Skip to contentSkip to conversation

tpm: ibmvtpm: Avoid -EINTR error when IMA talks to TPM

PATCH In review Stefan Berger 2 participants 3 messages Feb 9, 2021 – Feb 12, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 tpm: ibmvtpm: Avoid -EINTR error when IMA talks to TPM

When IMA is taking measurements during compilation for example and a user presses ctrl-c to abort the compilation, lots of these types of messages will appear in the kernel log:

[ 7406.275163] tpm tpm0: tpm_transmit: tpm_recv: error -4
[ 7406.275242] ima: Error Communicating to TPM chip, result: -4

The issue is caused by the fact that the IBM vTPM driver's recv() function is called immediately after send() without waiting for status on whether a response was received. It currently waits for the current command to finish using this call that ends up throwing these error messages because it is 'interruptible':

sig = wait_event_interruptible(ibmvtpm->wq,

                               !ibmvtpm->tpm_processing_cmd);

Instead, it should be using the polling loop in tpm_try_transmit() that uses a command's duration to poll until a result has been returned by the TPM, thus ending when the timeout has occurred but not responding to users' ctrl-c request anymore. To stay in this polling loop we now extend tpm_ibmvtpm_status() to return PM_STATUS_BUSY for as long as the vTPM is busy. Since we will need the timeouts in this loop now we get the TPM 1.2 and TPM 2 timeouts with tpm_get_timeouts().

We change tpm_processing_cmd to tpm_status and set the TPM_STATUS_BUSY flag while the vTPM is busy processing a command.

Fixes 18b3670d79ae9 ("tpm: ibmvtpm: Add support for TPM2")
Cc Nayna Jain <[email protected]>
Cc George Wilson <[email protected]>
---
 drivers/char/tpm/tpm_ibmvtpm.c | 31 ++++++++++++++++++-------------
 drivers/char/tpm/tpm_ibmvtpm.h |  3 ++-
 2 files changed, 20 insertions(+), 14 deletions(-)
drivers/char/tpm/tpm_ibmvtpm.c +18 -13
index 994385bf37c0..6290bd8889e4 100644
@@ -106,17 +106,12 @@ static int tpm_ibmvtpm_recv(struct tpm_chip *chip, u8 *buf, size_t count)
 {
 	struct ibmvtpm_dev *ibmvtpm = dev_get_drvdata(&chip->dev);
 	u16 len;
-	int sig;
 
 	if (!ibmvtpm->rtce_buf) {
 		dev_err(ibmvtpm->dev, "ibmvtpm device is not ready\n");
 		return 0;
 	}
 
-	sig = wait_event_interruptible(ibmvtpm->wq, !ibmvtpm->tpm_processing_cmd);
-	if (sig)
-		return -EINTR;
-
 	len = ibmvtpm->res_len;
 
 	if (count < len) {
@@ -220,11 +215,12 @@ static int tpm_ibmvtpm_send(struct tpm_chip *chip, u8 *buf, size_t count)
 		return -EIO;
 	}
 
-	if (ibmvtpm->tpm_processing_cmd) {
+	if ((ibmvtpm->tpm_status & TPM_STATUS_BUSY)) {
 		dev_info(ibmvtpm->dev,
 		         "Need to wait for TPM to finish\n");
 		/* wait for previous command to finish */
-		sig = wait_event_interruptible(ibmvtpm->wq, !ibmvtpm->tpm_processing_cmd);
+		sig = wait_event_interruptible(ibmvtpm->wq,
+				(ibmvtpm->tpm_status & TPM_STATUS_BUSY) == 0);
 		if (sig)
 			return -EINTR;
 	}
@@ -237,7 +233,7 @@ static int tpm_ibmvtpm_send(struct tpm_chip *chip, u8 *buf, size_t count)
 	 * set the processing flag before the Hcall, since we may get the
 	 * result (interrupt) before even being able to check rc.
 	 */
-	ibmvtpm->tpm_processing_cmd = true;
+	ibmvtpm->tpm_status |= TPM_STATUS_BUSY;
 
 again:
 	rc = ibmvtpm_send_crq(ibmvtpm->vdev,
@@ -255,7 +251,7 @@ static int tpm_ibmvtpm_send(struct tpm_chip *chip, u8 *buf, size_t count)
 			goto again;
 		}
 		dev_err(ibmvtpm->dev, "tpm_ibmvtpm_send failed rc=%d\n", rc);
-		ibmvtpm->tpm_processing_cmd = false;
+		ibmvtpm->tpm_status &= ~TPM_STATUS_BUSY;
 	}
 
 	spin_unlock(&ibmvtpm->rtce_lock);
@@ -269,7 +265,9 @@ static void tpm_ibmvtpm_cancel(struct tpm_chip *chip)
 
 static u8 tpm_ibmvtpm_status(struct tpm_chip *chip)
 {
-	return 0;
+	struct ibmvtpm_dev *ibmvtpm = dev_get_drvdata(&chip->dev);
+
+	return ibmvtpm->tpm_status;
 }
 
 /**
@@ -459,7 +457,7 @@ static const struct tpm_class_ops tpm_ibmvtpm = {
 	.send = tpm_ibmvtpm_send,
 	.cancel = tpm_ibmvtpm_cancel,
 	.status = tpm_ibmvtpm_status,
-	.req_complete_mask = 0,
+	.req_complete_mask = TPM_STATUS_BUSY,
 	.req_complete_val = 0,
 	.req_canceled = tpm_ibmvtpm_req_canceled,
 };
@@ -552,7 +550,7 @@ static void ibmvtpm_crq_process(struct ibmvtpm_crq *crq,
 		case VTPM_TPM_COMMAND_RES:
 			/* len of the data in rtce buffer */
 			ibmvtpm->res_len = be16_to_cpu(crq->len);
-			ibmvtpm->tpm_processing_cmd = false;
+			ibmvtpm->tpm_status &= ~TPM_STATUS_BUSY;
 			wake_up_interruptible(&ibmvtpm->wq);
 			return;
 		default:
@@ -690,8 +688,15 @@ static int tpm_ibmvtpm_probe(struct vio_dev *vio_dev,
 		goto init_irq_cleanup;
 	}
 
-	if (!strcmp(id->compat, "IBM,vtpm20")) {
+
+	if (!strcmp(id->compat, "IBM,vtpm20"))
 		chip->flags |= TPM_CHIP_FLAG_TPM2;
+
+	rc = tpm_get_timeouts(chip);
+	if (rc)
+		goto init_irq_cleanup;
+
+	if (chip->flags & TPM_CHIP_FLAG_TPM2) {
 		rc = tpm2_get_cc_attrs_tbl(chip);
 		if (rc)
 			goto init_irq_cleanup;
drivers/char/tpm/tpm_ibmvtpm.h +2 -1
index b92aa7d3e93e..252f1cccdfc5 100644
@@ -41,7 +41,8 @@ struct ibmvtpm_dev {
 	wait_queue_head_t wq;
 	u16 res_len;
 	u32 vtpm_version;
-	bool tpm_processing_cmd;
+	u8 tpm_status;
+#define TPM_STATUS_BUSY		(1 << 0) /* vtpm is processing a command */
 };
 
 #define CRQ_RES_BUF_SIZE	PAGE_SIZE
Signature · 1 line
2.29.2
maintainer
#1 lore raw
Quoted 4 lines · Stefan Berger

When IMA is taking measurements during compilation for example and a user presses ctrl-c to abort the compilation, lots of these types of messages will appear in the kernel log:

Quoted 8 lines
[ 7406.275163] tpm tpm0: tpm_transmit: tpm_recv: error -4
[ 7406.275242] ima: Error Communicating to TPM chip, result: -4

The issue is caused by the fact that the IBM vTPM driver's recv() function is called immediately after send() without waiting for status on whether a response was received. It currently waits for the current command to finish using this call that ends up throwing these error messages because it is 'interruptible':

Why it is an issue?

Quoted 12 lines

sig = wait_event_interruptible(ibmvtpm->wq,

                               !ibmvtpm->tpm_processing_cmd);

Instead, it should be using the polling loop in tpm_try_transmit() that uses a command's duration to poll until a result has been returned by the TPM, thus ending when the timeout has occurred but not responding to users' ctrl-c request anymore. To stay in this polling loop we now extend tpm_ibmvtpm_status() to return PM_STATUS_BUSY for as long as the vTPM is busy. Since we will need the timeouts in this loop now we get the TPM 1.2 and TPM 2 timeouts with tpm_get_timeouts().

Quoted 2 lines

We change tpm_processing_cmd to tpm_status and set the TPM_STATUS_BUSY flag while the vTPM is busy processing a command.

Please, never use "we". The commit message should describe in imperative form the action taken.

Quoted 5 lines
Fixes 18b3670d79ae9 ("tpm: ibmvtpm: Add support for TPM2")
Cc Nayna Jain <[email protected]>
Cc George Wilson <[email protected]>

Your SOB should be last.

Quoted 124 lines
---
 drivers/char/tpm/tpm_ibmvtpm.c | 31 ++++++++++++++++++-------------
 drivers/char/tpm/tpm_ibmvtpm.h |  3 ++-
 2 files changed, 20 insertions(+), 14 deletions(-)
drivers/char/tpm/tpm_ibmvtpm.c +18 -13
index 994385bf37c0..6290bd8889e4 100644
@@ -106,17 +106,12 @@ static int tpm_ibmvtpm_recv(struct tpm_chip *chip, u8 *buf, size_t count)
 {
 	struct ibmvtpm_dev *ibmvtpm = dev_get_drvdata(&chip->dev);
 	u16 len;
-	int sig;
 
 	if (!ibmvtpm->rtce_buf) {
 		dev_err(ibmvtpm->dev, "ibmvtpm device is not ready\n");
 		return 0;
 	}
 
-	sig = wait_event_interruptible(ibmvtpm->wq, !ibmvtpm->tpm_processing_cmd);
-	if (sig)
-		return -EINTR;
-
 	len = ibmvtpm->res_len;
 
 	if (count < len) {
@@ -220,11 +215,12 @@ static int tpm_ibmvtpm_send(struct tpm_chip *chip, u8 *buf, size_t count)
 		return -EIO;
 	}
 
-	if (ibmvtpm->tpm_processing_cmd) {
+	if ((ibmvtpm->tpm_status & TPM_STATUS_BUSY)) {
 		dev_info(ibmvtpm->dev,
 		         "Need to wait for TPM to finish\n");
 		/* wait for previous command to finish */
-		sig = wait_event_interruptible(ibmvtpm->wq, !ibmvtpm->tpm_processing_cmd);
+		sig = wait_event_interruptible(ibmvtpm->wq,
+				(ibmvtpm->tpm_status & TPM_STATUS_BUSY) == 0);
 		if (sig)
 			return -EINTR;
 	}
@@ -237,7 +233,7 @@ static int tpm_ibmvtpm_send(struct tpm_chip *chip, u8 *buf, size_t count)
 	 * set the processing flag before the Hcall, since we may get the
 	 * result (interrupt) before even being able to check rc.
 	 */
-	ibmvtpm->tpm_processing_cmd = true;
+	ibmvtpm->tpm_status |= TPM_STATUS_BUSY;
 
 again:
 	rc = ibmvtpm_send_crq(ibmvtpm->vdev,
@@ -255,7 +251,7 @@ static int tpm_ibmvtpm_send(struct tpm_chip *chip, u8 *buf, size_t count)
 			goto again;
 		}
 		dev_err(ibmvtpm->dev, "tpm_ibmvtpm_send failed rc=%d\n", rc);
-		ibmvtpm->tpm_processing_cmd = false;
+		ibmvtpm->tpm_status &= ~TPM_STATUS_BUSY;
 	}
 
 	spin_unlock(&ibmvtpm->rtce_lock);
@@ -269,7 +265,9 @@ static void tpm_ibmvtpm_cancel(struct tpm_chip *chip)
 
 static u8 tpm_ibmvtpm_status(struct tpm_chip *chip)
 {
-	return 0;
+	struct ibmvtpm_dev *ibmvtpm = dev_get_drvdata(&chip->dev);
+
+	return ibmvtpm->tpm_status;
 }
 
 /**
@@ -459,7 +457,7 @@ static const struct tpm_class_ops tpm_ibmvtpm = {
 	.send = tpm_ibmvtpm_send,
 	.cancel = tpm_ibmvtpm_cancel,
 	.status = tpm_ibmvtpm_status,
-	.req_complete_mask = 0,
+	.req_complete_mask = TPM_STATUS_BUSY,
 	.req_complete_val = 0,
 	.req_canceled = tpm_ibmvtpm_req_canceled,
 };
@@ -552,7 +550,7 @@ static void ibmvtpm_crq_process(struct ibmvtpm_crq *crq,
 		case VTPM_TPM_COMMAND_RES:
 			/* len of the data in rtce buffer */
 			ibmvtpm->res_len = be16_to_cpu(crq->len);
-			ibmvtpm->tpm_processing_cmd = false;
+			ibmvtpm->tpm_status &= ~TPM_STATUS_BUSY;
 			wake_up_interruptible(&ibmvtpm->wq);
 			return;
 		default:
@@ -690,8 +688,15 @@ static int tpm_ibmvtpm_probe(struct vio_dev *vio_dev,
 		goto init_irq_cleanup;
 	}
 
-	if (!strcmp(id->compat, "IBM,vtpm20")) {
+
+	if (!strcmp(id->compat, "IBM,vtpm20"))
 		chip->flags |= TPM_CHIP_FLAG_TPM2;
+
+	rc = tpm_get_timeouts(chip);
+	if (rc)
+		goto init_irq_cleanup;
+
+	if (chip->flags & TPM_CHIP_FLAG_TPM2) {
 		rc = tpm2_get_cc_attrs_tbl(chip);
 		if (rc)
 			goto init_irq_cleanup;
drivers/char/tpm/tpm_ibmvtpm.h +2 -1
index b92aa7d3e93e..252f1cccdfc5 100644
@@ -41,7 +41,8 @@ struct ibmvtpm_dev {
 	wait_queue_head_t wq;
 	u16 res_len;
 	u32 vtpm_version;
-	bool tpm_processing_cmd;
+	u8 tpm_status;
+#define TPM_STATUS_BUSY		(1 << 0) /* vtpm is processing a command */
 };
 
 #define CRQ_RES_BUF_SIZE	PAGE_SIZE
Signature · 1 line
2.29.2
#2 lore raw
Quoted 8 lines · Jarkko Sakkinen
Quoted 7 lines · Stefan Berger

When IMA is taking measurements during compilation for example and a user presses ctrl-c to abort the compilation, lots of these types of messages will appear in the kernel log:

[ 7406.275163] tpm tpm0: tpm_transmit: tpm_recv: error -4
[ 7406.275242] ima: Error Communicating to TPM chip, result: -4
Quoted 6 lines
Quoted 5 lines

The issue is caused by the fact that the IBM vTPM driver's recv() function is called immediately after send() without waiting for status on whether a response was received. It currently waits for the current command to finish using this call that ends up throwing these error messages because it is 'interruptible':

Why it is an issue?

The issue is the many kernel log entries we get when someone interrupts an application with ctrl-c while IMA is taking measurements of files it reads.I thought that was clear from the first paragraph.