Sail Support for Priv 1.13

[jjscheel]

Technical Group

Privileged Spec IC

ratification-pkg

Priv 1.13

Technical Liaison

Greg Favor

Task Category

SAIL model

Task Sub Category

• [ ] gcc
• [ ] binutils
• [ ] gdb
• [ ] intrinsics
• [ ] Java
• [ ] KVM
• [ ] ld
• [ ] llvm
• [ ] Linux kernel
• [ ] QEMU
• [ ] Spike

Ratification Target

3Q2024

Statement of Work (SOW)

Component names:

Requirements: Provide Sail model function for the following Priv 1.13 updates:

1. Defined the misa.V field to reflect that the V extension has been implemented.
2. Exposed count-overflow interrupts to VS-mode.
3. Redefined misa.MXL to be read-only, making MXLEN a constant.
4. Added the constraint that SXLEN≥UXLEN.
5. Defined the RV32-only medelegh and hedelegh CSRs. (1)
6. Allocated interrupt 13 for Sscofpmf LCOFI interrupt. (2)

The following footnotes clarify level of support:

1. Limitations in the hypervisor support will determine how much function can be enabled.
2. Limitations in the counter support will determine how much function can be enabled.

Note: the following features of Priv 1.13 require ACT infrastructure that prevent their implementation at this time:

• Defined hardware error and software check exception codes.
• Defined the misaligned atomicity granule PMA, superseding the proposed Zam extension.

Deliverables:

• Accepted Sail model updates (PRs) in support of the above Priv 1.13 features
• Accepted Sail build verification tests (PRs) to support the new model features

Acceptance Criteria:

• Passing of the ACT tests (regression)
• Passing of the build verification tests for Priv 1.13 (Note: the ACT tests for Priv 1.13 are requested to be self-checking and as such should be usable as BVTs. See #43 for more discussion.)

Projected timeframe: (best guess date) 2 person months (part time)

SOW Signoffs: (delete those not needed)

• Task group liaison sign-off
• Development partner sign-off
• Golden Model SIG sign-off (if SAIL work)

Waiver

• [X] Freeze
• [ ] Ratification

Pull Request Details

No response

[jjscheel]

@billmcspadden-riscv, I'd love your review and feedback of this SOW.

[billmcspadden-riscv]

With regards to acceptance criteria, passing of ACT tests: if the ACT tests are not self-checking, then the tests must be run against another model (perhaps Spike) and the signatures compared. If a second simulator is needed, then that model must also implement 1.13 features as well.

In short, unless the ACTs are self-checking for 1.13 features, I don't think this should be criteria for acceptance.

[jjscheel]

Thanks, @billmcspadden-riscv. So, you're advocating that I remove the second acceptance item and not replace it with anything?

[billmcspadden-riscv]

I think the BVTs for 1.13 must exist and must pass to meet acceptance criteria.

If the ACTs exist and are self-checking, then they should be part of the acceptance criteria. We need to ask Allen if the tests will be self-checking.

In general, I think we need to try and make the ACTs self-checking.

Bill Mc.

[jjscheel]

Thanks, @billmcspadden-riscv. I've updated the SOW description above w.r.t. ACT and BVTs. Please let me know if you'd like any additional clarifications.

[jjscheel]

@billmcspadden-riscv will be driving this work. If anyone has resources who can help, please reach out to him.

[billmcspadden-riscv]

Under way....