From: Mario Limonciello <mario.limonciello@amd.com>
To: Jarkko Sakkinen <jarkko@kernel.org>
Cc: peterhuewe@gmx.de, jgg@ziepe.ca, linux-integrity@vger.kernel.org
Subject: Re: [PATCH v2] tpm: Call cmd_ready/go_idle for each command transmission
Date: Tue, 1 Sep 2026 12:54:02 -0500 [thread overview]
Message-ID: <fb4affaa-0093-4a5d-9591-63700b46e64d@amd.com> (raw)
In-Reply-To: <apcNU_RLDkrfZraq@kernel.org>
On 9/1/26 12:37, Jarkko Sakkinen wrote:
> On Tue, Sep 01, 2026 at 11:32:15AM -0500, Mario Limonciello wrote:
>> Some TPM implementations, particularly fTPM using the CRB interface,
>> require the TPM to transition through idle and ready states for each
>> command rather than once per session.
>>
>> The current implementation calls cmd_ready once during tpm_chip_start()
>> and go_idle once during tpm_chip_stop(). For fTPM, when multiple commands
>> are sent without per-command idle transitions, subsequent commands timeout
>> as the TPM is waiting for the transition.
>>
>> Fix this by calling cmd_ready before each command and go_idle on all exit
>> paths in tpm_try_transmit(). For TPM implementations that don't require
>> per-command transitions, the callbacks return immediately based on the
>> start method.
>>
>> Remove the now-redundant per-session calls from tpm_chip_start() and
>> tpm_chip_stop() along with their helpers.
>>
>> This resolves timeout errors during TPM initialization on systems where
>> BIOS has already performed TPM startup.
>>
>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>> ---
>> v2:
>> * Drop redundant calls
>> * Update commit message
>> ---
>> drivers/char/tpm/tpm-chip.c | 24 ------------------------
>> drivers/char/tpm/tpm-interface.c | 26 ++++++++++++++++++++++----
>> 2 files changed, 22 insertions(+), 28 deletions(-)
>>
>> diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
>> index 12b7394b34bdc..0be5dbfaa72eb 100644
>> --- a/drivers/char/tpm/tpm-chip.c
>> +++ b/drivers/char/tpm/tpm-chip.c
>> @@ -66,22 +66,6 @@ static void tpm_relinquish_locality(struct tpm_chip *chip)
>> chip->locality = -1;
>> }
>>
>> -static int tpm_cmd_ready(struct tpm_chip *chip)
>> -{
>> - if (!chip->ops->cmd_ready)
>> - return 0;
>> -
>> - return chip->ops->cmd_ready(chip);
>> -}
>> -
>> -static int tpm_go_idle(struct tpm_chip *chip)
>> -{
>> - if (!chip->ops->go_idle)
>> - return 0;
>> -
>> - return chip->ops->go_idle(chip);
>> -}
>> -
>> static void tpm_clk_enable(struct tpm_chip *chip)
>> {
>> if (chip->ops->clk_enable)
>> @@ -116,13 +100,6 @@ int tpm_chip_start(struct tpm_chip *chip)
>> }
>> }
>>
>> - ret = tpm_cmd_ready(chip);
>> - if (ret) {
>> - tpm_relinquish_locality(chip);
>> - tpm_clk_disable(chip);
>> - return ret;
>> - }
>> -
>> return 0;
>> }
>> EXPORT_SYMBOL_GPL(tpm_chip_start);
>> @@ -137,7 +114,6 @@ EXPORT_SYMBOL_GPL(tpm_chip_start);
>> */
>> void tpm_chip_stop(struct tpm_chip *chip)
>> {
>> - tpm_go_idle(chip);
>> tpm_relinquish_locality(chip);
>> tpm_clk_disable(chip);
>> }
>> diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c
>> index f745a098908b3..c5b5b5ba41821 100644
>> --- a/drivers/char/tpm/tpm-interface.c
>> +++ b/drivers/char/tpm/tpm-interface.c
>> @@ -113,12 +113,21 @@ static ssize_t tpm_try_transmit(struct tpm_chip *chip, void *buf, size_t bufsiz)
>> return -E2BIG;
>> }
>>
>> + if (chip->ops->cmd_ready) {
>> + rc = chip->ops->cmd_ready(chip);
>> + if (rc) {
>> + dev_err(&chip->dev,
>> + "%s: cmd_ready(): error %d\n", __func__, rc);
>> + return rc;
>> + }
>> + }
>
> I'd consider retaining tpm_go_idle and do this after arming the command:
>
> chip = chip __free(tpm_go_idle);
>
Good idea; that would be a lot less cleanup changes. I'll modify and test.
>> +
>> rc = chip->ops->send(chip, buf, bufsiz, count);
>> if (rc < 0) {
>> if (rc != -EPIPE)
>> dev_err(&chip->dev,
>> "%s: send(): error %d\n", __func__, rc);
>> - return rc;
>> + goto out;
>> }
>>
>> /*
>> @@ -152,7 +161,8 @@ static ssize_t tpm_try_transmit(struct tpm_chip *chip, void *buf, size_t bufsiz)
>>
>> if (tpm_chip_req_canceled(chip, status)) {
>> dev_err(&chip->dev, "Operation Canceled\n");
>> - return -ECANCELED;
>> + rc = -ECANCELED;
>> + goto out;
>> }
>>
>> tpm_msleep(TPM_TIMEOUT_POLL);
>> @@ -168,19 +178,27 @@ static ssize_t tpm_try_transmit(struct tpm_chip *chip, void *buf, size_t bufsiz)
>>
>> tpm_chip_cancel(chip);
>> dev_err(&chip->dev, "Operation Timed out\n");
>> - return -ETIME;
>> + rc = -ETIME;
>> + goto out;
>>
>> out_recv:
>> len = chip->ops->recv(chip, buf, bufsiz);
>> if (len < 0) {
>> rc = len;
>> dev_err(&chip->dev, "tpm_transmit: tpm_recv: error %d\n", rc);
>> - return rc;
>> + goto out;
>> }
>> out_sync:
>> if (len < TPM_HEADER_SIZE || len != be32_to_cpu(header->length))
>> rc = -EFAULT;
>>
>> +out:
>> + if (chip->ops->go_idle) {
>> + int idle_rc = chip->ops->go_idle(chip);
>> + if (idle_rc && !rc)
>> + rc = idle_rc;
>> + }
>> +
>> return rc ? rc : len;
>> }
>>
>> --
>> 2.43.0
>>
>
> BR, Jarkko
prev parent reply other threads:[~2026-09-01 17:54 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-01 16:32 [PATCH v2] tpm: Call cmd_ready/go_idle for each command transmission Mario Limonciello
2026-09-01 17:37 ` Jarkko Sakkinen
2026-09-01 17:54 ` Mario Limonciello [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=fb4affaa-0093-4a5d-9591-63700b46e64d@amd.com \
--to=mario.limonciello@amd.com \
--cc=jarkko@kernel.org \
--cc=jgg@ziepe.ca \
--cc=linux-integrity@vger.kernel.org \
--cc=peterhuewe@gmx.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox