drivers: firmware: tisci: Add secure path support by Kronosblaster · Pull Request #101338 · zephyrproject-rtos/zephyr

Kronosblaster · 2025-12-19T11:28:43Z

Summary
Add support for secure TISCI communication paths to enable usage of secure threads for communication.

Changes:

Add is_secure flag to tisci_config for secure/non-secure path selection
Implement tisci_msg_requires_secure_path() to identify messages requiring secure path
Add secure header wrapping for outbound messages on secure path
Add secure header parsing for inbound messages on secure path
Improve error handling with proper semaphore cleanup in error paths
Add ti,is-secure device tree property to indicate secure host entities

Testing
Change the host-id and thread ids in the core specific dt file and run the /tests/drivers/tisci

soumya-TI · 2025-12-19T14:10:38Z

 	struct tisci_data *data = dev->data;
+	const struct tisci_config *config = dev->config;
+
+	if (!config->is_secure && tisci_msg_requires_secure_path(msg_type)) {


Why are we having a block here, not to send a message on a non-secure path, its job of server to nack it. isn't it? Client driver should be dumb enough to send the message, if that so, then all the messages should go as its is there shouldn't be a need to differentiate secure and non-secure messages. If we are differentiating we are building a intelligence, that can be added to support both secure and non-secure pathways on the same driver. Isn't it?

Resolving this comment, from a host perspective it can be either secure or non-secure, if non-secure we should not have secure messages sent out. The block seems okay in that perspective

I agree with your original assessment, a host can either be secure or non-secure. If secure then all messages go over the secure channel, non-secure -> all over non-secure. The TI-SCI client should not have to know details about which messages are "secure" messages, if they are and the host is not, then TI-SCI firmware will simply reject the message and we will return the appropriate error.

Suggest removing patch 2 from this series.

TISCI firmware doesnt reject the message. It silently drops it. This might lead to confusion if not blocked at send

TISCI firmware not responding will result in a timeout waiting for the result which is treated as an error the same as a reject message.

sonarqubecloud · 2026-01-08T08:47:06Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

glneo · 2026-01-23T14:39:19Z

+#define TISCI_MSG_ENTER_SLEEP 0x0301

-/* Clock requests */
+/* Security Messages */


Why are these message ID's put up here, why not down below with other high numbered messages (0x9000 messages should go almost at the bottom of the list)

These are still all out of order, why keep updating this series without first addressing my simple request to keep the definitions in numerical order?

glneo · 2026-01-23T14:45:15Z

+	case TISCI_MSG_READ_KEYCNT_KEYREV:
+	case TISCI_MSG_KEY_WRITER:
+	case TISCI_MSG_WRITE_KEYREV:
+	case TISCI_MSG_ENTER_SLEEP:


If a message is secure the core should not have any reason to use that message to begin with. This list is info that shouldn't need to be stored over here in applications using TI-SCI. I'd suggest just dropping this check for now as it is holding up this series that otherwise would almost ready.

LaurentiuM1234 · 2026-02-11T15:38:45Z

FYI got my eyes on this one but would like to see an ACK from TI before giving my +1 as everything here is TI-specific

glneo · 2026-02-13T15:36:49Z

+	/* Skip secure header (if present) and copy tisci_msg_hdr + payload */
+	memcpy(xfer->rx_message.buf,
+	       (uint8_t *)data->rx_message.buf + rx_offset,
+	       xfer->rx_message.size);


Why not just do this data move/shift up inside the if (config->is_secure) block, then you don't need to keep track of rx_offset. You could also just subtract out the secure header bytes from the xfer->rx_message.size at that point too so you can keep the old size check as it was (less churn).

- Add missing TISCI message type macros in tisci.h - Clean macro organisation - Add Secure indicator for message that require secure thread Signed-off-by: Dave Joseph <d-joseph@ti.com>

- Add ti,is-secure device tree property to indicate secure host entities - Add is_secure field in device config Signed-off-by: Dave Joseph <d-joseph@ti.com>

Add support for secure TISCI communication paths to enable message exchange with secure contexts on TI SoCs. - Add secure header wrapping for outbound messages on secure path - Add secure header parsing for inbound messages on secure path Signed-off-by: Dave Joseph <d-joseph@ti.com>

- Add proper semaphore cleanup in error paths where it was missing Signed-off-by: Dave Joseph <d-joseph@ti.com>

sonarqubecloud · 2026-02-16T10:57:50Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

zephyrbot added area: Devicetree Bindings area: Firmware area: Boards/SoCs platform: TI SimpleLink Texas Instruments SimpleLink MCU labels Dec 19, 2025

zephyrbot requested review from LaurentiuM1234, bogdanovs, jpanisbl, vaishnavachath and yongxu-wang15 December 19, 2025 11:30

zephyrbot assigned LaurentiuM1234 Dec 19, 2025

Kronosblaster force-pushed the tisci_fix branch from 35f700d to e408c96 Compare December 19, 2025 11:43

Kronosblaster marked this pull request as draft December 19, 2025 14:06

Kronosblaster force-pushed the tisci_fix branch from e408c96 to 1eaa1de Compare December 19, 2025 14:40

soumya-TI reviewed Dec 19, 2025

View reviewed changes

Kronosblaster force-pushed the tisci_fix branch 2 times, most recently from c7c2809 to af7060b Compare December 22, 2025 08:37

Kronosblaster marked this pull request as ready for review December 22, 2025 08:57

Kronosblaster force-pushed the tisci_fix branch from af7060b to 1eaad0d Compare January 8, 2026 08:42

zephyrbot added the area: Devicetree Binding PR modifies or adds a Device Tree binding label Jan 8, 2026

zephyrbot requested a review from JarmouniA January 8, 2026 08:44

Kronosblaster requested a review from glneo January 15, 2026 06:08

glneo reviewed Jan 23, 2026

View reviewed changes

LaurentiuM1234 mentioned this pull request Feb 11, 2026

mbox: mbox_ti_secproxy: Driver optimization and bugfixes #100642

Merged

Kronosblaster force-pushed the tisci_fix branch from 1eaad0d to 16cdcac Compare February 11, 2026 10:48

Kronosblaster force-pushed the tisci_fix branch from 16cdcac to 57bdb00 Compare February 12, 2026 05:13

Kronosblaster requested a review from glneo February 12, 2026 05:15

glneo reviewed Feb 13, 2026

View reviewed changes

Comment thread drivers/firmware/tisci/tisci.c

Kronosblaster force-pushed the tisci_fix branch from 57bdb00 to fa8c0c1 Compare February 13, 2026 05:17

Kronosblaster requested a review from glneo February 13, 2026 05:17

glneo reviewed Feb 13, 2026

View reviewed changes

Kronosblaster force-pushed the tisci_fix branch 2 times, most recently from 6fb7ff4 to 8b9057f Compare February 16, 2026 10:10

Kronosblaster requested a review from glneo February 16, 2026 10:10

Kronosblaster added 4 commits February 16, 2026 16:21

drivers: firmware: tisci: Reorganise macros

8bbad96

- Add missing TISCI message type macros in tisci.h - Clean macro organisation - Add Secure indicator for message that require secure thread Signed-off-by: Dave Joseph <d-joseph@ti.com>

drivers: firmware: tisci: Add secure property

c276132

- Add ti,is-secure device tree property to indicate secure host entities - Add is_secure field in device config Signed-off-by: Dave Joseph <d-joseph@ti.com>

drivers: firmware: tisci: Fix semaphore cleanup

3a1f337

- Add proper semaphore cleanup in error paths where it was missing Signed-off-by: Dave Joseph <d-joseph@ti.com>

Kronosblaster force-pushed the tisci_fix branch from 8b9057f to 3a1f337 Compare February 16, 2026 10:52

glneo approved these changes Feb 18, 2026

View reviewed changes

LaurentiuM1234 approved these changes Feb 18, 2026

View reviewed changes

aescolar merged commit 2f3af50 into zephyrproject-rtos:main Feb 19, 2026
27 checks passed

Kronosblaster mentioned this pull request Mar 18, 2026

BACKPORT: TISCI and SecProxy upstream patches TexasInstruments/zephyr#19

Merged

Conversation

Kronosblaster commented Dec 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sonarqubecloud Bot commented Jan 8, 2026

Quality Gate passed

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

LaurentiuM1234 commented Feb 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sonarqubecloud Bot commented Feb 16, 2026

Quality Gate passed

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Kronosblaster commented Dec 19, 2025 •

edited

Loading

LaurentiuM1234 commented Feb 11, 2026 •

edited

Loading