diff mbox series

[v6,1/8] crypto: ccp: Abort doing SEV INIT if SNP INIT fails

Message ID 9d8cae623934489b46dc5abdf65a3034800351d9.1741300901.git.ashish.kalra@amd.com (mailing list archive)
State New
Headers show
Series Move initializing SEV/SNP functionality to KVM | expand

Commit Message

Kalra, Ashish March 6, 2025, 11:09 p.m. UTC 
From: Ashish Kalra <ashish.kalra@amd.com>

If SNP host support (SYSCFG.SNPEn) is set, then RMP table must be
initialized up before calling SEV INIT.

In other words, if SNP_INIT(_EX) is not issued or fails then
SEV INIT will fail once SNP host support (SYSCFG.SNPEn) is enabled.

Fixes: 1ca5614b84eed ("crypto: ccp: Add support to initialize the AMD-SP for SEV-SNP")
Signed-off-by: Ashish Kalra <ashish.kalra@amd.com>
---
 drivers/crypto/ccp/sev-dev.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

Comments

Tom Lendacky March 7, 2025, 8:54 p.m. UTC | #1 
On 3/6/25 17:09, Ashish Kalra wrote:
> From: Ashish Kalra <ashish.kalra@amd.com>
> 
> If SNP host support (SYSCFG.SNPEn) is set, then RMP table must be

s/RMP/the RMP/

> initialized up before calling SEV INIT.

s/up//

> 
> In other words, if SNP_INIT(_EX) is not issued or fails then
> SEV INIT will fail once SNP host support (SYSCFG.SNPEn) is enabled.

s/once/if/

> 
> Fixes: 1ca5614b84eed ("crypto: ccp: Add support to initialize the AMD-SP for SEV-SNP")
> Signed-off-by: Ashish Kalra <ashish.kalra@amd.com>
> ---
>  drivers/crypto/ccp/sev-dev.c | 7 ++-----
>  1 file changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
> index 2e87ca0e292a..a0e3de94704e 100644
> --- a/drivers/crypto/ccp/sev-dev.c
> +++ b/drivers/crypto/ccp/sev-dev.c
> @@ -1112,7 +1112,7 @@ static int __sev_snp_init_locked(int *error)
>  	if (!sev_version_greater_or_equal(SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR)) {
>  		dev_dbg(sev->dev, "SEV-SNP support requires firmware version >= %d:%d\n",
>  			SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR);
> -		return 0;
> +		return -EOPNOTSUPP;
>  	}
>  
>  	/* SNP_INIT requires MSR_VM_HSAVE_PA to be cleared on all CPUs. */
> @@ -1325,12 +1325,9 @@ static int _sev_platform_init_locked(struct sev_platform_init_args *args)
>  	 */
>  	rc = __sev_snp_init_locked(&args->error);
>  	if (rc && rc != -ENODEV) {

Can we get ride of this extra -ENODEV check? It can only be returned
because of the same check that is made earlier in this function so it
doesn't really serve any purpose from what I can tell.

Just make this "if (rc) {"

Thanks,
Tom

> -		/*
> -		 * Don't abort the probe if SNP INIT failed,
> -		 * continue to initialize the legacy SEV firmware.
> -		 */
>  		dev_err(sev->dev, "SEV-SNP: failed to INIT rc %d, error %#x\n",
>  			rc, args->error);
> +		return rc;
>  	}
>  
>  	/* Defer legacy SEV/SEV-ES support if allowed by caller/module. */
Tom Lendacky March 7, 2025, 8:57 p.m. UTC | #2 
On 3/7/25 14:54, Tom Lendacky wrote:
> On 3/6/25 17:09, Ashish Kalra wrote:
>> From: Ashish Kalra <ashish.kalra@amd.com>
>>
>> If SNP host support (SYSCFG.SNPEn) is set, then RMP table must be
> 
> s/RMP/the RMP/
> 
>> initialized up before calling SEV INIT.
> 
> s/up//
> 
>>
>> In other words, if SNP_INIT(_EX) is not issued or fails then
>> SEV INIT will fail once SNP host support (SYSCFG.SNPEn) is enabled.
> 
> s/once/if/
> 
>>
>> Fixes: 1ca5614b84eed ("crypto: ccp: Add support to initialize the AMD-SP for SEV-SNP")
>> Signed-off-by: Ashish Kalra <ashish.kalra@amd.com>
>> ---
>>  drivers/crypto/ccp/sev-dev.c | 7 ++-----
>>  1 file changed, 2 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
>> index 2e87ca0e292a..a0e3de94704e 100644
>> --- a/drivers/crypto/ccp/sev-dev.c
>> +++ b/drivers/crypto/ccp/sev-dev.c
>> @@ -1112,7 +1112,7 @@ static int __sev_snp_init_locked(int *error)
>>  	if (!sev_version_greater_or_equal(SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR)) {
>>  		dev_dbg(sev->dev, "SEV-SNP support requires firmware version >= %d:%d\n",
>>  			SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR);
>> -		return 0;
>> +		return -EOPNOTSUPP;
>>  	}
>>  
>>  	/* SNP_INIT requires MSR_VM_HSAVE_PA to be cleared on all CPUs. */
>> @@ -1325,12 +1325,9 @@ static int _sev_platform_init_locked(struct sev_platform_init_args *args)
>>  	 */
>>  	rc = __sev_snp_init_locked(&args->error);
>>  	if (rc && rc != -ENODEV) {
> 
> Can we get ride of this extra -ENODEV check? It can only be returned
> because of the same check that is made earlier in this function so it
> doesn't really serve any purpose from what I can tell.
> 
> Just make this "if (rc) {"

My bad... -ENODEV is returned if cc_platform_has(CC_ATTR_HOST_SEV_SNP) is
false, never mind.

Thanks,
Tom

> 
> Thanks,
> Tom
> 
>> -		/*
>> -		 * Don't abort the probe if SNP INIT failed,
>> -		 * continue to initialize the legacy SEV firmware.
>> -		 */
>>  		dev_err(sev->dev, "SEV-SNP: failed to INIT rc %d, error %#x\n",
>>  			rc, args->error);
>> +		return rc;
>>  	}
>>  
>>  	/* Defer legacy SEV/SEV-ES support if allowed by caller/module. */
Kalra, Ashish March 7, 2025, 9:06 p.m. UTC | #3 
On 3/7/2025 2:57 PM, Tom Lendacky wrote:
> On 3/7/25 14:54, Tom Lendacky wrote:
>> On 3/6/25 17:09, Ashish Kalra wrote:
>>> From: Ashish Kalra <ashish.kalra@amd.com>
>>>
>>> If SNP host support (SYSCFG.SNPEn) is set, then RMP table must be
>>
>> s/RMP/the RMP/
>>
>>> initialized up before calling SEV INIT.
>>
>> s/up//
>>
>>>
>>> In other words, if SNP_INIT(_EX) is not issued or fails then
>>> SEV INIT will fail once SNP host support (SYSCFG.SNPEn) is enabled.
>>
>> s/once/if/
>>
>>>
>>> Fixes: 1ca5614b84eed ("crypto: ccp: Add support to initialize the AMD-SP for SEV-SNP")
>>> Signed-off-by: Ashish Kalra <ashish.kalra@amd.com>
>>> ---
>>>  drivers/crypto/ccp/sev-dev.c | 7 ++-----
>>>  1 file changed, 2 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
>>> index 2e87ca0e292a..a0e3de94704e 100644
>>> --- a/drivers/crypto/ccp/sev-dev.c
>>> +++ b/drivers/crypto/ccp/sev-dev.c
>>> @@ -1112,7 +1112,7 @@ static int __sev_snp_init_locked(int *error)
>>>  	if (!sev_version_greater_or_equal(SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR)) {
>>>  		dev_dbg(sev->dev, "SEV-SNP support requires firmware version >= %d:%d\n",
>>>  			SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR);
>>> -		return 0;
>>> +		return -EOPNOTSUPP;
>>>  	}
>>>  
>>>  	/* SNP_INIT requires MSR_VM_HSAVE_PA to be cleared on all CPUs. */
>>> @@ -1325,12 +1325,9 @@ static int _sev_platform_init_locked(struct sev_platform_init_args *args)
>>>  	 */
>>>  	rc = __sev_snp_init_locked(&args->error);
>>>  	if (rc && rc != -ENODEV) {
>>
>> Can we get ride of this extra -ENODEV check? It can only be returned
>> because of the same check that is made earlier in this function so it
>> doesn't really serve any purpose from what I can tell.
>>
>> Just make this "if (rc) {"
> 
> My bad... -ENODEV is returned if cc_platform_has(CC_ATTR_HOST_SEV_SNP) is
> false, never mind.

Yes, that's what i was going to reply with ... we want to continue with
SEV INIT if SNP host support is not enabled.

Thanks,
Ashish

> 
> Thanks,
> Tom
> 
>>
>> Thanks,
>> Tom
>>
>>> -		/*
>>> -		 * Don't abort the probe if SNP INIT failed,
>>> -		 * continue to initialize the legacy SEV firmware.
>>> -		 */
>>>  		dev_err(sev->dev, "SEV-SNP: failed to INIT rc %d, error %#x\n",
>>>  			rc, args->error);
>>> +		return rc;
>>>  	}
>>>  
>>>  	/* Defer legacy SEV/SEV-ES support if allowed by caller/module. */
Tom Lendacky March 7, 2025, 9:29 p.m. UTC | #4 
On 3/7/25 15:06, Kalra, Ashish wrote:
> On 3/7/2025 2:57 PM, Tom Lendacky wrote:
>> On 3/7/25 14:54, Tom Lendacky wrote:
>>> On 3/6/25 17:09, Ashish Kalra wrote:
>>>> From: Ashish Kalra <ashish.kalra@amd.com>
>>>>
>>>> If SNP host support (SYSCFG.SNPEn) is set, then RMP table must be
>>>
>>> s/RMP/the RMP/
>>>
>>>> initialized up before calling SEV INIT.
>>>
>>> s/up//
>>>
>>>>
>>>> In other words, if SNP_INIT(_EX) is not issued or fails then
>>>> SEV INIT will fail once SNP host support (SYSCFG.SNPEn) is enabled.
>>>
>>> s/once/if/
>>>
>>>>
>>>> Fixes: 1ca5614b84eed ("crypto: ccp: Add support to initialize the AMD-SP for SEV-SNP")
>>>> Signed-off-by: Ashish Kalra <ashish.kalra@amd.com>
>>>> ---
>>>>  drivers/crypto/ccp/sev-dev.c | 7 ++-----
>>>>  1 file changed, 2 insertions(+), 5 deletions(-)
>>>>
>>>> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
>>>> index 2e87ca0e292a..a0e3de94704e 100644
>>>> --- a/drivers/crypto/ccp/sev-dev.c
>>>> +++ b/drivers/crypto/ccp/sev-dev.c
>>>> @@ -1112,7 +1112,7 @@ static int __sev_snp_init_locked(int *error)
>>>>  	if (!sev_version_greater_or_equal(SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR)) {
>>>>  		dev_dbg(sev->dev, "SEV-SNP support requires firmware version >= %d:%d\n",
>>>>  			SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR);
>>>> -		return 0;
>>>> +		return -EOPNOTSUPP;
>>>>  	}
>>>>  
>>>>  	/* SNP_INIT requires MSR_VM_HSAVE_PA to be cleared on all CPUs. */
>>>> @@ -1325,12 +1325,9 @@ static int _sev_platform_init_locked(struct sev_platform_init_args *args)
>>>>  	 */
>>>>  	rc = __sev_snp_init_locked(&args->error);
>>>>  	if (rc && rc != -ENODEV) {
>>>
>>> Can we get ride of this extra -ENODEV check? It can only be returned
>>> because of the same check that is made earlier in this function so it
>>> doesn't really serve any purpose from what I can tell.
>>>
>>> Just make this "if (rc) {"
>>
>> My bad... -ENODEV is returned if cc_platform_has(CC_ATTR_HOST_SEV_SNP) is
>> false, never mind.
> 
> Yes, that's what i was going to reply with ... we want to continue with
> SEV INIT if SNP host support is not enabled.

Although we could get rid of that awkward if statement by doing...

	if (cc_platform_has(CC_ATTR_HOST_SEV_SNP)) {
		rc = __sev_snp_init_locked(&args->error);
		if (rc) {
			dev_err(sev->dev, ...
			return rc;
		}
	}

And deleting the cc_platform_has() check from __sev_snp_init_locked().

Thanks,
Tom

> 
> Thanks,
> Ashish
> 
>>
>> Thanks,
>> Tom
>>
>>>
>>> Thanks,
>>> Tom
>>>
>>>> -		/*
>>>> -		 * Don't abort the probe if SNP INIT failed,
>>>> -		 * continue to initialize the legacy SEV firmware.
>>>> -		 */
>>>>  		dev_err(sev->dev, "SEV-SNP: failed to INIT rc %d, error %#x\n",
>>>>  			rc, args->error);
>>>> +		return rc;
>>>>  	}
>>>>  
>>>>  	/* Defer legacy SEV/SEV-ES support if allowed by caller/module. */
>
Kalra, Ashish March 7, 2025, 9:39 p.m. UTC | #5 
On 3/7/2025 3:29 PM, Tom Lendacky wrote:
> On 3/7/25 15:06, Kalra, Ashish wrote:
>> On 3/7/2025 2:57 PM, Tom Lendacky wrote:
>>> On 3/7/25 14:54, Tom Lendacky wrote:
>>>> On 3/6/25 17:09, Ashish Kalra wrote:
>>>>> From: Ashish Kalra <ashish.kalra@amd.com>
>>>>>
>>>>> If SNP host support (SYSCFG.SNPEn) is set, then RMP table must be
>>>>
>>>> s/RMP/the RMP/
>>>>
>>>>> initialized up before calling SEV INIT.
>>>>
>>>> s/up//
>>>>
>>>>>
>>>>> In other words, if SNP_INIT(_EX) is not issued or fails then
>>>>> SEV INIT will fail once SNP host support (SYSCFG.SNPEn) is enabled.
>>>>
>>>> s/once/if/
>>>>
>>>>>
>>>>> Fixes: 1ca5614b84eed ("crypto: ccp: Add support to initialize the AMD-SP for SEV-SNP")
>>>>> Signed-off-by: Ashish Kalra <ashish.kalra@amd.com>
>>>>> ---
>>>>>  drivers/crypto/ccp/sev-dev.c | 7 ++-----
>>>>>  1 file changed, 2 insertions(+), 5 deletions(-)
>>>>>
>>>>> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
>>>>> index 2e87ca0e292a..a0e3de94704e 100644
>>>>> --- a/drivers/crypto/ccp/sev-dev.c
>>>>> +++ b/drivers/crypto/ccp/sev-dev.c
>>>>> @@ -1112,7 +1112,7 @@ static int __sev_snp_init_locked(int *error)
>>>>>  	if (!sev_version_greater_or_equal(SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR)) {
>>>>>  		dev_dbg(sev->dev, "SEV-SNP support requires firmware version >= %d:%d\n",
>>>>>  			SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR);
>>>>> -		return 0;
>>>>> +		return -EOPNOTSUPP;
>>>>>  	}
>>>>>  
>>>>>  	/* SNP_INIT requires MSR_VM_HSAVE_PA to be cleared on all CPUs. */
>>>>> @@ -1325,12 +1325,9 @@ static int _sev_platform_init_locked(struct sev_platform_init_args *args)
>>>>>  	 */
>>>>>  	rc = __sev_snp_init_locked(&args->error);
>>>>>  	if (rc && rc != -ENODEV) {
>>>>
>>>> Can we get ride of this extra -ENODEV check? It can only be returned
>>>> because of the same check that is made earlier in this function so it
>>>> doesn't really serve any purpose from what I can tell.
>>>>
>>>> Just make this "if (rc) {"
>>>
>>> My bad... -ENODEV is returned if cc_platform_has(CC_ATTR_HOST_SEV_SNP) is
>>> false, never mind.
>>
>> Yes, that's what i was going to reply with ... we want to continue with
>> SEV INIT if SNP host support is not enabled.
> 
> Although we could get rid of that awkward if statement by doing...
> 
> 	if (cc_platform_has(CC_ATTR_HOST_SEV_SNP)) {
> 		rc = __sev_snp_init_locked(&args->error);
> 		if (rc) {
> 			dev_err(sev->dev, ...
> 			return rc;
> 		}
> 	}
> 
> And deleting the cc_platform_has() check from __sev_snp_init_locked().

We probably have to keep the cc_platform_has() check in
__sev_snp_init_locked() for the implicit SNP INIT being issued
for various SNP specific ioctl's (via snp_move_to_init_state()).

Thanks,
Ashish

>>>>
>>>>> -		/*
>>>>> -		 * Don't abort the probe if SNP INIT failed,
>>>>> -		 * continue to initialize the legacy SEV firmware.
>>>>> -		 */
>>>>>  		dev_err(sev->dev, "SEV-SNP: failed to INIT rc %d, error %#x\n",
>>>>>  			rc, args->error);
>>>>> +		return rc;
>>>>>  	}
>>>>>  
>>>>>  	/* Defer legacy SEV/SEV-ES support if allowed by caller/module. */
>>
Tom Lendacky March 7, 2025, 9:43 p.m. UTC | #6 
On 3/7/25 15:39, Kalra, Ashish wrote:
> On 3/7/2025 3:29 PM, Tom Lendacky wrote:
>> On 3/7/25 15:06, Kalra, Ashish wrote:
>>> On 3/7/2025 2:57 PM, Tom Lendacky wrote:
>>>> On 3/7/25 14:54, Tom Lendacky wrote:
>>>>> On 3/6/25 17:09, Ashish Kalra wrote:
>>>>>> From: Ashish Kalra <ashish.kalra@amd.com>
>>>>>>
>>>>>> If SNP host support (SYSCFG.SNPEn) is set, then RMP table must be
>>>>>
>>>>> s/RMP/the RMP/
>>>>>
>>>>>> initialized up before calling SEV INIT.
>>>>>
>>>>> s/up//
>>>>>
>>>>>>
>>>>>> In other words, if SNP_INIT(_EX) is not issued or fails then
>>>>>> SEV INIT will fail once SNP host support (SYSCFG.SNPEn) is enabled.
>>>>>
>>>>> s/once/if/
>>>>>
>>>>>>
>>>>>> Fixes: 1ca5614b84eed ("crypto: ccp: Add support to initialize the AMD-SP for SEV-SNP")
>>>>>> Signed-off-by: Ashish Kalra <ashish.kalra@amd.com>
>>>>>> ---
>>>>>>  drivers/crypto/ccp/sev-dev.c | 7 ++-----
>>>>>>  1 file changed, 2 insertions(+), 5 deletions(-)
>>>>>>
>>>>>> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
>>>>>> index 2e87ca0e292a..a0e3de94704e 100644
>>>>>> --- a/drivers/crypto/ccp/sev-dev.c
>>>>>> +++ b/drivers/crypto/ccp/sev-dev.c
>>>>>> @@ -1112,7 +1112,7 @@ static int __sev_snp_init_locked(int *error)
>>>>>>  	if (!sev_version_greater_or_equal(SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR)) {
>>>>>>  		dev_dbg(sev->dev, "SEV-SNP support requires firmware version >= %d:%d\n",
>>>>>>  			SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR);
>>>>>> -		return 0;
>>>>>> +		return -EOPNOTSUPP;
>>>>>>  	}
>>>>>>  
>>>>>>  	/* SNP_INIT requires MSR_VM_HSAVE_PA to be cleared on all CPUs. */
>>>>>> @@ -1325,12 +1325,9 @@ static int _sev_platform_init_locked(struct sev_platform_init_args *args)
>>>>>>  	 */
>>>>>>  	rc = __sev_snp_init_locked(&args->error);
>>>>>>  	if (rc && rc != -ENODEV) {
>>>>>
>>>>> Can we get ride of this extra -ENODEV check? It can only be returned
>>>>> because of the same check that is made earlier in this function so it
>>>>> doesn't really serve any purpose from what I can tell.
>>>>>
>>>>> Just make this "if (rc) {"
>>>>
>>>> My bad... -ENODEV is returned if cc_platform_has(CC_ATTR_HOST_SEV_SNP) is
>>>> false, never mind.
>>>
>>> Yes, that's what i was going to reply with ... we want to continue with
>>> SEV INIT if SNP host support is not enabled.
>>
>> Although we could get rid of that awkward if statement by doing...
>>
>> 	if (cc_platform_has(CC_ATTR_HOST_SEV_SNP)) {
>> 		rc = __sev_snp_init_locked(&args->error);
>> 		if (rc) {
>> 			dev_err(sev->dev, ...
>> 			return rc;
>> 		}
>> 	}
>>
>> And deleting the cc_platform_has() check from __sev_snp_init_locked().
> 
> We probably have to keep the cc_platform_has() check in
> __sev_snp_init_locked() for the implicit SNP INIT being issued
> for various SNP specific ioctl's (via snp_move_to_init_state()).

Oh, yeah, the very next patch calls into __sev_snp_init_locked() now.

Thanks,
Tom

> 
> Thanks,
> Ashish
> 
>>>>>
>>>>>> -		/*
>>>>>> -		 * Don't abort the probe if SNP INIT failed,
>>>>>> -		 * continue to initialize the legacy SEV firmware.
>>>>>> -		 */
>>>>>>  		dev_err(sev->dev, "SEV-SNP: failed to INIT rc %d, error %#x\n",
>>>>>>  			rc, args->error);
>>>>>> +		return rc;
>>>>>>  	}
>>>>>>  
>>>>>>  	/* Defer legacy SEV/SEV-ES support if allowed by caller/module. */
>>>
>
diff mbox series

Patch

diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
index 2e87ca0e292a..a0e3de94704e 100644
--- a/drivers/crypto/ccp/sev-dev.c
+++ b/drivers/crypto/ccp/sev-dev.c
@@ -1112,7 +1112,7 @@  static int __sev_snp_init_locked(int *error)
 	if (!sev_version_greater_or_equal(SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR)) {
 		dev_dbg(sev->dev, "SEV-SNP support requires firmware version >= %d:%d\n",
 			SNP_MIN_API_MAJOR, SNP_MIN_API_MINOR);
-		return 0;
+		return -EOPNOTSUPP;
 	}
 
 	/* SNP_INIT requires MSR_VM_HSAVE_PA to be cleared on all CPUs. */
@@ -1325,12 +1325,9 @@  static int _sev_platform_init_locked(struct sev_platform_init_args *args)
 	 */
 	rc = __sev_snp_init_locked(&args->error);
 	if (rc && rc != -ENODEV) {
-		/*
-		 * Don't abort the probe if SNP INIT failed,
-		 * continue to initialize the legacy SEV firmware.
-		 */
 		dev_err(sev->dev, "SEV-SNP: failed to INIT rc %d, error %#x\n",
 			rc, args->error);
+		return rc;
 	}
 
 	/* Defer legacy SEV/SEV-ES support if allowed by caller/module. */