Re: [PATCH v11 08/26] gunyah: rsc_mgr: Add resource manager RPC core

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 





On 3/31/2023 7:25 AM, Alex Elder wrote:
On 3/3/23 7:06 PM, Elliot Berman wrote:
The resource manager is a special virtual machine which is always
running on a Gunyah system. It provides APIs for creating and destroying
VMs, secure memory management, sharing/lending of memory between VMs,
and setup of inter-VM communication. Calls to the resource manager are
made via message queues.

This patch implements the basic probing and RPC mechanism to make those
API calls. Request/response calls can be made with gh_rm_call.
Drivers can also register to notifications pushed by RM via
gh_rm_register_notifier

Specific API calls that resource manager supports will be implemented in
subsequent patches.

Mostly very simple issues noted here.    -Alex

Signed-off-by: Elliot Berman <quic_eberman@xxxxxxxxxxx>
---
  drivers/virt/gunyah/Makefile   |   3 +
  drivers/virt/gunyah/rsc_mgr.c  | 688 +++++++++++++++++++++++++++++++++
  drivers/virt/gunyah/rsc_mgr.h  |  16 +
  include/linux/gunyah_rsc_mgr.h |  21 +
  4 files changed, 728 insertions(+)
  create mode 100644 drivers/virt/gunyah/rsc_mgr.c
  create mode 100644 drivers/virt/gunyah/rsc_mgr.h
  create mode 100644 include/linux/gunyah_rsc_mgr.h

diff --git a/drivers/virt/gunyah/Makefile b/drivers/virt/gunyah/Makefile
index 34f32110faf9..cc864ff5abbb 100644
--- a/drivers/virt/gunyah/Makefile
+++ b/drivers/virt/gunyah/Makefile
@@ -1,3 +1,6 @@
  # SPDX-License-Identifier: GPL-2.0
  obj-$(CONFIG_GUNYAH) += gunyah.o
+
+gunyah_rsc_mgr-y += rsc_mgr.o
+obj-$(CONFIG_GUNYAH) += gunyah_rsc_mgr.o
diff --git a/drivers/virt/gunyah/rsc_mgr.c b/drivers/virt/gunyah/rsc_mgr.c
new file mode 100644
index 000000000000..67813c9a52db
--- /dev/null
+++ b/drivers/virt/gunyah/rsc_mgr.c
@@ -0,0 +1,688 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
+ */
+

. . .

+static void gh_rm_try_complete_connection(struct gh_rm *rm)
+{
+    struct gh_rm_connection *connection = rm->active_rx_connection;
+
+    if (!connection || connection->fragments_received != connection->num_fragments)
+        return;
+
+    switch (connection->type) {
+    case RM_RPC_TYPE_REPLY:
+        complete(&connection->reply.seq_done);
+        break;
+    case RM_RPC_TYPE_NOTIF:
+        schedule_work(&connection->notification.work);
+        break;
+    default:
+        dev_err_ratelimited(rm->dev, "Invalid message type (%d) received\n",

s/%d/%u/

+                    connection->type);
+        gh_rm_abort_connection(rm);
+        break;
+    }
+
+    rm->active_rx_connection = NULL;
+}
+
+static void gh_rm_msgq_rx_data(struct mbox_client *cl, void *mssg)
+{
+    struct gh_rm *rm = container_of(cl, struct gh_rm, msgq_client);
+    struct gh_msgq_rx_data *rx_data = mssg;
+    size_t msg_size = rx_data->length;
+    void *msg = rx_data->data;
+    struct gh_rm_rpc_hdr *hdr;
+
+    if (msg_size < sizeof(*hdr) || msg_size > GH_MSGQ_MAX_MSG_SIZE)
+        return;
+
+    hdr = msg;
+    if (hdr->api != RM_RPC_API) {
+        dev_err(rm->dev, "Unknown RM RPC API version: %x\n", hdr->api);
+        return;
+    }
+
+    switch (FIELD_GET(RM_RPC_TYPE_MASK, hdr->type)) {
+    case RM_RPC_TYPE_NOTIF:
+        gh_rm_process_notif(rm, msg, msg_size);
+        break;
+    case RM_RPC_TYPE_REPLY:
+        gh_rm_process_rply(rm, msg, msg_size);
+        break;
+    case RM_RPC_TYPE_CONTINUATION:
+        gh_rm_process_cont(rm, rm->active_rx_connection, msg, msg_size);
+        break;
+    default:
+        dev_err(rm->dev, "Invalid message type (%lu) received\n",
+            FIELD_GET(RM_RPC_TYPE_MASK, hdr->type));
+        return;
+    }
+
+    gh_rm_try_complete_connection(rm);
+}
+
+static void gh_rm_msgq_tx_done(struct mbox_client *cl, void *mssg, int r)
+{
+    struct gh_rm *rm = container_of(cl, struct gh_rm, msgq_client);
+
+    kmem_cache_free(rm->cache, mssg);
+    rm->last_tx_ret = r;
+}
+
+static int gh_rm_send_request(struct gh_rm *rm, u32 message_id,
+                  const void *req_buff, size_t req_buf_size,
+                  struct gh_rm_connection *connection)
+{
+    size_t buf_size_remaining = req_buf_size;
+    const void *req_buf_curr = req_buff;
+    struct gh_msgq_tx_data *msg;
+    struct gh_rm_rpc_hdr *hdr, hdr_template;
+    u32 cont_fragments = 0;
+    size_t payload_size;
+    void *payload;
+    int ret;
+
+    if (req_buf_size > GH_RM_MAX_NUM_FRAGMENTS * GH_RM_MAX_MSG_SIZE) {
+        dev_warn(rm->dev, "Limit exceeded for the number of fragments: %u\n",
+            cont_fragments);

You are printing the value of cont_fragments here when it's just zero.

+        dump_stack();
+        return -E2BIG;
+    }
+

Move the computation of cont_fragments prior to the block above.
You could use a ?: statement to assign it.

+    if (req_buf_size)
+        cont_fragments = (req_buf_size - 1) / GH_RM_MAX_MSG_SIZE;
+
+    hdr_template.api = RM_RPC_API;
+    hdr_template.type = FIELD_PREP(RM_RPC_TYPE_MASK, RM_RPC_TYPE_REQUEST) |
+            FIELD_PREP(RM_RPC_FRAGMENTS_MASK, cont_fragments);

The line above should be indented further.

+    hdr_template.seq = cpu_to_le16(connection->reply.seq);
+    hdr_template.msg_id = cpu_to_le32(message_id);
+
+    ret = mutex_lock_interruptible(&rm->send_lock);
+    if (ret)
+        return ret;
+
+    /* Consider also the 'request' packet for the loop count */

I don't think the comment above is helpful.

+    do {
+        msg = kmem_cache_zalloc(rm->cache, GFP_KERNEL);
+        if (!msg) {
+            ret = -ENOMEM;
+            goto out;
+        }
+
+        /* Fill header */
+        hdr = (struct gh_rm_rpc_hdr *)msg->data;

I personally would prefer &msg->data[0] in this case.

+        *hdr = hdr_template;
+
+        /* Copy payload */
+        payload = hdr + 1;

I think I might have suggested using "hdr + 1" here.

Elsewhere you use something like:
     payload = (char *)hdr + sizeof(hdr);
or something similar.  I suggest you choose one approach and use
it consistently througout the driver.  Either is fine, but I
have a slight preference for the "hdr + 1" way.


I think you might be referencing the memcpy in gh_rm_init_connection_payload. In the gh_rm_init_connection_payload, hdr_size is not fixed: for notifications, it's just the RPC header. For responses, there is the RPC header + the "RM error code". To be able to re-use same header processing, I'd have to do byte arithmetic rather than the "hdr + 1" way. I also prefer the "hdr + 1" way, but if I am going to be consistent, need to stick with byte arithmetic.

+        payload_size = min(buf_size_remaining, GH_RM_MAX_MSG_SIZE);
+        memcpy(payload, req_buf_curr, payload_size);
+        req_buf_curr += payload_size;
+        buf_size_remaining -= payload_size;
+
+        /* Force the last fragment to immediately alert the receiver */
+        msg->push = !buf_size_remaining;
+        msg->length = sizeof(*hdr) + payload_size;
+
+        ret = mbox_send_message(gh_msgq_chan(&rm->msgq), msg);
+        if (ret < 0) {
+            kmem_cache_free(rm->cache, msg);
+            break;
+        }
+
+        if (rm->last_tx_ret) {
+            ret = rm->last_tx_ret;
+            break;
+        }
+
+        hdr_template.type = FIELD_PREP(RM_RPC_TYPE_MASK, RM_RPC_TYPE_CONTINUATION) |
+                    FIELD_PREP(RM_RPC_FRAGMENTS_MASK, cont_fragments);
+    } while (buf_size_remaining);
+
+out:
+    mutex_unlock(&rm->send_lock);
+    return ret < 0 ? ret : 0;
+}
+
+/**
+ * gh_rm_call: Achieve request-response type communication with RPC
+ * @rm: Pointer to Gunyah resource manager internal data
+ * @message_id: The RM RPC message-id
+ * @req_buff: Request buffer that contains the payload
+ * @req_buf_size: Total size of the payload
+ * @resp_buf: Pointer to a response buffer
+ * @resp_buf_size: Size of the response buffer
+ *
+ * Make a request to the RM-VM and wait for reply back. For a successful

I think you could just say "to the RM and wait"...

Overall I suggest using "RM" or "RM VM" consistently when you talk
about the Resource Manager.  This is the only place I see "RM-VM".

+ * response, the function returns the payload. The size of the payload is set in + * resp_buf_size. The resp_buf should be freed by the caller when 0 is returned

s/should/must/

+ * and resp_buf_size != 0.
+ *
+ * req_buff should be not NULL for req_buf_size >0. If req_buf_size == 0,
+ * req_buff *can* be NULL and no additional payload is sent.

I'd say use "buf" or "buff" but not both in your naming
convention.


Not intentional -- will make it consistent.

+ *
+ * Context: Process context. Will sleep waiting for reply.
+ * Return: 0 on success. <0 if error.
+ */
+int gh_rm_call(struct gh_rm *rm, u32 message_id, void *req_buff, size_t req_buf_size,
+        void **resp_buf, size_t *resp_buf_size)

I suspect you could define the request buffer as a pointer to const;
can you?


I can!

+{
+    struct gh_rm_connection *connection;
+    u32 seq_id;
+    int ret;
+
+    /* message_id 0 is reserved. req_buf_size implies req_buf is not NULL */
+    if (!message_id || (!req_buff && req_buf_size) || !rm)

If you're going to check for a null RM pointer, I'd check it first.

+        return -EINVAL;
+
+
+    connection = kzalloc(sizeof(*connection), GFP_KERNEL);
+    if (!connection)
+        return -ENOMEM;
+
+    connection->type = RM_RPC_TYPE_REPLY;
+    connection->msg_id = cpu_to_le32(message_id);
+
+    init_completion(&connection->reply.seq_done);

. . .

diff --git a/include/linux/gunyah_rsc_mgr.h b/include/linux/gunyah_rsc_mgr.h
new file mode 100644
index 000000000000..deca9b3da541
--- /dev/null
+++ b/include/linux/gunyah_rsc_mgr.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
+ */
+
+#ifndef _GUNYAH_RSC_MGR_H
+#define _GUNYAH_RSC_MGR_H
+
+#include <linux/list.h>
+#include <linux/notifier.h>
+#include <linux/gunyah.h>
+
+#define GH_VMID_INVAL    U16_MAX

Add a tab before U16_MAX; it will line up more nicely
when you define GH_MEM_HANDLE_INVAL later.

+
+struct gh_rm;
+int gh_rm_notifier_register(struct gh_rm *rm, struct notifier_block *nb); +int gh_rm_notifier_unregister(struct gh_rm *rm, struct notifier_block *nb);
+struct device *gh_rm_get(struct gh_rm *rm);
+void gh_rm_put(struct gh_rm *rm);
+
+#endif




[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux