[Patch v5 6/8] firmware: qcom: scm: Add support for ARM64 SoCs
Andy Gross
andy.gross at linaro.org
Thu Jun 2 20:48:48 PDT 2016
On Thu, Jun 02, 2016 at 03:28:40PM -0700, Stephen Boyd wrote:
> On 05/12, Andy Gross wrote:
> > +
> > +#define MAX_QCOM_SCM_ARGS 10
> > +#define MAX_QCOM_SCM_RETS 3
> > +
> > +#define QCOM_SCM_ARGS_IMPL(num, a, b, c, d, e, f, g, h, i, j, ...) (\
> > + (((a) & 0x3) << 4) | \
> > + (((b) & 0x3) << 6) | \
> > + (((c) & 0x3) << 8) | \
> > + (((d) & 0x3) << 10) | \
> > + (((e) & 0x3) << 12) | \
> > + (((f) & 0x3) << 14) | \
> > + (((g) & 0x3) << 16) | \
> > + (((h) & 0x3) << 18) | \
> > + (((i) & 0x3) << 20) | \
> > + (((j) & 0x3) << 22) | \
> > + (num & 0xf))
>
> Parenthesis around num?
To be more precise, yes.
> > +
> > +#define QCOM_SCM_ARGS(...) QCOM_SCM_ARGS_IMPL(__VA_ARGS__, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)
> > +
> > +/**
> > + * struct qcom_scm_desc
> > + * @arginfo: Metadata describing the arguments in args[]
> > + * @args: The array of arguments for the secure syscall
> > + * @res: The values returned by the secure syscall
> > + */
> > +struct qcom_scm_desc {
> > + u32 arginfo;
> > + u64 args[MAX_QCOM_SCM_ARGS];
> > + struct arm_smccc_res res;
> > +};
>
> If we split the res from the descriptor structure we could make
> the qcom_scm_desc const in qcom_scm_call(). We would have to add
> another argument to the function though, not sure if this is a
> big win idea, but just an idea to keep things "safer".
I like that idea. The arginfo and args are immutable so this really locks down
the interface. I'll try it out and see how it turns out.
> > +
> > +static u64 qcom_smccc_convention = -1;
> > +static DEFINE_MUTEX(qcom_scm_lock);
> > +
> > +#define QCOM_SCM_EBUSY_WAIT_MS 30
> > +#define QCOM_SCM_EBUSY_MAX_RETRY 20
> > +
> > +#define N_EXT_QCOM_SCM_ARGS 7
> > +#define FIRST_EXT_ARG_IDX 3
> > +#define N_REGISTER_ARGS (MAX_QCOM_SCM_ARGS - N_EXT_QCOM_SCM_ARGS + 1)
> > +
> > +/**
> > + * qcom_scm_call() - Invoke a syscall in the secure world
> > + * @dev: device
> > + * @svc_id: service identifier
> > + * @cmd_id: command identifier
> > + * @desc: Descriptor structure containing arguments and return values
> > + *
> > + * Sends a command to the SCM and waits for the command to finish processing.
> > + * This should *only* be called in pre-emptible context.
> > +*/
> > +static int qcom_scm_call(struct device *dev, u32 svc_id, u32 cmd_id,
> > + struct qcom_scm_desc *desc)
> > +{
> > + int arglen = desc->arginfo & 0xf;
> > + int retry_count = 0, i;
> > + u32 fn_id = QCOM_SCM_FNID(svc_id, cmd_id);
> > + u64 cmd, x5 = desc->args[FIRST_EXT_ARG_IDX];
> > + dma_addr_t args_phys = 0;
> > + void *args_virt = NULL;
> > + size_t alloc_len;
> > +
> > + if (unlikely(arglen > N_REGISTER_ARGS)) {
> > + alloc_len = N_EXT_QCOM_SCM_ARGS * sizeof(u64);
> > + args_virt = kzalloc(PAGE_ALIGN(alloc_len), GFP_KERNEL);
> > +
> > + if (!args_virt)
> > + return qcom_scm_remap_error(-ENOMEM);
>
> Just return -ENOMEM here?
Ok. That would make it easier to follow.
> > +
> > + if (qcom_smccc_convention == ARM_SMCCC_SMC_32) {
> > + __le32 *args = args_virt;
> > +
> > + for (i = 0; i < N_EXT_QCOM_SCM_ARGS; i++)
> > + args[i] = cpu_to_le32(desc->args[i +
> > + FIRST_EXT_ARG_IDX]);
> > + } else {
> > + __le64 *args = args_virt;
> > +
> > + for (i = 0; i < N_EXT_QCOM_SCM_ARGS; i++)
> > + args[i] = cpu_to_le64(desc->args[i +
> > + FIRST_EXT_ARG_IDX]);
> > + }
> > +
> > + args_phys = dma_map_single(dev, args_virt, alloc_len,
> > + DMA_TO_DEVICE);
> > +
> > + if (dma_mapping_error(dev, args_phys)) {
> > + kfree(args_virt);
> > + return qcom_scm_remap_error(-ENOMEM);
>
> Just return -ENOMEM here?
Ok. That would make it easier to follow.
More information about the linux-arm-kernel
mailing list