> drivers/char/tpm/tpm_tis_core.c | 23 +++++++++++++++-------- >> drivers/char/tpm/tpm_tis_spi.c | 1 + >> 5 files changed, 24 insertions(+), 14 deletions(-) >> >> diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c >> index 0a62c19937b6..c88ec9a32a7e 100644 >> +++ b/drivers/char/tpm/tpm-chip.c >> @@ -346,6 +346,10 @@ static void tpm_del_char_device(struct tpm_chip >*chip) >> down_write(&chip->ops_sem); >> if (chip->flags & TPM_CHIP_FLAG_TPM2) >> tpm2_shutdown(chip, TPM2_SU_CLEAR); >> + if (chip->flags & TPM_CHIP_FLAG_DO_NOT_CLEAR_OPS) { >> + up_write(&chip->ops_sem); >> + return; >> + } > >Yikes, too ugly. Figure something else out.. ops->clk_enable is always >tpm_tis_clkrun_enable at this point, so maybe call it directly in the removal >path in tpm_tis_core? > Yes I thought about it too. But if some other chip->ops function in future, which *might* be in this same case, hence for that introduced this flag. I can use ops->clk_enable instead of introducing the new flag. >Jason Regards, Azhar Shaikh
