Hi Abdellatif On Mon, Nov 07, 2022 at 07:20:54PM +0000, Abdellatif El Khlifi wrote: > Add MM communication support using FF-A transport > > This feature allows accessing MM partitions services through > EFI MM communication protocol. MM partitions such as StandAlonneMM > or smm-gateway secure partitions which reside in secure world. > > An MM shared buffer and a door bell event are used to exchange > the data. > > The data is used by EFI services such as GetVariable()/SetVariable() > and copied from the communication buffer to the MM shared buffer. > > config EFI_MM_COMM_TEE > - bool "UEFI variables storage service via OP-TEE" > - depends on OPTEE > + bool "UEFI variables storage service via the trusted world" > + depends on OPTEE && ARM_FFA_TRANSPORT
This shouldn't rely on both. It's either OP-TEE or FF-A > > +#if (IS_ENABLED(CONFIG_OPTEE)) This separation is a bit problematic. A user can configure bot OP-TEE and FF-A. Those are not mutually exclusive, but for the EFI variables case they are. We need a better way to isolate the compilation choices. Why don't we make ffa_bus_discover() return -1 if FF-A isn't compiled in? > /** > * get_connection() - Retrieve OP-TEE session for a specific UUID. > * > @@ -143,13 +176,229 @@ static efi_status_t optee_mm_communicate(void > *comm_buf, ulong dsize) > > return ret; > } > +#endif > + > +#if (IS_ENABLED(CONFIG_ARM_FFA_TRANSPORT)) > + > +/** > + * ffa_notify_mm_sp() - Announce there is data in the shared buffer > + * > + * Notifies the MM partition in the trusted world that > + * data is available in the shared buffer. > + * This is a blocking call during which trusted world has exclusive access > + * to the MM shared buffer. > + * > + * Return: > + * > + * 0 on success > + */ > +static int ffa_notify_mm_sp(void) > +{ > + struct ffa_send_direct_data msg = {0}; > + int ret; > + int sp_event_ret = -1; > + > + if (!ffa_bus_ops_get()) > + return -EINVAL; > + > + msg.data0 = FFA_SHARED_MM_BUFFER_OFFSET; /* x3 */ > + > + ret = ffa_bus_ops_get()->sync_send_receive(mm_sp_id, &msg, 1); > + if (ret != 0) > + return ret; > + > + sp_event_ret = msg.data0; /* x3 */ > + > + if (sp_event_ret == MM_SUCCESS) > + return 0; > + > + /* > + * Failure to notify the MM SP > + */ > + > + return -EACCES; > +} > + > +/** > + * ffa_discover_mm_sp_id() - Query the MM partition ID > + * > + * Use the FF-A driver to get the MM partition ID. > + * If multiple partitions are found, use the first one. > + * This is a boot time function. > + * > + * Return: > + * > + * 0 on success > + */ > +static int ffa_discover_mm_sp_id(void) > +{ > + u32 count = 0, size = 0; > + int ret; > + struct ffa_partition_info *parts_info; > + > + if (!ffa_bus_ops_get()) > + return -EINVAL; > + > + /* > + * get from the driver the count of the SPs matching the UUID > + */ > + ret = ffa_bus_ops_get()->partition_info_get(mm_sp_svc_uuid, &count, > NULL); > + if (ret != 0) { > + log_err("EFI: Failure in querying partitions count (error code: > %d)\n", ret); > + return ret; > + } > + > + if (!count) { > + log_info("EFI: No MM partition found\n"); > + return ret; > + } > + > + /* > + * pre-allocate a buffer to be filled by the driver > + * with ffa_partition_info structs > + */ > + > + log_info("EFI: Pre-allocating %d partition(s) info structures\n", > count); > + > + parts_info = calloc(count, sizeof(struct ffa_partition_info)); I prefer sizeof(*parts_info). Same goes for all sizeof() calls. > + if (!parts_info) > + return -ENOMEM; > + > + size = count * sizeof(struct ffa_partition_info); > + > + /* > + * ask the driver to fill the > + * buffer with the SPs info > + */ > + ret = ffa_bus_ops_get()->partition_info_get(mm_sp_svc_uuid, &size, > parts_info); > + if (ret != 0) { if (!ret) [...] Thanks /Ilias