Re: [PATCH kernel v2 4/5] crypto: ccp: Enable SEV-TIO feature in the PSP when supported
From: Tom Lendacky
Date: Mon Dec 01 2025 - 09:32:52 EST
On 11/21/25 02:06, Alexey Kardashevskiy wrote:
> The PSP advertises the SEV-TIO support via the FEATURE_INFO command
> advertised via SNP_PLATFORM_STATUS.
>
> The BIOS advertises the SEV-TIO enablement via the IOMMU EFR2 register
> (added in an earlier patch).
>
> Enable SEV-TIO during the SNP_INIT_EX call if both the PSP and the BIOS
> advertise support for it.
>
> Signed-off-by: Alexey Kardashevskiy <aik@xxxxxxx>
> ---
> include/linux/psp-sev.h | 4 +++-
> drivers/crypto/ccp/sev-dev.c | 10 +++++++++-
> 2 files changed, 12 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/psp-sev.h b/include/linux/psp-sev.h
> index 34a25209f909..c0c817ca3615 100644
> --- a/include/linux/psp-sev.h
> +++ b/include/linux/psp-sev.h
> @@ -750,7 +750,8 @@ struct sev_data_snp_init_ex {
> u32 list_paddr_en:1;
> u32 rapl_dis:1;
> u32 ciphertext_hiding_en:1;
> - u32 rsvd:28;
> + u32 tio_en:1;
> + u32 rsvd:27;
> u32 rsvd1;
> u64 list_paddr;
> u16 max_snp_asid;
> @@ -850,6 +851,7 @@ struct snp_feature_info {
> } __packed;
>
> #define SNP_CIPHER_TEXT_HIDING_SUPPORTED BIT(3)
> +#define SNP_SEV_TIO_SUPPORTED BIT(1) /* EBX */
>
> #ifdef CONFIG_CRYPTO_DEV_SP_PSP
>
> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
> index 9e0c16b36f9c..2f1c9614d359 100644
> --- a/drivers/crypto/ccp/sev-dev.c
> +++ b/drivers/crypto/ccp/sev-dev.c
> @@ -1358,6 +1358,11 @@ static int snp_filter_reserved_mem_regions(struct resource *rs, void *arg)
> return 0;
> }
>
> +static bool sev_tio_present(struct sev_device *sev)
> +{
> + return (sev->snp_feat_info_0.ebx & SNP_SEV_TIO_SUPPORTED) != 0;
> +}
> +
> static int __sev_snp_init_locked(int *error, unsigned int max_snp_asid)
> {
> struct psp_device *psp = psp_master;
> @@ -1434,6 +1439,8 @@ static int __sev_snp_init_locked(int *error, unsigned int max_snp_asid)
> data.init_rmp = 1;
> data.list_paddr_en = 1;
> data.list_paddr = __psp_pa(snp_range_list);
> + data.tio_en = sev_tio_present(sev) &&
> + amd_iommu_sev_tio_supported();
Align amd_iommu_sev_tio_supported() with sev_tio_present() or just make
a single line. But...
I would move the IOMMU check into the sev_tio_present() function and
then rename sev_tio_present() to sev_tio_supported() unless
sev_tio_present() is going to be used more somewhere else?
Those can be follow-up patches, though, if this needs to get into 6.19.
Thanks,
Tom
> cmd = SEV_CMD_SNP_INIT_EX;
> } else {
> cmd = SEV_CMD_SNP_INIT;
> @@ -1471,7 +1478,8 @@ static int __sev_snp_init_locked(int *error, unsigned int max_snp_asid)
>
> snp_hv_fixed_pages_state_update(sev, HV_FIXED);
> sev->snp_initialized = true;
> - dev_dbg(sev->dev, "SEV-SNP firmware initialized\n");
> + dev_dbg(sev->dev, "SEV-SNP firmware initialized, SEV-TIO is %s\n",
> + data.tio_en ? "enabled" : "disabled");
>
> dev_info(sev->dev, "SEV-SNP API:%d.%d build:%d\n", sev->api_major,
> sev->api_minor, sev->build);