Re: [PATCH v5 05/43] arm64: RME: Add SMC definitions for calling the RMM

[Steven Price <steven.price@arm.com>]

On 25/10/2024 07:37, Gavin Shan wrote:
> On 10/5/24 1:27 AM, Steven Price wrote:
>> The RMM (Realm Management Monitor) provides functionality that can be
>> accessed by SMC calls from the host.
>>
>> The SMC definitions are based on DEN0137[1] version 1.0-rel0
>>
>> [1] https://developer.arm.com/documentation/den0137/1-0rel0/
>>
>> Signed-off-by: Steven Price <steven.price@arm.com>
>> ---
>> Changes since v4:
>>   * Update to point to final released RMM spec.
>>   * Minor rearrangements.
>> Changes since v3:
>>   * Update to match RMM spec v1.0-rel0-rc1.
>> Changes since v2:
>>   * Fix specification link.
>>   * Rename rec_entry->rec_enter to match spec.
>>   * Fix size of pmu_ovf_status to match spec.
>> ---
>>   arch/arm64/include/asm/rmi_smc.h | 255 +++++++++++++++++++++++++++++++
>>   1 file changed, 255 insertions(+)
>>   create mode 100644 arch/arm64/include/asm/rmi_smc.h
>>
>> diff --git a/arch/arm64/include/asm/rmi_smc.h
>> b/arch/arm64/include/asm/rmi_smc.h
>> new file mode 100644
>> index 000000000000..0fde2e06d275
>> --- /dev/null
>> +++ b/arch/arm64/include/asm/rmi_smc.h
>> @@ -0,0 +1,255 @@
>> +/* SPDX-License-Identifier: GPL-2.0 */
>> +/*
>> + * Copyright (C) 2023-2024 ARM Ltd.
>> + *
>> + * The values and structures in this file are from the Realm
>> Management Monitor
>> + * specification (DEN0137) version 1.0-rel0:
>> + * https://developer.arm.com/documentation/den0137/1-0rel0/
>> + */
>> +
>> +#ifndef __ASM_RME_SMC_H
>> +#define __ASM_RME_SMC_H
>> +
>> +#include <linux/arm-smccc.h>
>> +
>> +#define SMC_RxI_CALL(func)                \
>> +    ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,        \
>> +               ARM_SMCCC_SMC_64,        \
>> +               ARM_SMCCC_OWNER_STANDARD,    \
>> +               (func))
>> +
> 
> I guess the 'x' of 'RxI' here can be 'M' or 'S'. We already had similar
> macro
> (SMC_RSI_FID) in rsi_smc.h, so 'RMI' sounds more appropriate to me since
> this
> macro is only used to define those RMI function calls. SMC_RMI_FID is
> the name
> consistent to SMC_RSI_FID in rsi_smc.h.

Yeah, considering this is only in the R*M*I header I'll rename to
SMC_RMI_CALL().

>> +#define SMC_RMI_DATA_CREATE        SMC_RxI_CALL(0x0153)
>> +#define SMC_RMI_DATA_CREATE_UNKNOWN    SMC_RxI_CALL(0x0154)
>> +#define SMC_RMI_DATA_DESTROY        SMC_RxI_CALL(0x0155)
>> +#define SMC_RMI_FEATURES        SMC_RxI_CALL(0x0165)
>> +#define SMC_RMI_GRANULE_DELEGATE    SMC_RxI_CALL(0x0151)
>> +#define SMC_RMI_GRANULE_UNDELEGATE    SMC_RxI_CALL(0x0152)
>> +#define SMC_RMI_PSCI_COMPLETE        SMC_RxI_CALL(0x0164)
>> +#define SMC_RMI_REALM_ACTIVATE        SMC_RxI_CALL(0x0157)
>> +#define SMC_RMI_REALM_CREATE        SMC_RxI_CALL(0x0158)
>> +#define SMC_RMI_REALM_DESTROY        SMC_RxI_CALL(0x0159)
>> +#define SMC_RMI_REC_AUX_COUNT        SMC_RxI_CALL(0x0167)
>> +#define SMC_RMI_REC_CREATE        SMC_RxI_CALL(0x015a)
>> +#define SMC_RMI_REC_DESTROY        SMC_RxI_CALL(0x015b)
>> +#define SMC_RMI_REC_ENTER        SMC_RxI_CALL(0x015c)
>> +#define SMC_RMI_RTT_CREATE        SMC_RxI_CALL(0x015d)
>> +#define SMC_RMI_RTT_DESTROY        SMC_RxI_CALL(0x015e)
>> +#define SMC_RMI_RTT_FOLD        SMC_RxI_CALL(0x0166)
>> +#define SMC_RMI_RTT_INIT_RIPAS        SMC_RxI_CALL(0x0168)
>> +#define SMC_RMI_RTT_MAP_UNPROTECTED    SMC_RxI_CALL(0x015f)
>> +#define SMC_RMI_RTT_READ_ENTRY        SMC_RxI_CALL(0x0161)
>> +#define SMC_RMI_RTT_SET_RIPAS        SMC_RxI_CALL(0x0169)
>> +#define SMC_RMI_RTT_UNMAP_UNPROTECTED    SMC_RxI_CALL(0x0162)
>> +#define SMC_RMI_VERSION            SMC_RxI_CALL(0x0150)
>> +
> 
> Similar to what we had in rsi_smc.h, it may be good idea to have those
> definitions
> in the ascending order of the function ID (number). It will help readers
> to search
> based on the function ID (number) if you agree.

Ah, good point. I'd been matching the specification, but that has now
been updated to numeric order now. I'll reorder.

>> +#define RMI_ABI_MAJOR_VERSION    1
>> +#define RMI_ABI_MINOR_VERSION    0
>> +
>> +#define RMI_ABI_VERSION_GET_MAJOR(version) ((version) >> 16)
>> +#define RMI_ABI_VERSION_GET_MINOR(version) ((version) & 0xFFFF)
>> +#define RMI_ABI_VERSION(major, minor)      (((major) << 16) | (minor))
>> +
>> +#define RMI_UNASSIGNED            0
>> +#define RMI_ASSIGNED            1
>> +#define RMI_TABLE            2
>> +
> 
> Those RTT entry states are associated with struct rtt_entry::state only.
> So the best
> place to have those definiation would be rmi_cmds.h where 'struct
> rtt_entry' is
> declared. Besides, there is a enumeration RmiRttEntryState for them as
> stated in
> the specifiction (B4.4.24).

The struct rtt_entry is a Linux struct and not part of the specification
(e.g. reordering the fields would't break anything), whereas these
enumeration values are part of the spec. I have been trying to keep
values from the spec in rmi_smc.h.

I could convert the #defines to a proper enum (like enum rmi_ripas), but
to be honest I'd been wondering more about switching the rmi_ripas
values to #defines... I'm not a great fan of C's enums. On the other
hand I guess it would provide some documentation of what these values
are (which is the main reason I've left enum rmi_ripas as an enum).

>> +#define RMI_RETURN_STATUS(ret)        ((ret) & 0xFF)
>> +#define RMI_RETURN_INDEX(ret)        (((ret) >> 8) & 0xFF)
>> +
>> +#define RMI_SUCCESS        0
>> +#define RMI_ERROR_INPUT        1
>> +#define RMI_ERROR_REALM        2
>> +#define RMI_ERROR_REC        3
>> +#define RMI_ERROR_RTT        4
>> +
>> +enum rmi_ripas {
>> +    RMI_EMPTY = 0,
>> +    RMI_RAM = 1,
>> +    RMI_DESTROYED = 2,
>> +};
>> +
>> +#define RMI_NO_MEASURE_CONTENT    0
>> +#define RMI_MEASURE_CONTENT    1
>> +
>> +#define RMI_FEATURE_REGISTER_0_S2SZ        GENMASK(7, 0)
>> +#define RMI_FEATURE_REGISTER_0_LPA2        BIT(8)
>> +#define RMI_FEATURE_REGISTER_0_SVE_EN        BIT(9)
>> +#define RMI_FEATURE_REGISTER_0_SVE_VL        GENMASK(13, 10)
>> +#define RMI_FEATURE_REGISTER_0_NUM_BPS        GENMASK(19, 14)
>> +#define RMI_FEATURE_REGISTER_0_NUM_WPS        GENMASK(25, 20)
>> +#define RMI_FEATURE_REGISTER_0_PMU_EN        BIT(26)
>> +#define RMI_FEATURE_REGISTER_0_PMU_NUM_CTRS    GENMASK(31, 27)
>> +#define RMI_FEATURE_REGISTER_0_HASH_SHA_256    BIT(32)
>> +#define RMI_FEATURE_REGISTER_0_HASH_SHA_512    BIT(33)
>> +#define RMI_FEATURE_REGISTER_0_GICV3_NUM_LRS    GENMASK(37, 34)
>> +#define RMI_FEATURE_REGISTER_0_MAX_RECS_ORDER    GENMASK(41, 38)
>> +
> 
> The 'Reserved' field can be defined as well so that the definitions are
> complete
> if you agree.
> 
>    #define RMI_FEATGURE_REGISTER_0_Reserved    GENMASK(63, 42)

Fair enough - although I'm not sure when we'll ever use that define.

>> +#define RMI_REALM_PARAM_FLAG_LPA2        BIT(0)
>> +#define RMI_REALM_PARAM_FLAG_SVE        BIT(1)
>> +#define RMI_REALM_PARAM_FLAG_PMU        BIT(2)
>> +
>> +/*
>> + * Note many of these fields are smaller than u64 but all fields have
>> u64
>> + * alignment, so use u64 to ensure correct alignment.
>> + */
>> +struct realm_params {
>> +    union { /* 0x0 */
>> +        struct {
>> +            u64 flags;
>> +            u64 s2sz;
>> +            u64 sve_vl;
>> +            u64 num_bps;
>> +            u64 num_wps;
>> +            u64 pmu_num_ctrs;
>> +            u64 hash_algo;
>> +        };
>> +        u8 padding1[0x400];
>> +    };
>> +    union { /* 0x400 */
>> +        u8 rpv[64];
>> +        u8 padding2[0x400];
>> +    };
>> +    union { /* 0x800 */
>> +        struct {
>> +            u64 vmid;
>> +            u64 rtt_base;
>> +            s64 rtt_level_start;
>> +            u64 rtt_num_start;
>> +        };
>> +        u8 padding3[0x800];
>> +    };
>> +};
>> +
>> +/*
>> + * The number of GPRs (starting from X0) that are
>> + * configured by the host when a REC is created.
>> + */
>> +#define REC_CREATE_NR_GPRS        8
>> +
>> +#define REC_PARAMS_FLAG_RUNNABLE    BIT_ULL(0)
>> +
>> +#define REC_PARAMS_AUX_GRANULES        16
>> +
>> +struct rec_params {
>> +    union { /* 0x0 */
>> +        u64 flags;
>> +        u8 padding1[0x100];
>> +    };
>> +    union { /* 0x100 */
>> +        u64 mpidr;
>> +        u8 padding2[0x100];
>> +    };
>> +    union { /* 0x200 */
>> +        u64 pc;
>> +        u8 padding3[0x100];
>> +    };
>> +    union { /* 0x300 */
>> +        u64 gprs[REC_CREATE_NR_GPRS];
>> +        u8 padding4[0x500];
>> +    };
>> +    union { /* 0x800 */
>> +        struct {
>> +            u64 num_rec_aux;
>> +            u64 aux[REC_PARAMS_AUX_GRANULES];
>> +        };
>> +        u8 padding5[0x800];
>> +    };
>> +};
>> +
>> +#define REC_ENTER_EMULATED_MMIO        BIT(0)
>> +#define REC_ENTER_INJECT_SEA        BIT(1)
>> +#define REC_ENTER_TRAP_WFI        BIT(2)
>> +#define REC_ENTER_TRAP_WFE        BIT(3)
>> +#define REC_ENTER_RIPAS_RESPONSE    BIT(4)
>> +
>> +#define REC_RUN_GPRS            31
>> +#define REC_GIC_NUM_LRS            16
>> +
>> +struct rec_enter {
>> +    union { /* 0x000 */
>> +        u64 flags;
>> +        u8 padding0[0x200];
>> +    };
>> +    union { /* 0x200 */
>> +        u64 gprs[REC_RUN_GPRS];
>> +        u8 padding2[0x100];
>> +    };
>> +    union { /* 0x300 */
>> +        struct {
>> +            u64 gicv3_hcr;
>> +            u64 gicv3_lrs[REC_GIC_NUM_LRS];
>> +        };
>> +        u8 padding3[0x100];
>> +    };
>> +    u8 padding4[0x400];
>> +};
>> +
>> +#define RMI_EXIT_SYNC            0x00
>> +#define RMI_EXIT_IRQ            0x01
>> +#define RMI_EXIT_FIQ            0x02
>> +#define RMI_EXIT_PSCI            0x03
>> +#define RMI_EXIT_RIPAS_CHANGE        0x04
>> +#define RMI_EXIT_HOST_CALL        0x05
>> +#define RMI_EXIT_SERROR            0x06
>> +
>> +struct rec_exit {
>> +    union { /* 0x000 */
>> +        u8 exit_reason;
>> +        u8 padding0[0x100];
>> +    };
>> +    union { /* 0x100 */
>> +        struct {
>> +            u64 esr;
>> +            u64 far;
>> +            u64 hpfar;
>> +        };
>> +        u8 padding1[0x100];
>> +    };
>> +    union { /* 0x200 */
>> +        u64 gprs[REC_RUN_GPRS];
>> +        u8 padding2[0x100];
>> +    };
>> +    union { /* 0x300 */
>> +        struct {
>> +            u64 gicv3_hcr;
>> +            u64 gicv3_lrs[REC_GIC_NUM_LRS];
>> +            u64 gicv3_misr;
>> +            u64 gicv3_vmcr;
>> +        };
>> +        u8 padding3[0x100];
>> +    };
>> +    union { /* 0x400 */
>> +        struct {
>> +            u64 cntp_ctl;
>> +            u64 cntp_cval;
>> +            u64 cntv_ctl;
>> +            u64 cntv_cval;
>> +        };
>> +        u8 padding4[0x100];
>> +    };
>> +    union { /* 0x500 */
>> +        struct {
>> +            u64 ripas_base;
>> +            u64 ripas_top;
>> +            u64 ripas_value;
>> +        };
>> +        u8 padding5[0x100];
>> +    };
>> +    union { /* 0x600 */
>> +        u16 imm;
>> +        u8 padding6[0x100];
>> +    };
>> +    union { /* 0x700 */
>> +        struct {
>> +            u8 pmu_ovf_status;
>> +        };
>> +        u8 padding7[0x100];
>> +    };
>> +};
>> +
> 
> The names for the 'padding' field starts from 'padding1' instead of
> 'padding0'
> as we did for other structures.

Actually this file is really quite inconsitent! E.g. struct rec_enter
starts from 0 and then skips 1... I'll clean this up!

Thanks,
Steve

>> +struct rec_run {
>> +    struct rec_enter enter;
>> +    struct rec_exit exit;
>> +};
>> +
>> +#endif
> 
> #endif /* __ASM_RME_SMC_H */
> 
> Thanks,
> Gavin
>