On Tue, Jun 13, 2023 at 10:20:33AM -0700, Elliot Berman wrote:
> Add hypercalls to send and receive messages on a Gunyah message queue.
>
> Reviewed-by: Alex Elder <elder@xxxxxxxxxx>
> Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxx>
> Signed-off-by: Elliot Berman <quic_eberman@xxxxxxxxxxx>
> ---
> arch/arm64/gunyah/gunyah_hypercall.c | 31 ++++++++++++++++++++++++++++
> include/linux/gunyah.h | 6 ++++++
> 2 files changed, 37 insertions(+)
>
> diff --git a/arch/arm64/gunyah/gunyah_hypercall.c b/arch/arm64/gunyah/gunyah_hypercall.c
> index ef48e0b8b5448..c5bf0dd468ec9 100644
> --- a/arch/arm64/gunyah/gunyah_hypercall.c
> +++ b/arch/arm64/gunyah/gunyah_hypercall.c
> @@ -35,6 +35,8 @@ EXPORT_SYMBOL_GPL(arch_is_gh_guest);
> fn)
>
> #define GH_HYPERCALL_HYP_IDENTIFY GH_HYPERCALL(0x8000)
> +#define GH_HYPERCALL_MSGQ_SEND GH_HYPERCALL(0x801B)
> +#define GH_HYPERCALL_MSGQ_RECV GH_HYPERCALL(0x801C)
>
> /**
> * gh_hypercall_hyp_identify() - Returns build information and feature flags
> @@ -54,5 +56,34 @@ void gh_hypercall_hyp_identify(struct gh_hypercall_hyp_identify_resp *hyp_identi
> }
> EXPORT_SYMBOL_GPL(gh_hypercall_hyp_identify);
>
> +enum gh_error gh_hypercall_msgq_send(u64 capid, size_t size, void *buff, u64 tx_flags, bool *ready)
I find the name "ready" unclear, I believe it will be true to indicate
that the message queue is not full (i.e. Gunyah is ready to receive
another message?)
Naming it "full" and reversing the values would make this immediately
obvious.
> +{
> + struct arm_smccc_res res;
> +
> + arm_smccc_1_1_hvc(GH_HYPERCALL_MSGQ_SEND, capid, size, (uintptr_t)buff, tx_flags, 0, &res);
> +
> + if (res.a0 == GH_ERROR_OK)
> + *ready = !!res.a1;
> +
> + return res.a0;
> +}
> +EXPORT_SYMBOL_GPL(gh_hypercall_msgq_send);
> +
> +enum gh_error gh_hypercall_msgq_recv(u64 capid, void *buff, size_t size, size_t *recv_size,
> + bool *ready)
And this "ready" seems to be "more data available". Naming it "empty",
and reversing the values, would make this obvious.
Perhaps when working with the hypercalls on a regular basis it makes
sense to not invert the value space...renaming the two "ready" would be
beneficial regardless.
> +{
> + struct arm_smccc_res res;
> +
> + arm_smccc_1_1_hvc(GH_HYPERCALL_MSGQ_RECV, capid, (uintptr_t)buff, size, 0, &res);
> +
> + if (res.a0 == GH_ERROR_OK) {
> + *recv_size = res.a1;
> + *ready = !!res.a2;
> + }
> +
> + return res.a0;
> +}
> +EXPORT_SYMBOL_GPL(gh_hypercall_msgq_recv);
> +
> MODULE_LICENSE("GPL");
> MODULE_DESCRIPTION("Gunyah Hypervisor Hypercalls");
> diff --git a/include/linux/gunyah.h b/include/linux/gunyah.h
> index 6b36cf4787efb..01a6f202d037e 100644
> --- a/include/linux/gunyah.h
> +++ b/include/linux/gunyah.h
> @@ -111,4 +111,10 @@ static inline u16 gh_api_version(const struct gh_hypercall_hyp_identify_resp *gh
>
> void gh_hypercall_hyp_identify(struct gh_hypercall_hyp_identify_resp *hyp_identity);
>
> +#define GH_HYPERCALL_MSGQ_TX_FLAGS_PUSH BIT(0)
Unless the flags are growing, passing this as a bool push instead of a
flag field seems reasonable.
> +
> +enum gh_error gh_hypercall_msgq_send(u64 capid, size_t size, void *buff, u64 tx_flags, bool *ready);
> +enum gh_error gh_hypercall_msgq_recv(u64 capid, void *buff, size_t size, size_t *recv_size,
> + bool *ready);
Please wrap things to 80 characters, unless going a little bit over
makes things more readable.
Regards,
Bjorn
