From: Tom Lendacky <thomas.lendacky@amd.com>
To: "Mario Limonciello" <superm1@kernel.org>,
"Herbert Xu" <herbert@gondor.apana.org.au>,
"Shyam Sundar S K" <Shyam-sundar.S-k@amd.com>,
"Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>,
"Rijo Thomas" <Rijo-john.Thomas@amd.com>
Cc: John Allen <john.allen@amd.com>,
"David S . Miller" <davem@davemloft.net>,
"open list:AMD CRYPTOGRAPHIC COPROCESSOR (CCP) DRIVER"
<linux-crypto@vger.kernel.org>, Hans de Goede <hansg@kernel.org>,
"open list:AMD PMF DRIVER" <platform-driver-x86@vger.kernel.org>,
Lars Francke <lars.francke@gmail.com>,
Yijun Shen <Yijun.Shen@dell.com>,
Devaraj Rangasamy <Devaraj.Rangasamy@amd.com>
Subject: Re: [PATCH v5 3/5] crypto: ccp - Add an S4 restore flow
Date: Mon, 12 Jan 2026 16:06:20 -0600 [thread overview]
Message-ID: <4505efb0-d44c-4fe8-b10a-cf4c6f9b17a4@amd.com> (raw)
In-Reply-To: <20acfb14-02a0-4f7a-9924-fa70fb485418@kernel.org>
On 1/12/26 15:25, Mario Limonciello wrote:
> On 1/12/26 10:33 AM, Tom Lendacky wrote:
>> On 1/11/26 23:36, Mario Limonciello (AMD) wrote:
>>> The system will have lost power during S4. The ring used for TEE
>>> communications needs to be initialized before use.
>>>
>>> Fixes: f892a21f51162 ("crypto: ccp - use generic power management")
>>> Reported-by: Lars Francke <lars.francke@gmail.com>
>>> Closes: https://lore.kernel.org/platform-driver-x86/CAD-
>>> Ua_gfJnQSo8ucS_7ZwzuhoBRJ14zXP7s8b-zX3ZcxcyWePw@mail.gmail.com/
>>> Tested-by: Yijun Shen <Yijun.Shen@Dell.com>
>>> Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
>>> ---
>>> v5:
>>> * Fix return for psp_restore() (LKP robot, Shyam)
>>> v4:
>>> * Add tag (Yijun)
>>> * Remove blank line (Ilpo)
>>> * Adjust whitespace in psp_restore() (Ilpo)
>>> ---
>>> drivers/crypto/ccp/psp-dev.c | 11 +++++++++++
>>> drivers/crypto/ccp/sp-dev.h | 2 ++
>>> drivers/crypto/ccp/sp-pci.c | 24 +++++++++++++++++++++++-
>>> drivers/crypto/ccp/tee-dev.c | 5 +++++
>>> drivers/crypto/ccp/tee-dev.h | 1 +
>>> 5 files changed, 42 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/crypto/ccp/psp-dev.c b/drivers/crypto/ccp/psp-dev.c
>>> index 9e21da0e298ad..5c7f7e02a7d8a 100644
>>> --- a/drivers/crypto/ccp/psp-dev.c
>>> +++ b/drivers/crypto/ccp/psp-dev.c
>>> @@ -351,6 +351,17 @@ struct psp_device *psp_get_master_device(void)
>>> return sp ? sp->psp_data : NULL;
>>> }
>>> +int psp_restore(struct sp_device *sp)
>>> +{
>>> + struct psp_device *psp = sp->psp_data;
>>> + int ret = 0;
>>> +
>>> + if (psp->tee_data)
>>> + ret = tee_restore(psp);
>>> +
>>> + return ret;
>>> +}
>>> +
>>> void psp_pci_init(void)
>>> {
>>> psp_master = psp_get_master_device();
>>> diff --git a/drivers/crypto/ccp/sp-dev.h b/drivers/crypto/ccp/sp-dev.h
>>> index 1335a83fe052e..0deea1a399e47 100644
>>> --- a/drivers/crypto/ccp/sp-dev.h
>>> +++ b/drivers/crypto/ccp/sp-dev.h
>>> @@ -174,6 +174,7 @@ int psp_dev_init(struct sp_device *sp);
>>> void psp_pci_init(void);
>>> void psp_dev_destroy(struct sp_device *sp);
>>> void psp_pci_exit(void);
>>> +int psp_restore(struct sp_device *sp);
>>> #else /* !CONFIG_CRYPTO_DEV_SP_PSP */
>>> @@ -181,6 +182,7 @@ static inline int psp_dev_init(struct sp_device
>>> *sp) { return 0; }
>>> static inline void psp_pci_init(void) { }
>>> static inline void psp_dev_destroy(struct sp_device *sp) { }
>>> static inline void psp_pci_exit(void) { }
>>> +static inline int psp_restore(struct sp_device *sp) { return 0; }
>>> #endif /* CONFIG_CRYPTO_DEV_SP_PSP */
>>> diff --git a/drivers/crypto/ccp/sp-pci.c b/drivers/crypto/ccp/sp-pci.c
>>> index 8891ceee1d7d0..931ec6bf2cec6 100644
>>> --- a/drivers/crypto/ccp/sp-pci.c
>>> +++ b/drivers/crypto/ccp/sp-pci.c
>>> @@ -353,6 +353,21 @@ static int __maybe_unused sp_pci_resume(struct
>>> device *dev)
>>> return sp_resume(sp);
>>> }
>>> +static int __maybe_unused sp_pci_restore(struct device *dev)
>>> +{
>>> + struct sp_device *sp = dev_get_drvdata(dev);
>>> +
>>> +#ifdef CONFIG_CRYPTO_DEV_SP_PSP
>>
>> This isn't needed since you have the "#else" definition in sp-dev.h.
>
> Ack.
>
>>
>>> + if (sp->psp_data) {
>>> + int ret = psp_restore(sp);
>>> +
>>
>> Remove blank line.
>
> Then checkpatch complains:
>
> Missing a blank line after declarations
Ah, yeah, bummer...
>
>>
>>> + if (ret)
>>> + return ret;
>>> + }
>>> +#endif
>>
>> This isn't the way the other functions work. This function should just
>> be calling to sp_resume() and then sp_resume() should call psp_restore()
>> (or psp_dev_restore() ?) if the psp_data is present.
>>
>> And if you need to distinguish between the resume and restore case in
>> sp_resume(), then it sounds like you need an sp_restore().
>
> We do need a different path for resume and restore.
>
> I guess the call path you're looking for is:
>
> sp_pci_restore()
> -> sp_restore()
> ->-> psp_restore()
> ->-> sp_resume()
>
> Basically move the contents of sp_pci_restore() into a new sp_restore().
Right, without the #ifdef's in the .c file but in the .h file instead.
Thanks,
Tom
>
>>
>> Thanks,
>> Tom
>>
>>> + return sp_resume(sp);
>>> +}
>>> +
>>> #ifdef CONFIG_CRYPTO_DEV_SP_PSP
>>> static const struct sev_vdata sevv1 = {
>>> .cmdresp_reg = 0x10580, /* C2PMSG_32 */
>>> @@ -563,7 +578,14 @@ static const struct pci_device_id sp_pci_table[]
>>> = {
>>> };
>>> MODULE_DEVICE_TABLE(pci, sp_pci_table);
>>> -static SIMPLE_DEV_PM_OPS(sp_pci_pm_ops, sp_pci_suspend,
>>> sp_pci_resume);
>>> +static const struct dev_pm_ops sp_pci_pm_ops = {
>>> + .suspend = pm_sleep_ptr(sp_pci_suspend),
>>> + .resume = pm_sleep_ptr(sp_pci_resume),
>>> + .freeze = pm_sleep_ptr(sp_pci_suspend),
>>> + .thaw = pm_sleep_ptr(sp_pci_resume),
>>> + .poweroff = pm_sleep_ptr(sp_pci_suspend),
>>> + .restore_early = pm_sleep_ptr(sp_pci_restore),
>>> +};
>>> static struct pci_driver sp_pci_driver = {
>>> .name = "ccp",
>>> diff --git a/drivers/crypto/ccp/tee-dev.c b/drivers/crypto/ccp/tee-dev.c
>>> index af881daa5855b..11c4b05e2f3a2 100644
>>> --- a/drivers/crypto/ccp/tee-dev.c
>>> +++ b/drivers/crypto/ccp/tee-dev.c
>>> @@ -366,3 +366,8 @@ int psp_check_tee_status(void)
>>> return 0;
>>> }
>>> EXPORT_SYMBOL(psp_check_tee_status);
>>> +
>>> +int tee_restore(struct psp_device *psp)
>>> +{
>>> + return tee_init_ring(psp->tee_data);
>>> +}
>>> diff --git a/drivers/crypto/ccp/tee-dev.h b/drivers/crypto/ccp/tee-dev.h
>>> index ea9a2b7c05f57..c23416cb7bb37 100644
>>> --- a/drivers/crypto/ccp/tee-dev.h
>>> +++ b/drivers/crypto/ccp/tee-dev.h
>>> @@ -111,5 +111,6 @@ struct tee_ring_cmd {
>>> int tee_dev_init(struct psp_device *psp);
>>> void tee_dev_destroy(struct psp_device *psp);
>>> +int tee_restore(struct psp_device *psp);
>>> #endif /* __TEE_DEV_H__ */
>>
>>
>
next prev parent reply other threads:[~2026-01-12 22:06 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-01-12 5:36 [PATCH v5 0/5] Fixes for PMF and CCP drivers after S4 Mario Limonciello (AMD)
2026-01-12 5:36 ` [PATCH v5 1/5] platform/x86/amd/pmf: Prevent TEE errors after hibernate Mario Limonciello (AMD)
2026-01-12 5:36 ` [PATCH v5 2/5] crypto: ccp - Declare PSP dead if PSP_CMD_TEE_RING_INIT fails Mario Limonciello (AMD)
2026-01-12 5:36 ` [PATCH v5 3/5] crypto: ccp - Add an S4 restore flow Mario Limonciello (AMD)
2026-01-12 16:33 ` Tom Lendacky
2026-01-12 21:25 ` Mario Limonciello
2026-01-12 22:06 ` Tom Lendacky [this message]
2026-01-12 5:36 ` [PATCH v5 4/5] crypto: ccp - Factor out ring destroy handling to a helper Mario Limonciello (AMD)
2026-01-12 5:36 ` [PATCH v5 5/5] crypto: ccp - Send PSP_CMD_TEE_RING_DESTROY when PSP_CMD_TEE_RING_INIT fails Mario Limonciello (AMD)
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=4505efb0-d44c-4fe8-b10a-cf4c6f9b17a4@amd.com \
--to=thomas.lendacky@amd.com \
--cc=Devaraj.Rangasamy@amd.com \
--cc=Rijo-john.Thomas@amd.com \
--cc=Shyam-sundar.S-k@amd.com \
--cc=Yijun.Shen@dell.com \
--cc=davem@davemloft.net \
--cc=hansg@kernel.org \
--cc=herbert@gondor.apana.org.au \
--cc=ilpo.jarvinen@linux.intel.com \
--cc=john.allen@amd.com \
--cc=lars.francke@gmail.com \
--cc=linux-crypto@vger.kernel.org \
--cc=platform-driver-x86@vger.kernel.org \
--cc=superm1@kernel.org \
/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