[PATCH v4 1/2] tpm: Disable a TPM that stops responding
From: Breno Leitao
Date: Fri Nov 20 2026 - 15:51:45 EST
Some TPMs stop responding at runtime, hours to months after they passed
probe. The first sign is usually "invalid TPM_STS.x 0xff", and from then
on every command fails only after the transport timeout. A TPM that is
already dead at boot fails to probe and never registers, which this
patch does not change.
The callers keep sending more commands: the hwrng thread every 10
seconds forever, and userspace through /dev/tpmrm0. Each attempt holds
tpm_mutex for the whole timeout and logs an error:
tpm tpm0: invalid TPM_STS.x 0xff, dumping stack for forensics
tpm tpm0: tpm_try_transmit: send(): error -62
tpm tpm0: tpm_try_transmit: send(): error -62
...
The core already stops using a TPM it cannot trust: tpm_try_get_ops()
fails with -EIO once TPM_CHIP_FLAG_DISABLE is set, which
tpm2_load_null() does when it cannot restore the null key. Set the same
flag when two commands in a row time out.
The first timeout sets TPM_CHIP_FLAG_MISSED_RESPONSE, and a command that
gets an answer, whatever its TPM return code, clears it.
With interrupts, the wait helpers in tpm_tis_core, st33zp24 and
xen-tpmfront return -ETIME as soon as a signal is pending, so an -ETIME
with a signal pending is not counted. Otherwise any user could disable
the TPM by interrupting two reads of a PCR file in sysfs.
Some TPM 1.2 chips do not get ready while their self test runs, and
tpm1_do_selftest() keeps polling through those timeouts, see
commit 4643826a3da1 ("TPM: Work around buggy TPMs that block during
continue self test"). It now clears the flag after each of them.
A disabled TPM stays disabled until the driver is rebound or the machine
reboots.
Suggested-by: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
Signed-off-by: Breno Leitao <leitao@xxxxxxxxxx>
---
drivers/char/tpm/tpm-interface.c | 15 +++++++++++++++
drivers/char/tpm/tpm1-cmd.c | 1 +
include/linux/tpm.h | 1 +
3 files changed, 17 insertions(+)
diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c
index e0385b16616b46..cb794cfc5f63e8 100644
--- a/drivers/char/tpm/tpm-interface.c
+++ b/drivers/char/tpm/tpm-interface.c
@@ -26,6 +26,7 @@
#include <linux/spinlock.h>
#include <linux/suspend.h>
#include <linux/freezer.h>
+#include <linux/sched/signal.h>
#include <linux/tpm_eventlog.h>
#include "tpm.h"
@@ -268,6 +269,20 @@ ssize_t tpm_transmit(struct tpm_chip *chip, u8 *buf, size_t bufsiz)
delay_msec *= 2;
memcpy(buf, save, save_size);
}
+
+ if (ret >= 0)
+ chip->flags &= ~TPM_CHIP_FLAG_MISSED_RESPONSE;
+
+ /* Some drivers also return -ETIME when a signal interrupts the wait. */
+ if (ret == -ETIME && !signal_pending(current)) {
+ if (!(chip->flags & TPM_CHIP_FLAG_MISSED_RESPONSE)) {
+ chip->flags |= TPM_CHIP_FLAG_MISSED_RESPONSE;
+ } else if (!(chip->flags & TPM_CHIP_FLAG_DISABLE)) {
+ dev_err(&chip->dev,
+ "two commands in a row timed out, disabling the TPM\n");
+ chip->flags |= TPM_CHIP_FLAG_DISABLE;
+ }
+ }
return ret;
}
diff --git a/drivers/char/tpm/tpm1-cmd.c b/drivers/char/tpm/tpm1-cmd.c
index 7a8606e8c7117b..e4e1f0799da196 100644
--- a/drivers/char/tpm/tpm1-cmd.c
+++ b/drivers/char/tpm/tpm1-cmd.c
@@ -647,6 +647,7 @@ int tpm1_do_selftest(struct tpm_chip *chip)
* until the self test duration expires.
*/
if (rc == -ETIME) {
+ chip->flags &= ~TPM_CHIP_FLAG_MISSED_RESPONSE;
dev_info(&chip->dev, HW_ERR "TPM command timed out during continue self test");
tpm_msleep(delay_msec);
continue;
diff --git a/include/linux/tpm.h b/include/linux/tpm.h
index 0db277af45c393..0f97abc82846ae 100644
--- a/include/linux/tpm.h
+++ b/include/linux/tpm.h
@@ -197,6 +197,7 @@ enum tpm_chip_flags {
TPM_CHIP_FLAG_HWRNG_DISABLED = BIT(9),
TPM_CHIP_FLAG_DISABLE = BIT(10),
TPM_CHIP_FLAG_SYNC = BIT(11),
+ TPM_CHIP_FLAG_MISSED_RESPONSE = BIT(12),
};
#define to_tpm_chip(d) container_of(d, struct tpm_chip, dev)
--
2.53.1-Meta