[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <1572378065-4490-16-git-send-email-eberman@codeaurora.org>
Date: Tue, 29 Oct 2019 12:41:03 -0700
From: Elliot Berman <eberman@...eaurora.org>
To: bjorn.andersson@...aro.org, saipraka@...eaurora.org,
agross@...nel.org
Cc: tsoni@...eaurora.org, sidgup@...eaurora.org,
psodagud@...eaurora.org, linux-arm-msm@...r.kernel.org,
linux-kernel@...r.kernel.org,
Elliot Berman <eberman@...eaurora.org>
Subject: [RFC 15/17] firmware: qcom_scm: Merge legacy and SMCCC conventions
Copy/paste legacy SCM driver into qcom_scm-64 with the following notes:
- Renamed qcom_scm_call in qcom_scm-32 to qcom_scm_call_legacy in the
copy.
- Renamed qcom_scm_call_atomic in qcom_scm 32 to
qcom_scm_call_atomic_legacy in the copy.
- Rename __qcom_scm_call_smccc to qcom_scm_call_smccc.
- Filled in implementations set_cold_boot_addr, set_warm_boot_addr, and
cpu_power_down from qcom_scm-32.c.
- set_dload_mode, io_writel, io_readl now use atomic variants as in
qcom_scm-32.
Signed-off-by: Elliot Berman <eberman@...eaurora.org>
---
drivers/firmware/qcom_scm-64.c | 322 ++++++++++++++++++++++++++++++++++++++++-
1 file changed, 314 insertions(+), 8 deletions(-)
diff --git a/drivers/firmware/qcom_scm-64.c b/drivers/firmware/qcom_scm-64.c
index 9f39483..71b07b8 100644
--- a/drivers/firmware/qcom_scm-64.c
+++ b/drivers/firmware/qcom_scm-64.c
@@ -95,7 +95,7 @@ static void __qcom_scm_call_do_quirk(const struct arm_smccc_args *smc,
} while (res->a0 == QCOM_SCM_INTERRUPTED);
}
-static int ___qcom_scm_call_smccc(struct device *dev,
+static int qcom_scm_call_smccc(struct device *dev,
struct qcom_scm_desc *desc, bool atomic)
{
int arglen = desc->arginfo & 0xf;
@@ -181,6 +181,223 @@ static int ___qcom_scm_call_smccc(struct device *dev,
return 0;
}
+#define LEGACY_FUNCNUM(s, c) (((s) << 10) | ((c) & 0x3ff))
+
+/**
+ * struct legacy_command - one SCM command buffer
+ * @len: total available memory for command and response
+ * @buf_offset: start of command buffer
+ * @resp_hdr_offset: start of response buffer
+ * @id: command to be executed
+ * @buf: buffer returned from legacy_get_command_buffer()
+ *
+ * An SCM command is laid out in memory as follows:
+ *
+ * ------------------- <--- struct legacy_command
+ * | command header |
+ * ------------------- <--- legacy_get_command_buffer()
+ * | command buffer |
+ * ------------------- <--- struct legacy_response and
+ * | response header | legacy_command_to_response()
+ * ------------------- <--- legacy_get_response_buffer()
+ * | response buffer |
+ * -------------------
+ *
+ * There can be arbitrary padding between the headers and buffers so
+ * you should always use the appropriate qcom_scm_get_*_buffer() routines
+ * to access the buffers in a safe manner.
+ */
+struct legacy_command {
+ __le32 len;
+ __le32 buf_offset;
+ __le32 resp_hdr_offset;
+ __le32 id;
+ __le32 buf[0];
+};
+
+/**
+ * struct legacy_response - one SCM response buffer
+ * @len: total available memory for response
+ * @buf_offset: start of response data relative to start of legacy_response
+ * @is_complete: indicates if the command has finished processing
+ */
+struct legacy_response {
+ __le32 len;
+ __le32 buf_offset;
+ __le32 is_complete;
+};
+
+/**
+ * legacy_command_to_response() - Get a pointer to a legacy_response
+ * @cmd: command
+ *
+ * Returns a pointer to a response for a command.
+ */
+static inline struct legacy_response *legacy_command_to_response(
+ const struct legacy_command *cmd)
+{
+ return (void *)cmd + le32_to_cpu(cmd->resp_hdr_offset);
+}
+
+/**
+ * legacy_get_command_buffer() - Get a pointer to a command buffer
+ * @cmd: command
+ *
+ * Returns a pointer to the command buffer of a command.
+ */
+static inline void *legacy_get_command_buffer(const struct legacy_command *cmd)
+{
+ return (void *)cmd->buf;
+}
+
+/**
+ * legacy_get_response_buffer() - Get a pointer to a response buffer
+ * @rsp: response
+ *
+ * Returns a pointer to a response buffer of a response.
+ */
+static inline void *legacy_get_response_buffer(const struct legacy_response *rsp)
+{
+ return (void *)rsp + le32_to_cpu(rsp->buf_offset);
+}
+
+static void __qcom_scm_call_do(const struct arm_smccc_args *smc,
+ struct arm_smccc_res *res)
+{
+ do {
+ arm_smccc_smc(smc->a[0], smc->a[1], smc->a[2], smc->a[3],
+ smc->a[4], smc->a[5], smc->a[6], smc->a[7], res);
+ } while (res->a0 == QCOM_SCM_INTERRUPTED);
+}
+
+/**
+ * qcom_scm_call_legacy() - Send an SCM command
+ * @dev: struct device
+ * @svc_id: service identifier
+ * @cmd_id: command identifier
+ * @cmd_buf: command buffer
+ * @cmd_len: length of the command buffer
+ * @resp_buf: response buffer
+ * @resp_len: length of the response buffer
+ *
+ * Sends a command to the SCM and waits for the command to finish processing.
+ *
+ * A note on cache maintenance:
+ * Note that any buffers that are expected to be accessed by the secure world
+ * must be flushed before invoking qcom_scm_call and invalidated in the cache
+ * immediately after qcom_scm_call returns. Cache maintenance on the command
+ * and response buffers is taken care of by qcom_scm_call; however, callers are
+ * responsible for any other cached buffers passed over to the secure world.
+ */
+static int qcom_scm_call_legacy(struct device *dev, struct qcom_scm_desc *desc)
+{
+ int arglen = desc->arginfo & 0xf;
+ int ret = 0, context_id;
+ size_t i;
+ struct legacy_command *cmd;
+ struct legacy_response *rsp;
+ struct arm_smccc_args smc = {0};
+ struct arm_smccc_res res;
+ const size_t cmd_len = arglen * sizeof(__le32);
+ const size_t resp_len = MAX_QCOM_SCM_RETS * sizeof(__le32);
+ size_t alloc_len = sizeof(*cmd) + cmd_len + sizeof(*rsp) + resp_len;
+ dma_addr_t cmd_phys;
+ __le32 *arg_buf;
+ __le32 *res_buf;
+
+ cmd = kzalloc(PAGE_ALIGN(alloc_len), GFP_KERNEL);
+ if (!cmd)
+ return -ENOMEM;
+
+ cmd->len = cpu_to_le32(alloc_len);
+ cmd->buf_offset = cpu_to_le32(sizeof(*cmd));
+ cmd->resp_hdr_offset = cpu_to_le32(sizeof(*cmd) + cmd_len);
+ cmd->id = cpu_to_le32(LEGACY_FUNCNUM(desc->svc, desc->cmd));
+
+ arg_buf = legacy_get_command_buffer(cmd);
+ for (i = 0; i < arglen; i++)
+ arg_buf[i] = cpu_to_le32(desc->args[i]);
+
+ rsp = legacy_command_to_response(cmd);
+
+ cmd_phys = dma_map_single(dev, cmd, alloc_len, DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, cmd_phys)) {
+ kfree(cmd);
+ return -ENOMEM;
+ }
+
+ smc.a[0] = 1;
+ smc.a[1] = (unsigned long)&context_id;
+ smc.a[2] = cmd_phys;
+
+ mutex_lock(&qcom_scm_lock);
+ __qcom_scm_call_do(&smc, &res);
+ if (res.a0 < 0)
+ ret = qcom_scm_remap_error(res.a0);
+ mutex_unlock(&qcom_scm_lock);
+ if (ret)
+ goto out;
+
+ do {
+ dma_sync_single_for_cpu(dev, cmd_phys + sizeof(*cmd) + cmd_len,
+ sizeof(*rsp), DMA_FROM_DEVICE);
+ } while (!rsp->is_complete);
+
+ dma_sync_single_for_cpu(dev, cmd_phys + sizeof(*cmd) + cmd_len +
+ le32_to_cpu(rsp->buf_offset),
+ resp_len, DMA_FROM_DEVICE);
+
+ res_buf = legacy_get_response_buffer(rsp);
+ for (i = 0; i < MAX_QCOM_SCM_RETS; i++)
+ desc->res[i] = le32_to_cpu(res_buf[i]);
+out:
+ dma_unmap_single(dev, cmd_phys, alloc_len, DMA_TO_DEVICE);
+ kfree(cmd);
+ return ret;
+}
+
+#define LEGACY_ATOMIC_N_REG_ARGS 5
+#define LEGACY_ATOMIC_FIRST_REG_IDX 2
+#define LEGACY_CLASS_REGISTER (0x2 << 8)
+#define LEGACY_MASK_IRQS BIT(5)
+#define LEGACY_ATOMIC(svc, cmd, n) ((LEGACY_FUNCNUM(svc, cmd) << 12) | \
+ LEGACY_CLASS_REGISTER | \
+ LEGACY_MASK_IRQS | \
+ (n & 0xf))
+
+/**
+ * qcom_scm_call_atomic_legacy() - Send an atomic SCM command with up to
+ * 5 arguments and 3 return values
+ *
+ * This shall only be used with commands that are guaranteed to be
+ * uninterruptable, atomic and SMP safe.
+ */
+static int qcom_scm_call_atomic_legacy(struct device *dev,
+ struct qcom_scm_desc *desc)
+{
+ int context_id;
+ struct arm_smccc_args smc = {0};
+ struct arm_smccc_res res;
+ size_t i, arglen = desc->arginfo & 0xf;
+
+ BUG_ON(arglen > LEGACY_ATOMIC_N_REG_ARGS);
+
+ smc.a[0] = LEGACY_ATOMIC(desc->svc, desc->cmd, arglen);
+ smc.a[1] = (unsigned long)&context_id;
+
+ for(i = 0; i < arglen; i++)
+ smc.a[i + LEGACY_ATOMIC_FIRST_REG_IDX] = desc->args[i];
+
+ arm_smccc_smc(smc.a[0], smc.a[1], smc.a[2], smc.a[3],
+ smc.a[4], smc.a[5], smc.a[6], smc.a[7], &res);
+
+ desc->res[0] = res.a1;
+ desc->res[1] = res.a2;
+ desc->res[2] = res.a3;
+
+ return res.a0;
+}
+
/**
* qcom_scm_call() - Invoke a syscall in the secure world
* @dev: device
@@ -194,7 +411,7 @@ static int ___qcom_scm_call_smccc(struct device *dev,
static int qcom_scm_call(struct device *dev, struct qcom_scm_desc *desc)
{
might_sleep();
- return ___qcom_scm_call_smccc(dev, desc, false);
+ return qcom_scm_call_smccc(dev, desc, false);
}
/**
@@ -210,9 +427,14 @@ static int qcom_scm_call(struct device *dev, struct qcom_scm_desc *desc)
*/
static int qcom_scm_call_atomic(struct device *dev, struct qcom_scm_desc *desc)
{
- return ___qcom_scm_call_smccc(dev, desc, true);
+ return qcom_scm_call_smccc(dev, desc, true);
}
+#define QCOM_SCM_FLAG_COLDBOOT_CPU0 0x00
+#define QCOM_SCM_FLAG_COLDBOOT_CPU1 0x01
+#define QCOM_SCM_FLAG_COLDBOOT_CPU2 0x08
+#define QCOM_SCM_FLAG_COLDBOOT_CPU3 0x20
+
/**
* qcom_scm_set_cold_boot_addr() - Set the cold boot address for cpus
* @entry: Entry point function for the cpus
@@ -224,9 +446,54 @@ static int qcom_scm_call_atomic(struct device *dev, struct qcom_scm_desc *desc)
int __qcom_scm_set_cold_boot_addr(struct device *dev, void *entry,
const cpumask_t *cpus)
{
- return -ENOTSUPP;
+ int flags = 0;
+ int cpu;
+ int scm_cb_flags[] = {
+ QCOM_SCM_FLAG_COLDBOOT_CPU0,
+ QCOM_SCM_FLAG_COLDBOOT_CPU1,
+ QCOM_SCM_FLAG_COLDBOOT_CPU2,
+ QCOM_SCM_FLAG_COLDBOOT_CPU3,
+ };
+ struct qcom_scm_desc desc = {
+ .svc = QCOM_SCM_SVC_BOOT,
+ .cmd = QCOM_SCM_BOOT_SET_ADDR,
+ .owner = ARM_SMCCC_OWNER_SIP,
+ };
+
+ if (!cpus || (cpus && cpumask_empty(cpus)))
+ return -EINVAL;
+
+ for_each_cpu(cpu, cpus) {
+ if (cpu < ARRAY_SIZE(scm_cb_flags))
+ flags |= scm_cb_flags[cpu];
+ else
+ set_cpu_present(cpu, false);
+ }
+
+ desc.args[0] = flags;
+ desc.args[1] = virt_to_phys(entry);
+ desc.arginfo = QCOM_SCM_ARGS(2);
+
+ return qcom_scm_call_atomic(dev, &desc);
}
+#define QCOM_SCM_FLAG_WARMBOOT_CPU0 0x04
+#define QCOM_SCM_FLAG_WARMBOOT_CPU1 0x02
+#define QCOM_SCM_FLAG_WARMBOOT_CPU2 0x10
+#define QCOM_SCM_FLAG_WARMBOOT_CPU3 0x40
+
+struct qcom_scm_entry {
+ int flag;
+ void *entry;
+};
+
+static struct qcom_scm_entry qcom_scm_wb[] = {
+ { .flag = QCOM_SCM_FLAG_WARMBOOT_CPU0 },
+ { .flag = QCOM_SCM_FLAG_WARMBOOT_CPU1 },
+ { .flag = QCOM_SCM_FLAG_WARMBOOT_CPU2 },
+ { .flag = QCOM_SCM_FLAG_WARMBOOT_CPU3 },
+};
+
/**
* qcom_scm_set_warm_boot_addr() - Set the warm boot address for cpus
* @dev: Device pointer
@@ -239,7 +506,37 @@ int __qcom_scm_set_cold_boot_addr(struct device *dev, void *entry,
int __qcom_scm_set_warm_boot_addr(struct device *dev, void *entry,
const cpumask_t *cpus)
{
- return -ENOTSUPP;
+ int ret;
+ int flags = 0;
+ int cpu;
+ struct qcom_scm_desc desc = {
+ .svc = QCOM_SCM_SVC_BOOT,
+ .cmd = QCOM_SCM_BOOT_SET_ADDR,
+ };
+
+ /*
+ * Reassign only if we are switching from hotplug entry point
+ * to cpuidle entry point or vice versa.
+ */
+ for_each_cpu(cpu, cpus) {
+ if (entry == qcom_scm_wb[cpu].entry)
+ continue;
+ flags |= qcom_scm_wb[cpu].flag;
+ }
+
+ /* No change in entry function */
+ if (!flags)
+ return 0;
+
+ desc.args[0] = virt_to_phys(entry);
+ desc.args[1] = flags;
+ ret = qcom_scm_call(dev, &desc);
+ if (!ret) {
+ for_each_cpu(cpu, cpus)
+ qcom_scm_wb[cpu].entry = entry;
+ }
+
+ return ret;
}
/**
@@ -252,6 +549,15 @@ int __qcom_scm_set_warm_boot_addr(struct device *dev, void *entry,
*/
void __qcom_scm_cpu_power_down(struct device *dev, u32 flags)
{
+ struct qcom_scm_desc desc = {
+ .svc = QCOM_SCM_SVC_BOOT,
+ .cmd = QCOM_SCM_BOOT_TERMINATE_PC,
+ .args[0] = flags & QCOM_SCM_FLUSH_FLAG_MASK,
+ .arginfo = QCOM_SCM_ARGS(1),
+ .owner = ARM_SMCCC_OWNER_SIP,
+ };
+
+ qcom_scm_call_atomic(dev, &desc);
}
int __qcom_scm_set_remote_state(struct device *dev, u32 state, u32 id)
@@ -284,7 +590,7 @@ int __qcom_scm_set_dload_mode(struct device *dev, bool enable)
desc.args[1] = enable ? QCOM_SCM_BOOT_SET_DLOAD_MODE : 0;
desc.arginfo = QCOM_SCM_ARGS(2);
- return qcom_scm_call(dev, &desc);
+ return qcom_scm_call_atomic(dev, &desc);
}
bool __qcom_scm_pas_supported(struct device *dev, u32 peripheral)
@@ -408,7 +714,7 @@ int __qcom_scm_io_readl(struct device *dev, phys_addr_t addr,
desc.args[0] = addr;
desc.arginfo = QCOM_SCM_ARGS(1);
- ret = qcom_scm_call(dev, &desc);
+ ret = qcom_scm_call_atomic(dev, &desc);
if (ret >= 0)
*val = desc.res[0];
@@ -427,7 +733,7 @@ int __qcom_scm_io_writel(struct device *dev, phys_addr_t addr, unsigned int val)
desc.args[1] = val;
desc.arginfo = QCOM_SCM_ARGS(2);
- return qcom_scm_call(dev, &desc);
+ return qcom_scm_call_atomic(dev, &desc);
}
int __qcom_scm_is_call_available(struct device *dev, u32 svc_id, u32 cmd_id)
--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
Powered by blists - more mailing lists