From: dave.patel@riscstar.com
To: opensbi@lists.infradead.org
Cc: Scott Bambrough <scott@riscstar.com>,
Robin Randhawa <robin.randhawa@sifive.com>,
Anup Patel <anup.patel@qti.qualcomm.com>,
Samuel Holland <samuel.holland@sifive.com>,
Dave Patel <dave.patel@riscstar.com>,
Ray Mao <raymond.mao@riscstar.com>,
Anup Patel <anuppate@qti.qualcomm.com>,
Dhaval <dhaval@rivosinc.com>, Peter Lin <peter.lin@sifive.com>
Subject: [PATCH v1 0/4] Add eager FP and RISC-V vector context switching support
Date: Fri, 20 Mar 2026 14:23:45 +0000 [thread overview]
Message-ID: <20260320142349.72980-1-dave.patel@riscstar.com> (raw)
Hi,
This patch series adds support for eager floating-point and RISC-V vector
context save/restore in OpenSBI.
The goal is to ensure that machine-mode trap handling preserves the full
architectural state of lower privilege software (e.g., S-mode), providing
correct isolation across trap boundaries.
### Overview
This series introduces:
* Per-hart FP and vector context pointers in `sbi_scratch`
* FP context save/restore implementation
* RISC-V vector context save/restore implementation
* Integration into the OpenSBI trap handler
### Design
The implementation follows an eager context switching model:
* FP and vector state is saved on every trap entry
* FP and vector state is restored before returning from traps
This ensures correctness and avoids the need for lazy switching or
trap-on-first-use mechanisms.
### Context Storage
* FP context:
* f0–f31 registers
* fcsr
* Vector context:
* v0–v31 registers
* vl, vtype, vcsr
Stored per-hart via `sbi_scratch`.
### Notes / Assumptions
* Assumes FP and vector units are enabled when used (mstatus.FS/VS)
* Assumes RV64 for floating-point implementation (fsd/fld)
* Vector maximum size is bounded by SBI_MAX_VLENB
* Context pointers must be initialized before use
### Performance
This approach introduces additional overhead on every trap due to full
state save and restore, especially for vector state, but provides a
simple and deterministic model.
### Patch Breakdown
* Patch 1: Add FP context save/restore support
* Patch 2: Add vector context save/restore support
* Patch 3: Extend sbi_scratch with FP/vector context pointers
* Patch 4: Integrate into trap handler
Feedback is welcome, especially regarding:
* CSR restore ordering (vtype/vl/vcsr)
* Trap performance implications
Thanks,
Dave Patel
--
opensbi mailing list
opensbi@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/opensbi
next reply other threads:[~2026-03-20 14:24 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-20 14:23 dave.patel [this message]
2026-03-20 14:23 ` [PATCH 1/4] lib: sbi: domain: ensure boot_hartid is assigned dave.patel
2026-03-20 14:23 ` [PATCH 2/4] lib: sbi: Add RISC-V vector context save/restore support (eager switching) dave.patel
2026-03-20 14:23 ` [PATCH 3/4] lib: sbi: Add floating-point context save/restore support dave.patel
2026-03-20 14:23 ` [PATCH 4/4] include: sbi: scratch: Add per-hart FP and vector context pointers in scratch dave.patel
-- strict thread matches above, loose matches on Subject: below --
2026-03-20 14:30 [PATCH v1 0/4] Add eager FP and RISC-V vector context switching support dave.patel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260320142349.72980-1-dave.patel@riscstar.com \
--to=dave.patel@riscstar.com \
--cc=anup.patel@qti.qualcomm.com \
--cc=anuppate@qti.qualcomm.com \
--cc=dhaval@rivosinc.com \
--cc=opensbi@lists.infradead.org \
--cc=peter.lin@sifive.com \
--cc=raymond.mao@riscstar.com \
--cc=robin.randhawa@sifive.com \
--cc=samuel.holland@sifive.com \
--cc=scott@riscstar.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox