[PATCH v18 1/7] firmware: arm_rmm: Add SMC definitions for calling the RMM

[Jonathan Cameron]

On Mon, 21 Sep 2026 10:27:46 +0100
Suzuki K Poulose <suzuki.poulose at arm.com> wrote:

> On 19/09/2026 02:27, Jonathan Cameron 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 2.0-bet3
> >>
> >> [1] https://developer.arm.com/documentation/den0137/2-0bet3/
> >>
> >> Signed-off-by: Steven Price <steven.price at arm.com>
> >> Signed-off-by: Suzuki K Poulose <suzuki.poulose at arm.com>  
> > 
> > With Gavin's nitpicks and the GENMASK_ULL() from sashiko, just a few
> > comments inline.  Mostly on subtle inconsistencies that really don't
> > matter that much.
> >   
> >>   include/linux/arm-smccc-rmi.h | 497 ++++++++++++++++++++++++++++++++++
> >>   1 file changed, 497 insertions(+)
> >>   create mode 100644 include/linux/arm-smccc-rmi.h
> >>
> >> diff --git a/include/linux/arm-smccc-rmi.h b/include/linux/arm-smccc-rmi.h
> >> new file mode 100644
> >> index 000000000000..214d6228dfc2
> >> --- /dev/null
> >> +++ b/include/linux/arm-smccc-rmi.h
> >> @@ -0,0 +1,497 @@
> >> +/* SPDX-License-Identifier: GPL-2.0 */
> >> +/*
> >> + * Copyright (C) 2023-2026 ARM Ltd.
> >> + *
> >> + * The values and structures in this file are from the Realm Management Monitor
> >> + * specification (DEN0137) version 2.0-bet3:
> >> + * https://developer.arm.com/documentation/den0137/2-0bet3/
> >> + */
> >> +
> >> +#ifndef __LINUX_ARM_SMCCC_RMI_H_
> >> +#define __LINUX_ARM_SMCCC_RMI_H_
> >> +
> >> +#include <linux/arm-smccc.h>
> >> +#include <linux/bitfield.h>
> >> +#include <linux/bits.h>
> >> +#include <linux/build_bug.h>
> >> +#include <linux/sizes.h>
> >> +
> >> +#include <asm/page.h>
> >> +
> >> +#define SMC_RMI_CALL(func)				\
> >> +	ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,		\
> >> +			   ARM_SMCCC_SMC_64,		\
> >> +			   ARM_SMCCC_OWNER_STANDARD,	\
> >> +			   (func))  
> > 
> > Obviously it is v18 so probably a future thing but nothing about this
> > is RMI specific.  Could be used for ARM_SMCCC_TRNG_RND64 for instance.
> > I'm not entirely sure what we'd call such a macro
> > 
> > ARM_SMCCC_CALL_VAL64_STD() maybe?  
> 
> ARM_SMCCC_STD_CALL64_VAL() ?
> 
> But, I would leave it as a wider cleanup in the tree as a separate
> series.

Ok.  A follow up would be fine I guess.


> >> +
> >> +#define RMI_RETURN_STATUS_MASK		GENMASK(7, 0)
> >> +#define RMI_RETURN_INDEX_MASK		GENMASK(15, 8)
> >> +#define RMI_RETURN_MEMREQ_MASK		GENMASK(9, 8)
> >> +#define RMI_RETURN_CAN_CANCEL_MASK	BIT(10)
> >> +
> >> +#define RMI_RETURN_STATUS(ret)		FIELD_GET(RMI_RETURN_STATUS_MASK, ret)
> >> +#define RMI_RETURN_INDEX(ret)		FIELD_GET(RMI_RETURN_INDEX_MASK, ret)  
> > 
> > What's this one?  I can't find anything in the spec that matches it
> > and as far as I can tell you don't use it in this series.  
> 
> This is coming from RmiResultDataLevel. See RmiResult type.
> This was renamed after we introduce the RmiResultDataIncomplete.
> It is used in the KVM code to find the "level" where a command
> failed/walked.
> 
> I could rename it to RMI_RESULT_DATA_LEVEL() ?
> Similarly RMI_RESULT_STATUS instead of RMI_RETURN_*

Yes, that would make tracking it down easier. Thanks.

> 
> >   
> >> +#define RMI_RETURN_MEMREQ(ret)		FIELD_GET(RMI_RETURN_MEMREQ_MASK, ret)
> >> +#define RMI_RETURN_CAN_CANCEL(ret)	FIELD_GET(RMI_RETURN_CAN_CANCEL_MASK, ret)  
> 
> > These are obscure enough to find in the spec I'd give a comment just
> > to save the sanity of anyone looking for them.  
> 
> As above, they are really RMI_RESULT_DATA_INCOMPLETE_*
> 
> >   
> >> +/*
> >> + * Note many of these fields are smaller than u64 but all fields have u64
> >> + * alignment, so use u64 to ensure correct alignment.  
> > 
> > Obviously this is only going to run on arm64 so it's not critical, but
> > more generally u64s aren't always 64 bit aligned.  So if you 'really'
> > care aligned_u64 is there to ensure it.  Meh, arm64 so fine.  
> 
> Agreed, I am worried about the churn in the consumer code. Also, like
> you said, this is only for ARM64. So, I would pass it.

Ok. A tiny bit ugly but x86_32 adoption of RMM 2.0 is likely to be minimal :)


> >> +
> >> +struct rec_params {
> >> +	union { /* 0x0 */
> >> +		u64 flags;
> >> +		u8 padding0[0x100];
> >> +	};
> >> +	union { /* 0x100 */
> >> +		u64 mpidr;
> >> +		u8 padding1[0x100];
> >> +	};
> >> +	union { /* 0x200 */
> >> +		u64 pc;
> >> +		u8 padding2[0x100];
> >> +	};
> >> +	union { /* 0x300 */
> >> +		u64 gprs[REC_CREATE_NR_GPRS];
> >> +		u8 padding3[0xd00];
> >> +	};
> >> +};
> >> +
> >> +static_assert(sizeof(struct rec_params) == SZ_4K);  
> > Whilst the assert works and is need to prevent oversized the
> > dos never seem to provide any indication of the final trailing
> > padding other than indirectly and I don't like maths on Fridays ;).  
> 
> Agree it is a bit obscure, but ...
> 
> > Maybe union the inner union set with a u8 [SZ_4K]?
> >   
> 
> That doesn't help if the other one runs past SZ_4K ?

You still need the assert so I guess that is already providing
the documentation indirectly so indeed little purpose in the
extra union beyond removing need to have magic padding in the
last element.  Mind you not obvious what that last pad should be
if it wasn't just 'the rest'. So, I think this is fine as is.


Thanks,

Jonathan