diff options
| author | Randall Spangler <rspangler@chromium.org> | 2016-08-05 10:25:42 -0700 |
|---|---|---|
| committer | chrome-bot <chrome-bot@chromium.org> | 2016-09-24 13:53:26 -0700 |
| commit | b94747d501ad7e4347b50c3ea55e615422d96212 (patch) | |
| tree | b2c078ca5b87a8af29827474789475903f5b316f /chip/g/sps_tpm.c | |
| parent | 5fa727805ca3222e9268755c54c15bbf87c5b2bd (diff) | |
| download | chrome-ec-b94747d501ad7e4347b50c3ea55e615422d96212.tar.gz | |
cr50: reset only the TPM state on SYS_RST_L
Rather than doing a full cr50 reset when the system reset SYS_RST_L is
asserted, just reset the state of the TPM task and library. Re-clear
.bss for those modules, then re-initialize.
BRANCH=none
BUG=chrome-os-partner:52366
CQ-DEPEND=CL:366792
TEST=make buildall; test on Gru
Trigger a SYS_RST_L by using the AP's reboot command, power
off/on, log in/out/in.
See that the Cr50 does not reboot and the firmware and userspace
are still happy about the TPM.
Change-Id: I32cd2bb72316f68c74db77a20a8d09112b402d4b
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/361680
Diffstat (limited to 'chip/g/sps_tpm.c')
| -rw-r--r-- | chip/g/sps_tpm.c | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/chip/g/sps_tpm.c b/chip/g/sps_tpm.c index 2c3f6876e7..3ef12994e6 100644 --- a/chip/g/sps_tpm.c +++ b/chip/g/sps_tpm.c @@ -271,3 +271,12 @@ void sps_tpm_enable(void) sps_register_rx_handler(SPS_GENERIC_MODE, tpm_rx_handler, 3); init_new_cycle(); } + +void sps_tpm_disable(void) +{ + sps_tpm_state = SPS_TPM_STATE_PONDERING; + sps_unregister_rx_handler(); + /* We don't care anymore, so we can sleep whenever */ + delay_sleep_by(0); + enable_sleep(SLEEP_MASK_SPI); +} |