Cr50 firmware has a requirement to wait for the TPM to wakeup before sending commands over the SPI bus. Otherwise, the firmware could be in deep sleep and not respond. Add a hook to tpm_tis_spi_transfer() before we start a SPI transfer so we can keep track of the last time the TPM driver accessed the SPI bus.
Cc: Andrey Pronin <apro...@chromium.org> Cc: Duncan Laurie <dlau...@chromium.org> Cc: Jason Gunthorpe <j...@ziepe.ca> Cc: Arnd Bergmann <a...@arndb.de> Cc: Greg Kroah-Hartman <gre...@linuxfoundation.org> Cc: Guenter Roeck <gro...@chromium.org> Cc: Alexander Steffen <alexander.stef...@infineon.com> Signed-off-by: Stephen Boyd <swb...@chromium.org> --- drivers/char/tpm/tpm_tis_spi.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/char/tpm/tpm_tis_spi.c b/drivers/char/tpm/tpm_tis_spi.c index 819602e85b34..93f49b1941f0 100644 --- a/drivers/char/tpm/tpm_tis_spi.c +++ b/drivers/char/tpm/tpm_tis_spi.c @@ -44,6 +44,7 @@ struct tpm_tis_spi_phy { struct spi_device *spi_device; int (*flow_control)(struct tpm_tis_spi_phy *phy, struct spi_transfer *xfer); + void (*pre_transfer)(struct tpm_tis_spi_phy *phy); u8 *iobuf; }; @@ -129,6 +130,8 @@ static int tpm_tis_spi_transfer(struct tpm_tis_data *data, u32 addr, u16 len, spi_message_init(&m); spi_message_add_tail(&spi_xfer, &m); + if (phy->pre_transfer) + phy->pre_transfer(phy); ret = spi_sync_locked(phy->spi_device, &m); if (ret < 0) goto exit; -- Sent by a computer through tubes