soc/intel/common/block/tco: clear TCO1_STS register, too

The register TCO1_STS is never cleared, which will cause SMIs to either
retrigger over and over again (e.g. TIMEOUT) or prevent concurrent
interrupt events, depending on which event triggered.

Clear both TCO2_STS and TCO1_STS.

This also fixes the issue where SECOND_TO_STS will always end up set in
the SMI handler by unconditionally (re)setting it.

Tested on X11SSM-F, where enabling TCO caused the terminal to get
flooded with SMI debug messages. With this patch, a message gets written
every ~1 second.

Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Change-Id: Ia57c203a672fdd0095355a7e2a0e01aaa6657968
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39259
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-by: Andrey Petrov <anpetrov@fb.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>

1 files changed, 5 insertions, 5 deletions

diff --git a/src/soc/intel/common/block/smbus/tco.c b/src/soc/intel/common/block/smbus/tco.c
index 1a215eb69d..bd8790aa6e 100644
--- a/src/soc/intel/common/block/smbus/tco.c
+++ b/src/soc/intel/common/block/smbus/tco.c
@@ -74,13 +74,13 @@ uint32_t tco_reset_status(void)
 	uint16_t tco1_sts;
 	uint16_t tco2_sts;
 
-	/* TCO Status 2 register */
-	tco2_sts = tco_read_reg(TCO2_STS);
-	tco2_sts |= TCO_STS_SECOND_TO;
-	tco_write_reg(TCO2_STS, tco2_sts);
-
 	/* TCO Status 1 register */
 	tco1_sts = tco_read_reg(TCO1_STS);
+	tco_write_reg(TCO1_STS, tco1_sts);
+
+	/* TCO Status 2 register */
+	tco2_sts = tco_read_reg(TCO2_STS);
+	tco_write_reg(TCO2_STS, tco2_sts | TCO_STS_SECOND_TO);
 
 	return (tco2_sts << 16) | tco1_sts;
 }