src/sys/dev/tpm, branch stable/14 FreeBSD source tree tpm20: fix suspend/resume and entropy harvesting 2026-04-12T13:43:44+00:00 Kyle Evans kevans@FreeBSD.org 2026-02-05T03:35:01+00:00 f329d69b9176a1d407276a010c2b87a07a2bbc5d There were a few problem here: - TPM2_Shutdown results in a response that we need to either process or ignore, otherwise any tpm20_write or tpm20_harvest call will trivially hang on an `sc->pending_data_length != 0` - We should have a matching TPM2_Startup upon resume to restore any state that should have persisted - We must drain the harvest task before we suspend to avoid problems there This commit is sufficient to avoid breaking suspend/resume. Co-authored-by: markj Tested by: garga (cherry picked from commit 38a4995eb52db21116f8b37ed942e66a8c2f050f) 
There were a few problem here:
  - TPM2_Shutdown results in a response that we need to either process
    or ignore, otherwise any tpm20_write or tpm20_harvest call will
    trivially hang on an `sc->pending_data_length != 0`
  - We should have a matching TPM2_Startup upon resume to restore any
    state that should have persisted
  - We must drain the harvest task before we suspend to avoid problems
    there

This commit is sufficient to avoid breaking suspend/resume.

Co-authored-by:	markj
Tested by:	garga

(cherry picked from commit 38a4995eb52db21116f8b37ed942e66a8c2f050f)
tpm: Register TPM as entropy source 2024-09-07T05:13:14+00:00 Jean-François Hren jean-francois.hren@stormshield.eu 2024-08-31T17:53:02+00:00 0b683336b211e74b6d7f21ecf517c6591370b793 TPM is defined as an entropy and is called every 10 seconds. However it was not registered and calls were discarded. Signed-off-by: Jean-François Hren <jean-francois.hren@stormshield.eu> Pull Request: https://github.com/freebsd/freebsd-src/pull/1398 (cherry picked from commit d7c87526b1c3d32d3ac3fee8c59d92405b7621b9) 
TPM is defined as an entropy and is called every 10 seconds. However it
was not registered and calls were discarded.

Signed-off-by: Jean-François Hren <jean-francois.hren@stormshield.eu>

Pull Request:	https://github.com/freebsd/freebsd-src/pull/1398

(cherry picked from commit d7c87526b1c3d32d3ac3fee8c59d92405b7621b9)
sys: Remove $FreeBSD$: one-line .c pattern 2023-08-16T17:54:36+00:00 Warner Losh imp@FreeBSD.org 2023-08-16T17:54:36+00:00 685dc743dc3b5645e34836464128e1c0558b404b Remove /^[\s*]*__FBSDID\("\$FreeBSD\$"\);?\s*\n/ 
Remove /^[\s*]*__FBSDID\("\$FreeBSD\$"\);?\s*\n/
sys: Remove $FreeBSD$: two-line .h pattern 2023-08-16T17:54:11+00:00 Warner Losh imp@FreeBSD.org 2023-08-16T17:54:11+00:00 95ee2897e98f5d444f26ed2334cc7c439f9c16c6 Remove /^\s*\*\n \*\s+\$FreeBSD\$$\n/ 
Remove /^\s*\*\n \*\s+\$FreeBSD\$$\n/
tpm: Fix entropy harvesting logic 2022-09-02T11:09:41+00:00 Kornel Dulęba kd@FreeBSD.org 2022-09-02T11:07:03+00:00 e23ad9c4ba2d2fd235fc3dc988ca422ef4e7ca3b Use a taskqueue instead of a callout. Callout functions mustn't sleep, where as the TPM driver uses a sx lock in order to serialize accesses to the device. Since the entropy harvesting feature is not enabled by default, this commit should not bring any functional changes to the GENERIC kernel. Approved by: mw(mentor) Sponsored by: Stormshield Obtained from: Semihalf MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D36412 
Use a taskqueue instead of a callout.
Callout functions mustn't sleep, where as the TPM driver
uses a sx lock in order to serialize accesses to the device.
Since the entropy harvesting feature is not enabled by default,
this commit should not bring any functional changes to the GENERIC
kernel.

Approved by:	mw(mentor)
Sponsored by:	Stormshield
Obtained from:	Semihalf
MFC after:	1 week
Differential Revision:	https://reviews.freebsd.org/D36412
TPM: do not set device description if probe fails 2022-08-16T09:07:12+00:00 Konrad Sewiłło-Jopek kjopek@gmail.com 2022-08-16T07:27:36+00:00 31ee3a33aa612294db302828d4d1f6f0c83bef4c device_set_desc should be called only if driver probes successfully. Approved by: mw(mentor) Reviewed by: mw, kd MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D35998 
device_set_desc should be called only if driver probes successfully.

Approved by:	mw(mentor)
Reviewed by:	mw, kd
MFC after:	2 weeks
Differential Revision:	https://reviews.freebsd.org/D35998
tpm: Remove unused devclass arguments to DRIVER_MODULE. 2022-05-09T19:22:04+00:00 John Baldwin jhb@FreeBSD.org 2022-05-09T19:22:04+00:00 f5dfa6b98901e161bb7bac46034c6955e5baa0ee 






tpm: Remove unused variables.
2022-04-08T00:01:29+00:00

John Baldwin
jhb@FreeBSD.org

2022-04-08T00:01:29+00:00

08ff11e524230bf9bf4c2adfd8f094af64c5f956

All of the legacy config stuff in tpm seems broken.  Nothing ever sets
tpm_enabled to a non-zero value and the legacy bits expect tpm_enabled
to be set to the base IO port address (I think).





All of the legacy config stuff in tpm seems broken.  Nothing ever sets
tpm_enabled to a non-zero value and the legacy bits expect tpm_enabled
to be set to the base IO port address (I think).







tpm_tis: Improve interrupt allocation
2021-08-16T04:28:33+00:00

Kornel Duleba
mindal@semihalf.com

2021-08-16T04:26:50+00:00

59d1661cd8b02d13bc697e9dc0ff97844e719d9b

Validate the irq received from ACPI. Test if it works by sending a simple command and checking if the interrupt handler was executed.
Internal buffer allocation was moved away from common code to tis and crb parts - in order to test the interrupt we need to have it allocated early.

Obtained from:		Semihalf
Differential revision:	https://reviews.freebsd.org/D31395





Validate the irq received from ACPI. Test if it works by sending a simple command and checking if the interrupt handler was executed.
Internal buffer allocation was moved away from common code to tis and crb parts - in order to test the interrupt we need to have it allocated early.

Obtained from:		Semihalf
Differential revision:	https://reviews.freebsd.org/D31395







unifdef -D__FreeBSD__ to remove the OpenBSD support
2021-01-14T06:35:22+00:00

David E. O'Brien
obrien@FreeBSD.org

2021-01-14T04:45:54+00:00

9a01a25516367d1deeb48484d38a66137b01ebb0

OpenBSD never accepted this driver, and instead wrote their
own minimal one (sys/dev/acpi/tpm.c for suspending the device).

Reviewed by:    stevek, emaste
Differential Revision:  D10321





OpenBSD never accepted this driver, and instead wrote their
own minimal one (sys/dev/acpi/tpm.c for suspending the device).

Reviewed by:    stevek, emaste
Differential Revision:  D10321