On 2/2/26 3:08 PM, Jan Beulich wrote:
On 02.02.2026 13:57, Oleksii Kurochko wrote:
--- a/xen/arch/riscv/include/asm/domain.h
+++ b/xen/arch/riscv/include/asm/domain.h
@@ -24,6 +24,55 @@ struct arch_vcpu_io {
struct arch_vcpu {
struct vcpu_vmid vmid;
Nit: This being already here contradicts the title saying "introduce".
I'll update then the commit message.
+
+ /*
+ * Callee saved registers for Xen's state used to switch from
+ * prev's stack to the next's stack during context switch.
+ */
+ struct
+ {
+ register_t s0;
+ register_t s1;
+ register_t s2;
+ register_t s3;
+ register_t s4;
+ register_t s5;
+ register_t s6;
+ register_t s7;
+ register_t s8;
+ register_t s9;
+ register_t s10;
+ register_t s11;
+ register_t sp;
+ register_t ra;
+ } xen_saved_context;
+
+ register_t hedeleg;
+ register_t hideleg;
+ register_t hvip;
+ register_t hip;
+ register_t hie;
+ register_t hgeie;
+ register_t henvcfg;
+ register_t hcounteren;
+ register_t htimedelta;
+ register_t htval;
+ register_t htinst;
+ register_t hstateen0;
+#ifdef CONFIG_RISCV_32
+ register_t henvcfgh;
+ register_t htimedeltah;
+#endif
When you add these right away, any reason other high-half registers aren't added
as well? HVIP (above) and VSIE / VSIP at the very least have such counterparts,
iirc.
No specific reason. It was added early in the RISC-V port when I was considering
supporting RV32, but as you know I later decided to focus only on RV64, at least
for now. Therefore, it seems better to drop this `#ifdef` from the `arch_vcpu`
structure and re-introduce when RV32 support will be added.
+ register_t vsstatus;
+ register_t vsip;
+ register_t vsie;
+ register_t vstvec;
+ register_t vsscratch;
+ register_t vscause;
+ register_t vstval;
+ register_t vsatp;
+ register_t vsepc;
};
What I'm generally concerned of: With the registers all being put here in one
go,
without any of them actually being used in this patch, it'll likely be harder to
track that they're saved / restored / initialized correctly (once uses appear).
Then if it will really better I'll introduce in this patch series only vsatp
register
as it is used in the further patches. (or even will just introduce vsatp in
Patch 3)
I will keep the patch [1] separately in the patch series with introduction of
vtimer
things and drop only an introduction of vsatp there.
[1]
https://lore.kernel.org/xen-devel/[email protected]/T/#ma33af3ff4b2b033f86aa964902ebcab8705fd114
~ Oleksii
